netcdf-c/libnczarr/zxcache.c

934 lines
27 KiB
C
Raw Normal View History

/* Copyright 2018, University Corporation for Atmospheric
* Research. See COPYRIGHT file for copying and redistribution
* conditions. */
/**
* @file @internal The functions which control NCZ
* caching. These caching controls allow the user to change the cache
* sizes of ZARR before opening files.
*
* @author Dennis Heimbigner, Ed Hartnett
*/
#include "zincludes.h"
#include "zcache.h"
#include "ncxcache.h"
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
#include "zfilter.h"
#include <stddef.h>
#undef DEBUG
#undef FLUSH
#define LEAFLEN 32
#define USEPARAMSIZE 0xffffffffffffffff
/* Forward */
static int get_chunk(NCZChunkCache* cache, NCZCacheEntry* entry);
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
static int put_chunk(NCZChunkCache* cache, NCZCacheEntry*);
static int verifycache(NCZChunkCache* cache);
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
static int flushcache(NCZChunkCache* cache);
static int constraincache(NCZChunkCache* cache, size64_t needed);
Mitigate S3 test interference + Unlimited Dimensions in NCZarr This PR started as an attempt to add unlimited dimensions to NCZarr. It did that, but this exposed significant problems with test interference. So this PR is mostly about fixing -- well mitigating anyway -- test interference. The problem of test interference is now documented in the document docs/internal.md. The solutions implemented here are also describe in that document. The solution is somewhat fragile but multiple cleanup mechanisms are provided. Note that this feature requires that the AWS command line utility must be installed. ## Unlimited Dimensions. The existing NCZarr extensions to Zarr are modified to support unlimited dimensions. NCzarr extends the Zarr meta-data for the ".zgroup" object to include netcdf-4 model extensions. This information is stored in ".zgroup" as dictionary named "_nczarr_group". Inside "_nczarr_group", there is a key named "dims" that stores information about netcdf-4 named dimensions. The value of "dims" is a dictionary whose keys are the named dimensions. The value associated with each dimension name has one of two forms Form 1 is a special case of form 2, and is kept for backward compatibility. Whenever a new file is written, it uses format 1 if possible, otherwise format 2. * Form 1: An integer representing the size of the dimension, which is used for simple named dimensions. * Form 2: A dictionary with the following keys and values" - "size" with an integer value representing the (current) size of the dimension. - "unlimited" with a value of either "1" or "0" to indicate if this dimension is an unlimited dimension. For Unlimited dimensions, the size is initially zero, and as variables extend the length of that dimension, the size value for the dimension increases. That dimension size is shared by all arrays referencing that dimension, so if one array extends an unlimited dimension, it is implicitly extended for all other arrays that reference that dimension. This is the standard semantics for unlimited dimensions. Adding unlimited dimensions required a number of other changes to the NCZarr code-base. These included the following. * Did a partial refactor of the slice handling code in zwalk.c to clean it up. * Added a number of tests for unlimited dimensions derived from the same test in nc_test4. * Added several NCZarr specific unlimited tests; more are needed. * Add test of endianness. ## Misc. Other Changes * Modify libdispatch/ncs3sdk_aws.cpp to optionally support use of the AWS Transfer Utility mechanism. This is controlled by the ```#define TRANSFER```` command in that file. It defaults to being disabled. * Parameterize both the standard Unidata S3 bucket (S3TESTBUCKET) and the netcdf-c test data prefix (S3TESTSUBTREE). * Fixed an obscure memory leak in ncdump. * Removed some obsolete unit testing code and test cases. * Uncovered a bug in the netcdf-c handling of big-endian floats and doubles. Have not fixed yet. See tst_h5_endians.c. * Renamed some nczarr_tests testcases to avoid name conflicts with nc_test4. * Modify the semantics of zmap\#ncsmap_write to only allow total rewrite of objects. * Modify the semantics of zodom to properly handle stride > 1. * Add a truncate operation to the libnczarr zmap code.
2023-09-27 06:56:48 +08:00
static void
setmodified(NCZCacheEntry* e, int tf)
{
e->modified = tf;
}
/**************************************************/
/* Dispatch table per-var cache functions */
/**
* @internal Set chunk cache size for a variable. This is the internal
* function called by nc_set_var_chunk_cache().
*
* @param ncid File ID.
* @param varid Variable ID.
* @param size Size in bytes to set cache.
* @param nelems # of entries in cache
* @param preemption Controls cache swapping.
*
* @returns ::NC_NOERR No error.
* @returns ::NC_EBADID Bad ncid.
* @returns ::NC_ENOTVAR Invalid variable ID.
* @returns ::NC_ESTRICTNC3 Attempting netcdf-4 operation on strict
* nc3 netcdf-4 file.
* @returns ::NC_EINVAL Invalid input.
* @returns ::NC_EHDFERR HDF5 error.
* @author Ed Hartnett
*/
int
NCZ_set_var_chunk_cache(int ncid, int varid, size_t cachesize, size_t nelems, float preemption)
{
NC_GRP_INFO_T *grp;
NC_FILE_INFO_T *h5;
NC_VAR_INFO_T *var;
NCZ_VAR_INFO_T *zvar;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
int retval = NC_NOERR;
/* Check input for validity. */
if (preemption < 0 || preemption > 1)
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
{retval = NC_EINVAL; goto done;}
/* Find info for this file and group, and set pointer to each. */
if ((retval = nc4_find_nc_grp_h5(ncid, NULL, &grp, &h5)))
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
goto done;
assert(grp && h5);
/* Find the var. */
if (!(var = (NC_VAR_INFO_T *)ncindexith(grp->vars, varid)))
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
{retval = NC_ENOTVAR; goto done;}
assert(var && var->hdr.id == varid);
zvar = (NCZ_VAR_INFO_T*)var->format_var_info;
assert(zvar != NULL && zvar->cache != NULL);
/* Set the values. */
Add support for setting HDF5 alignment property when creating a file re: https://github.com/Unidata/netcdf-c/issues/2177 re: https://github.com/Unidata/netcdf-c/pull/2178 Provide get/set functions to store global data alignment information and apply it when a file is created. The api is as follows: ```` int nc_set_alignment(int threshold, int alignment); int nc_get_alignment(int* thresholdp, int* alignmentp); ```` If defined, then for every file created opened after the call to nc_set_alignment, for every new variable added to the file, the most recently set threshold and alignment values will be applied to that variable. The nc_get_alignment function return the last values set by nc_set_alignment. If nc_set_alignment has not been called, then it returns the value 0 for both threshold and alignment. The alignment parameters are stored in the NCglobalstate object (see below) for use as needed. Repeated calls to nc_set_alignment will overwrite any existing values in NCglobalstate. The alignment parameters are applied in libhdf5/hdf5create.c and libhdf5/hdf5open.c The set/get alignment functions are defined in libsrc4/nc4internal.c. A test program was added as nc_test4/tst_alignment.c. ## Misc. Changes Unrelated to Alignment * The NCRCglobalstate type was renamed to NCglobalstate to indicate that it represented more general global state than just .rc data. It was also moved to nc4internal.h. This led to a large number of small changes: mostly renaming. The global state management functions were moved to nc4internal.c. * The global chunk cache variables have been moved into NCglobalstate. As warranted, other global state will be moved as well. * Some misc. problems with the nczarr performance tests were corrected.
2022-01-30 06:27:52 +08:00
var->chunkcache.size = cachesize;
var->chunkcache.nelems = nelems;
var->chunkcache.preemption = preemption;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* Fix up cache */
if((retval = NCZ_adjust_var_cache(var))) goto done;
done:
return retval;
}
/**
* @internal Adjust the chunk cache of a var for better
* performance.
*
* @note For contiguous and compact storage vars, or when parallel I/O
* is in use, this function will do nothing and return ::NC_NOERR;
*
* @param grp Pointer to group info struct.
* @param var Pointer to var info struct.
*
* @return ::NC_NOERR No error.
* @author Ed Hartnett
*/
int
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
NCZ_adjust_var_cache(NC_VAR_INFO_T *var)
{
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
int stat = NC_NOERR;
NCZ_VAR_INFO_T* zvar = (NCZ_VAR_INFO_T*)var->format_var_info;
NCZChunkCache* zcache = NULL;
zcache = zvar->cache;
if(zcache->valid) goto done;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
#ifdef DEBUG
fprintf(stderr,"xxx: adjusting cache for: %s\n",var->hdr.name);
#endif
/* completely empty the cache */
flushcache(zcache);
/* Reclaim any existing fill_chunk */
if((stat = NCZ_reclaim_fill_chunk(zcache))) goto done;
/* Reset the parameters */
zvar->cache->params.size = var->chunkcache.size;
zvar->cache->params.nelems = var->chunkcache.nelems;
zvar->cache->params.preemption = var->chunkcache.preemption;
#ifdef DEBUG
fprintf(stderr,"%s.cache.adjust: size=%ld nelems=%ld\n",
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
var->hdr.name,(unsigned long)zvar->cache->maxsize,(unsigned long)zvar->cache->maxentries);
#endif
/* One more thing, adjust the chunksize and count*/
zcache->chunksize = zvar->chunksize;
zcache->chunkcount = 1;
if(var->ndims > 0) {
int i;
for(i=0;i<var->ndims;i++) {
zcache->chunkcount *= var->chunksizes[i];
}
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
}
zcache->valid = 1;
done:
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
return stat;
}
/**************************************************/
/**
* Create a chunk cache object
*
* @param var containing var
* @param entrysize Size in bytes of an entry
* @param cachep return cache pointer
*
* @return ::NC_NOERR No error.
* @return ::NC_EINVAL Bad preemption.
* @author Dennis Heimbigner, Ed Hartnett
*/
int
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
NCZ_create_chunk_cache(NC_VAR_INFO_T* var, size64_t chunksize, char dimsep, NCZChunkCache** cachep)
{
int stat = NC_NOERR;
NCZChunkCache* cache = NULL;
void* fill = NULL;
NCZ_VAR_INFO_T* zvar = NULL;
if(chunksize == 0) return NC_EINVAL;
zvar = (NCZ_VAR_INFO_T*)var->format_var_info;
if((cache = calloc(1,sizeof(NCZChunkCache))) == NULL)
{stat = NC_ENOMEM; goto done;}
cache->var = var;
cache->ndims = var->ndims + zvar->scalar;
cache->fillchunk = NULL;
cache->chunksize = chunksize;
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
cache->dimension_separator = dimsep;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
zvar->cache = cache;
cache->chunkcount = 1;
if(var->ndims > 0) {
int i;
for(i=0;i<var->ndims;i++) {
cache->chunkcount *= var->chunksizes[i];
}
}
/* Set default cache parameters */
cache->params = NC_getglobalstate()->chunkcache;
#ifdef FLUSH
cache->maxentries = 1;
#endif
#ifdef DEBUG
fprintf(stderr,"%s.cache: nelems=%ld size=%ld\n",
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
var->hdr.name,(unsigned long)cache->maxentries,(unsigned long)cache->maxsize);
#endif
if((stat = ncxcachenew(LEAFLEN,&cache->xcache))) goto done;
if((cache->mru = nclistnew()) == NULL)
{stat = NC_ENOMEM; goto done;}
nclistsetalloc(cache->mru,cache->params.nelems);
if(cachep) {*cachep = cache; cache = NULL;}
done:
nullfree(fill);
NCZ_free_chunk_cache(cache);
return THROW(stat);
}
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
static void
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
free_cache_entry(NCZChunkCache* cache, NCZCacheEntry* entry)
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
{
if(entry) {
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
int tid = cache->var->type_info->hdr.id;
if(tid == NC_STRING && !entry->isfixedstring) {
Improve performance of the nc_reclaim_data and nc_copy_data functions. re: Issue https://github.com/Unidata/netcdf-c/issues/2685 re: PR https://github.com/Unidata/netcdf-c/pull/2179 As noted in PR https://github.com/Unidata/netcdf-c/pull/2179, the old code did not allow for reclaiming instances of types, nor for properly copying them. That PR provided new functions capable of reclaiming/copying instances of arbitrary types. However, as noted by Issue https://github.com/Unidata/netcdf-c/issues/2685, using these most general functions resulted in a significant performance degradation, even for common cases. This PR attempts to mitigate the cost of using the general reclaim/copy functions in two ways. First, the previous functions operating at the top level by using ncid and typeid arguments. These functions were augmented with equivalent versions that used the netcdf-c library internal data structures to allow direct access to needed information. These new functions are used internally to the library. The second mitigation involves optimizing the internal functions by providing early tests for common cases. This avoids unnecessary recursive function calls. The overall result is a significant improvement in speed by a factor of roughly twenty -- your mileage may vary. These optimized functions are still not as fast as the original (more limited) functions, but they are getting close. Additional optimizations are possible. But the cost is a significant "uglification" of the code that I deemed a step too far, at least for now. ## Misc. Changes 1. Added a test case to check the proper reclamation/copy of complex types. 2. Found and fixed some places where nc_reclaim/copy should have been used. 3. Replaced, in the netcdf-c library, (almost all) occurrences of nc_reclaim_copy with calls to NC_reclaim/copy. This plus the optimizations is the primary speed-up mechanism. 4. In DAP4, the metadata is held in a substrate in-memory file; this required some changes so that the reclaim/copy code accessed that substrate dispatcher rather than the DAP4 dispatcher. 5. Re-factored and isolated the code that computes if a type is (transitively) variable-sized or not. 6. Clean up the reclamation code in ncgen; adding the use of nc_reclaim exposed some memory problems.
2023-05-21 07:11:25 +08:00
NC_reclaim_data(cache->var->container->nc4_info->controller,tid,entry->data,cache->chunkcount);
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
}
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
nullfree(entry->data);
nullfree(entry->key.varkey);
nullfree(entry->key.chunkkey);
nullfree(entry);
}
}
void
NCZ_free_chunk_cache(NCZChunkCache* cache)
{
if(cache == NULL) return;
ZTRACE(4,"cache.var=%s",cache->var->hdr.name);
/* Iterate over the entries */
while(nclistlength(cache->mru) > 0) {
void* ptr;
NCZCacheEntry* entry = nclistremove(cache->mru,0);
(void)ncxcacheremove(cache->xcache,entry->hashkey,&ptr);
assert(ptr == entry);
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
free_cache_entry(cache,entry);
}
#ifdef DEBUG
fprintf(stderr,"|cache.free|=%ld\n",nclistlength(cache->mru));
#endif
ncxcachefree(cache->xcache);
nclistfree(cache->mru);
cache->mru = NULL;
(void)NCZ_reclaim_fill_chunk(cache);
nullfree(cache);
(void)ZUNTRACE(NC_NOERR);
}
size64_t
NCZ_cache_entrysize(NCZChunkCache* cache)
{
assert(cache);
return cache->chunksize;
}
/* Return number of active entries in cache */
size64_t
NCZ_cache_size(NCZChunkCache* cache)
{
assert(cache);
return nclistlength(cache->mru);
}
int
NCZ_read_cache_chunk(NCZChunkCache* cache, const size64_t* indices, void** datap)
{
int stat = NC_NOERR;
int rank = cache->ndims;
NCZCacheEntry* entry = NULL;
ncexhashkey_t hkey = 0;
int created = 0;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* the hash key */
hkey = ncxcachekey(indices,sizeof(size64_t)*cache->ndims);
/* See if already in cache */
stat = ncxcachelookup(cache->xcache,hkey,(void**)&entry);
switch(stat) {
case NC_NOERR:
/* Move to front of the lru */
(void)ncxcachetouch(cache->xcache,hkey);
break;
case NC_ENOOBJECT:
entry = NULL; /* not found; */
break;
default: goto done;
}
if(entry == NULL) { /*!found*/
/* Create a new entry */
if((entry = calloc(1,sizeof(NCZCacheEntry)))==NULL)
{stat = NC_ENOMEM; goto done;}
memcpy(entry->indices,indices,rank*sizeof(size64_t));
/* Create the key for this cache */
if((stat = NCZ_buildchunkpath(cache,indices,&entry->key))) goto done;
entry->hashkey = hkey;
assert(entry->data == NULL && entry->size == 0);
/* Try to read the object from "disk"; might change size; will create if non-existent */
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
if((stat=get_chunk(cache,entry))) goto done;
assert(entry->data != NULL);
/* Ensure cache constraints not violated; but do it before entry is added */
if((stat=verifycache(cache))) goto done;
nclistpush(cache->mru,entry);
if((stat = ncxcacheinsert(cache->xcache,entry->hashkey,entry))) goto done;
}
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
#ifdef DEBUG
fprintf(stderr,"|cache.read.lru|=%ld\n",nclistlength(cache->mru));
#endif
if(datap) *datap = entry->data;
entry = NULL;
done:
if(created && stat == NC_NOERR) stat = NC_EEMPTY; /* tell upper layers */
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
if(entry) free_cache_entry(cache,entry);
return THROW(stat);
}
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
#if 0
int
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
NCZ_write_cache_chunk(NCZChunkCache* cache, const size64_t* indices, void* content)
{
int stat = NC_NOERR;
int rank = cache->ndims;
NCZCacheEntry* entry = NULL;
ncexhashkey_t hkey;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* create the hash key */
hkey = ncxcachekey(indices,sizeof(size64_t)*cache->ndims);
if(entry == NULL) { /*!found*/
/* Create a new entry */
if((entry = calloc(1,sizeof(NCZCacheEntry)))==NULL)
{stat = NC_ENOMEM; goto done;}
memcpy(entry->indices,indices,rank*sizeof(size64_t));
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
if((stat = NCZ_buildchunkpath(cache,indices,&entry->key))) goto done;
entry->hashkey = hkey;
/* Create the local copy space */
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
entry->size = cache->chunksize;
if((entry->data = calloc(1,cache->chunksize)) == NULL)
{stat = NC_ENOMEM; goto done;}
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
memcpy(entry->data,content,cache->chunksize);
}
Mitigate S3 test interference + Unlimited Dimensions in NCZarr This PR started as an attempt to add unlimited dimensions to NCZarr. It did that, but this exposed significant problems with test interference. So this PR is mostly about fixing -- well mitigating anyway -- test interference. The problem of test interference is now documented in the document docs/internal.md. The solutions implemented here are also describe in that document. The solution is somewhat fragile but multiple cleanup mechanisms are provided. Note that this feature requires that the AWS command line utility must be installed. ## Unlimited Dimensions. The existing NCZarr extensions to Zarr are modified to support unlimited dimensions. NCzarr extends the Zarr meta-data for the ".zgroup" object to include netcdf-4 model extensions. This information is stored in ".zgroup" as dictionary named "_nczarr_group". Inside "_nczarr_group", there is a key named "dims" that stores information about netcdf-4 named dimensions. The value of "dims" is a dictionary whose keys are the named dimensions. The value associated with each dimension name has one of two forms Form 1 is a special case of form 2, and is kept for backward compatibility. Whenever a new file is written, it uses format 1 if possible, otherwise format 2. * Form 1: An integer representing the size of the dimension, which is used for simple named dimensions. * Form 2: A dictionary with the following keys and values" - "size" with an integer value representing the (current) size of the dimension. - "unlimited" with a value of either "1" or "0" to indicate if this dimension is an unlimited dimension. For Unlimited dimensions, the size is initially zero, and as variables extend the length of that dimension, the size value for the dimension increases. That dimension size is shared by all arrays referencing that dimension, so if one array extends an unlimited dimension, it is implicitly extended for all other arrays that reference that dimension. This is the standard semantics for unlimited dimensions. Adding unlimited dimensions required a number of other changes to the NCZarr code-base. These included the following. * Did a partial refactor of the slice handling code in zwalk.c to clean it up. * Added a number of tests for unlimited dimensions derived from the same test in nc_test4. * Added several NCZarr specific unlimited tests; more are needed. * Add test of endianness. ## Misc. Other Changes * Modify libdispatch/ncs3sdk_aws.cpp to optionally support use of the AWS Transfer Utility mechanism. This is controlled by the ```#define TRANSFER```` command in that file. It defaults to being disabled. * Parameterize both the standard Unidata S3 bucket (S3TESTBUCKET) and the netcdf-c test data prefix (S3TESTSUBTREE). * Fixed an obscure memory leak in ncdump. * Removed some obsolete unit testing code and test cases. * Uncovered a bug in the netcdf-c handling of big-endian floats and doubles. Have not fixed yet. See tst_h5_endians.c. * Renamed some nczarr_tests testcases to avoid name conflicts with nc_test4. * Modify the semantics of zmap\#ncsmap_write to only allow total rewrite of objects. * Modify the semantics of zodom to properly handle stride > 1. * Add a truncate operation to the libnczarr zmap code.
2023-09-27 06:56:48 +08:00
setmodified(entry,1);
nclistpush(cache->mru,entry); /* MRU order */
#ifdef DEBUG
fprintf(stderr,"|cache.write|=%ld\n",nclistlength(cache->mru));
#endif
entry = NULL;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* Ensure cache constraints not violated */
if((stat=verifycache(cache))) goto done;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
done:
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
if(entry) free_cache_entry(cache,entry);
return THROW(stat);
}
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
#endif
/* Constrain cache */
static int
verifycache(NCZChunkCache* cache)
{
int stat = NC_NOERR;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
#if 0
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* Sanity check; make sure at least one entry is always allowed */
if(nclistlength(cache->mru) == 1)
goto done;
#endif
if((stat = constraincache(cache,USEPARAMSIZE))) goto done;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
done:
return stat;
}
/* Completely flush cache */
static int
flushcache(NCZChunkCache* cache)
{
int stat = NC_NOERR;
#if 0
size_t oldsize = cache->params.size;
cache->params.size = 0;
stat = constraincache(cache,USEPARAMSIZE);
cache->params.size = oldsize;
#else
stat = constraincache(cache,USEPARAMSIZE);
#endif
return stat;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
}
/* Remove entries to ensure cache is not
violating any of its constraints.
On entry, constraints might be violated.
Make sure that the entryinuse (NULL => no constraint) is not reclaimed.
@param cache
@param needed make sure there is room for this much space; USEPARAMSIZE => ensure no more than cache params is used.
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
*/
static int
constraincache(NCZChunkCache* cache, size64_t needed)
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
{
int stat = NC_NOERR;
size64_t final_size;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* If the cache is empty then do nothing */
if(cache->used == 0) goto done;
if(needed == USEPARAMSIZE)
final_size = cache->params.size;
else if(cache->used > needed)
final_size = cache->used - needed;
else
final_size = 0;
/* Flush from LRU end if we are at capacity */
while(nclistlength(cache->mru) > cache->params.nelems || cache->used > final_size) {
size_t i;
void* ptr;
NCZCacheEntry* e = ncxcachelast(cache->xcache); /* last entry is the least recently used */
if(e == NULL) break;
if((stat = ncxcacheremove(cache->xcache,e->hashkey,&ptr))) goto done;
assert(e == ptr);
for(i=0;i<nclistlength(cache->mru);i++) {
e = nclistget(cache->mru,i);
if(ptr == e) break;
}
assert(e != NULL);
assert(i >= 0 && i < nclistlength(cache->mru));
nclistremove(cache->mru,i);
assert(cache->used >= e->size);
/* Note that |old chunk data| may not be same as |new chunk data| because of filters */
cache->used -= e->size; /* old size */
if(e->modified) /* flush to file */
stat=put_chunk(cache,e);
/* reclaim */
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
nullfree(e->data); nullfree(e->key.varkey); nullfree(e->key.chunkkey); nullfree(e);
}
#ifdef DEBUG
fprintf(stderr,"|cache.makeroom|=%ld\n",nclistlength(cache->mru));
#endif
done:
return stat;
}
/**
Push modified cache entries to disk.
Also make sure the cache size is correct.
@param cache
@return NC_EXXX error
*/
int
NCZ_flush_chunk_cache(NCZChunkCache* cache)
{
int stat = NC_NOERR;
size_t i;
ZTRACE(4,"cache.var=%s |cache|=%d",cache->var->hdr.name,(int)nclistlength(cache->mru));
if(NCZ_cache_size(cache) == 0) goto done;
/* Iterate over the entries in hashmap */
for(i=0;i<nclistlength(cache->mru);i++) {
NCZCacheEntry* entry = nclistget(cache->mru,i);
if(entry->modified) {
/* Write out this chunk in toto*/
if((stat=put_chunk(cache,entry)))
goto done;
}
Mitigate S3 test interference + Unlimited Dimensions in NCZarr This PR started as an attempt to add unlimited dimensions to NCZarr. It did that, but this exposed significant problems with test interference. So this PR is mostly about fixing -- well mitigating anyway -- test interference. The problem of test interference is now documented in the document docs/internal.md. The solutions implemented here are also describe in that document. The solution is somewhat fragile but multiple cleanup mechanisms are provided. Note that this feature requires that the AWS command line utility must be installed. ## Unlimited Dimensions. The existing NCZarr extensions to Zarr are modified to support unlimited dimensions. NCzarr extends the Zarr meta-data for the ".zgroup" object to include netcdf-4 model extensions. This information is stored in ".zgroup" as dictionary named "_nczarr_group". Inside "_nczarr_group", there is a key named "dims" that stores information about netcdf-4 named dimensions. The value of "dims" is a dictionary whose keys are the named dimensions. The value associated with each dimension name has one of two forms Form 1 is a special case of form 2, and is kept for backward compatibility. Whenever a new file is written, it uses format 1 if possible, otherwise format 2. * Form 1: An integer representing the size of the dimension, which is used for simple named dimensions. * Form 2: A dictionary with the following keys and values" - "size" with an integer value representing the (current) size of the dimension. - "unlimited" with a value of either "1" or "0" to indicate if this dimension is an unlimited dimension. For Unlimited dimensions, the size is initially zero, and as variables extend the length of that dimension, the size value for the dimension increases. That dimension size is shared by all arrays referencing that dimension, so if one array extends an unlimited dimension, it is implicitly extended for all other arrays that reference that dimension. This is the standard semantics for unlimited dimensions. Adding unlimited dimensions required a number of other changes to the NCZarr code-base. These included the following. * Did a partial refactor of the slice handling code in zwalk.c to clean it up. * Added a number of tests for unlimited dimensions derived from the same test in nc_test4. * Added several NCZarr specific unlimited tests; more are needed. * Add test of endianness. ## Misc. Other Changes * Modify libdispatch/ncs3sdk_aws.cpp to optionally support use of the AWS Transfer Utility mechanism. This is controlled by the ```#define TRANSFER```` command in that file. It defaults to being disabled. * Parameterize both the standard Unidata S3 bucket (S3TESTBUCKET) and the netcdf-c test data prefix (S3TESTSUBTREE). * Fixed an obscure memory leak in ncdump. * Removed some obsolete unit testing code and test cases. * Uncovered a bug in the netcdf-c handling of big-endian floats and doubles. Have not fixed yet. See tst_h5_endians.c. * Renamed some nczarr_tests testcases to avoid name conflicts with nc_test4. * Modify the semantics of zmap\#ncsmap_write to only allow total rewrite of objects. * Modify the semantics of zodom to properly handle stride > 1. * Add a truncate operation to the libnczarr zmap code.
2023-09-27 06:56:48 +08:00
setmodified(entry,0);
}
/* Re-compute space used */
cache->used = 0;
for(i=0;i<nclistlength(cache->mru);i++) {
NCZCacheEntry* entry = nclistget(cache->mru,i);
cache->used += entry->size;
}
/* Make sure cache size and nelems are correct */
if((stat=verifycache(cache))) goto done;
done:
return ZUNTRACE(stat);
}
/* Ensure existence of some kind of fill chunk */
int
NCZ_ensure_fill_chunk(NCZChunkCache* cache)
{
int i, stat = NC_NOERR;
NC_VAR_INFO_T* var = cache->var;
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
nc_type typeid = var->type_info->hdr.id;
size_t typesize = var->type_info->size;
if(cache->fillchunk) goto done;
if((cache->fillchunk = malloc(cache->chunksize))==NULL)
{stat = NC_ENOMEM; goto done;}
if(var->no_fill) {
/* use zeros */
memset(cache->fillchunk,0,cache->chunksize);
goto done;
}
if((stat = NCZ_ensure_fill_value(var))) goto done;
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
if(typeid == NC_STRING) {
char* src = *((char**)(var->fill_value));
char** dst = (char**)(cache->fillchunk);
for(i=0;i<cache->chunkcount;i++) dst[i] = strdup(src);
} else
switch (typesize) {
case 1: {
unsigned char c = *((unsigned char*)var->fill_value);
memset(cache->fillchunk,c,cache->chunksize);
} break;
case 2: {
unsigned short fv = *((unsigned short*)var->fill_value);
unsigned short* p2 = (unsigned short*)cache->fillchunk;
for(i=0;i<cache->chunksize;i+=typesize) *p2++ = fv;
} break;
case 4: {
unsigned int fv = *((unsigned int*)var->fill_value);
unsigned int* p4 = (unsigned int*)cache->fillchunk;
for(i=0;i<cache->chunksize;i+=typesize) *p4++ = fv;
} break;
case 8: {
unsigned long long fv = *((unsigned long long*)var->fill_value);
unsigned long long* p8 = (unsigned long long*)cache->fillchunk;
for(i=0;i<cache->chunksize;i+=typesize) *p8++ = fv;
} break;
default: {
unsigned char* p;
for(p=cache->fillchunk,i=0;i<cache->chunksize;i+=typesize,p+=typesize)
memcpy(p,var->fill_value,typesize);
} break;
}
done:
return NC_NOERR;
}
int
NCZ_reclaim_fill_chunk(NCZChunkCache* zcache)
{
int stat = NC_NOERR;
if(zcache && zcache->fillchunk) {
NC_VAR_INFO_T* var = zcache->var;
int tid = var->type_info->hdr.id;
size_t chunkcount = zcache->chunkcount;
Improve performance of the nc_reclaim_data and nc_copy_data functions. re: Issue https://github.com/Unidata/netcdf-c/issues/2685 re: PR https://github.com/Unidata/netcdf-c/pull/2179 As noted in PR https://github.com/Unidata/netcdf-c/pull/2179, the old code did not allow for reclaiming instances of types, nor for properly copying them. That PR provided new functions capable of reclaiming/copying instances of arbitrary types. However, as noted by Issue https://github.com/Unidata/netcdf-c/issues/2685, using these most general functions resulted in a significant performance degradation, even for common cases. This PR attempts to mitigate the cost of using the general reclaim/copy functions in two ways. First, the previous functions operating at the top level by using ncid and typeid arguments. These functions were augmented with equivalent versions that used the netcdf-c library internal data structures to allow direct access to needed information. These new functions are used internally to the library. The second mitigation involves optimizing the internal functions by providing early tests for common cases. This avoids unnecessary recursive function calls. The overall result is a significant improvement in speed by a factor of roughly twenty -- your mileage may vary. These optimized functions are still not as fast as the original (more limited) functions, but they are getting close. Additional optimizations are possible. But the cost is a significant "uglification" of the code that I deemed a step too far, at least for now. ## Misc. Changes 1. Added a test case to check the proper reclamation/copy of complex types. 2. Found and fixed some places where nc_reclaim/copy should have been used. 3. Replaced, in the netcdf-c library, (almost all) occurrences of nc_reclaim_copy with calls to NC_reclaim/copy. This plus the optimizations is the primary speed-up mechanism. 4. In DAP4, the metadata is held in a substrate in-memory file; this required some changes so that the reclaim/copy code accessed that substrate dispatcher rather than the DAP4 dispatcher. 5. Re-factored and isolated the code that computes if a type is (transitively) variable-sized or not. 6. Clean up the reclamation code in ncgen; adding the use of nc_reclaim exposed some memory problems.
2023-05-21 07:11:25 +08:00
stat = NC_reclaim_data_all(var->container->nc4_info->controller,tid,zcache->fillchunk,chunkcount);
zcache->fillchunk = NULL;
}
return stat;
}
int
Mitigate S3 test interference + Unlimited Dimensions in NCZarr This PR started as an attempt to add unlimited dimensions to NCZarr. It did that, but this exposed significant problems with test interference. So this PR is mostly about fixing -- well mitigating anyway -- test interference. The problem of test interference is now documented in the document docs/internal.md. The solutions implemented here are also describe in that document. The solution is somewhat fragile but multiple cleanup mechanisms are provided. Note that this feature requires that the AWS command line utility must be installed. ## Unlimited Dimensions. The existing NCZarr extensions to Zarr are modified to support unlimited dimensions. NCzarr extends the Zarr meta-data for the ".zgroup" object to include netcdf-4 model extensions. This information is stored in ".zgroup" as dictionary named "_nczarr_group". Inside "_nczarr_group", there is a key named "dims" that stores information about netcdf-4 named dimensions. The value of "dims" is a dictionary whose keys are the named dimensions. The value associated with each dimension name has one of two forms Form 1 is a special case of form 2, and is kept for backward compatibility. Whenever a new file is written, it uses format 1 if possible, otherwise format 2. * Form 1: An integer representing the size of the dimension, which is used for simple named dimensions. * Form 2: A dictionary with the following keys and values" - "size" with an integer value representing the (current) size of the dimension. - "unlimited" with a value of either "1" or "0" to indicate if this dimension is an unlimited dimension. For Unlimited dimensions, the size is initially zero, and as variables extend the length of that dimension, the size value for the dimension increases. That dimension size is shared by all arrays referencing that dimension, so if one array extends an unlimited dimension, it is implicitly extended for all other arrays that reference that dimension. This is the standard semantics for unlimited dimensions. Adding unlimited dimensions required a number of other changes to the NCZarr code-base. These included the following. * Did a partial refactor of the slice handling code in zwalk.c to clean it up. * Added a number of tests for unlimited dimensions derived from the same test in nc_test4. * Added several NCZarr specific unlimited tests; more are needed. * Add test of endianness. ## Misc. Other Changes * Modify libdispatch/ncs3sdk_aws.cpp to optionally support use of the AWS Transfer Utility mechanism. This is controlled by the ```#define TRANSFER```` command in that file. It defaults to being disabled. * Parameterize both the standard Unidata S3 bucket (S3TESTBUCKET) and the netcdf-c test data prefix (S3TESTSUBTREE). * Fixed an obscure memory leak in ncdump. * Removed some obsolete unit testing code and test cases. * Uncovered a bug in the netcdf-c handling of big-endian floats and doubles. Have not fixed yet. See tst_h5_endians.c. * Renamed some nczarr_tests testcases to avoid name conflicts with nc_test4. * Modify the semantics of zmap\#ncsmap_write to only allow total rewrite of objects. * Modify the semantics of zodom to properly handle stride > 1. * Add a truncate operation to the libnczarr zmap code.
2023-09-27 06:56:48 +08:00
NCZ_chunk_cache_modify(NCZChunkCache* cache, const size64_t* indices)
{
int stat = NC_NOERR;
Mitigate S3 test interference + Unlimited Dimensions in NCZarr This PR started as an attempt to add unlimited dimensions to NCZarr. It did that, but this exposed significant problems with test interference. So this PR is mostly about fixing -- well mitigating anyway -- test interference. The problem of test interference is now documented in the document docs/internal.md. The solutions implemented here are also describe in that document. The solution is somewhat fragile but multiple cleanup mechanisms are provided. Note that this feature requires that the AWS command line utility must be installed. ## Unlimited Dimensions. The existing NCZarr extensions to Zarr are modified to support unlimited dimensions. NCzarr extends the Zarr meta-data for the ".zgroup" object to include netcdf-4 model extensions. This information is stored in ".zgroup" as dictionary named "_nczarr_group". Inside "_nczarr_group", there is a key named "dims" that stores information about netcdf-4 named dimensions. The value of "dims" is a dictionary whose keys are the named dimensions. The value associated with each dimension name has one of two forms Form 1 is a special case of form 2, and is kept for backward compatibility. Whenever a new file is written, it uses format 1 if possible, otherwise format 2. * Form 1: An integer representing the size of the dimension, which is used for simple named dimensions. * Form 2: A dictionary with the following keys and values" - "size" with an integer value representing the (current) size of the dimension. - "unlimited" with a value of either "1" or "0" to indicate if this dimension is an unlimited dimension. For Unlimited dimensions, the size is initially zero, and as variables extend the length of that dimension, the size value for the dimension increases. That dimension size is shared by all arrays referencing that dimension, so if one array extends an unlimited dimension, it is implicitly extended for all other arrays that reference that dimension. This is the standard semantics for unlimited dimensions. Adding unlimited dimensions required a number of other changes to the NCZarr code-base. These included the following. * Did a partial refactor of the slice handling code in zwalk.c to clean it up. * Added a number of tests for unlimited dimensions derived from the same test in nc_test4. * Added several NCZarr specific unlimited tests; more are needed. * Add test of endianness. ## Misc. Other Changes * Modify libdispatch/ncs3sdk_aws.cpp to optionally support use of the AWS Transfer Utility mechanism. This is controlled by the ```#define TRANSFER```` command in that file. It defaults to being disabled. * Parameterize both the standard Unidata S3 bucket (S3TESTBUCKET) and the netcdf-c test data prefix (S3TESTSUBTREE). * Fixed an obscure memory leak in ncdump. * Removed some obsolete unit testing code and test cases. * Uncovered a bug in the netcdf-c handling of big-endian floats and doubles. Have not fixed yet. See tst_h5_endians.c. * Renamed some nczarr_tests testcases to avoid name conflicts with nc_test4. * Modify the semantics of zmap\#ncsmap_write to only allow total rewrite of objects. * Modify the semantics of zodom to properly handle stride > 1. * Add a truncate operation to the libnczarr zmap code.
2023-09-27 06:56:48 +08:00
ncexhashkey_t hkey = 0;
NCZCacheEntry* entry = NULL;
Mitigate S3 test interference + Unlimited Dimensions in NCZarr This PR started as an attempt to add unlimited dimensions to NCZarr. It did that, but this exposed significant problems with test interference. So this PR is mostly about fixing -- well mitigating anyway -- test interference. The problem of test interference is now documented in the document docs/internal.md. The solutions implemented here are also describe in that document. The solution is somewhat fragile but multiple cleanup mechanisms are provided. Note that this feature requires that the AWS command line utility must be installed. ## Unlimited Dimensions. The existing NCZarr extensions to Zarr are modified to support unlimited dimensions. NCzarr extends the Zarr meta-data for the ".zgroup" object to include netcdf-4 model extensions. This information is stored in ".zgroup" as dictionary named "_nczarr_group". Inside "_nczarr_group", there is a key named "dims" that stores information about netcdf-4 named dimensions. The value of "dims" is a dictionary whose keys are the named dimensions. The value associated with each dimension name has one of two forms Form 1 is a special case of form 2, and is kept for backward compatibility. Whenever a new file is written, it uses format 1 if possible, otherwise format 2. * Form 1: An integer representing the size of the dimension, which is used for simple named dimensions. * Form 2: A dictionary with the following keys and values" - "size" with an integer value representing the (current) size of the dimension. - "unlimited" with a value of either "1" or "0" to indicate if this dimension is an unlimited dimension. For Unlimited dimensions, the size is initially zero, and as variables extend the length of that dimension, the size value for the dimension increases. That dimension size is shared by all arrays referencing that dimension, so if one array extends an unlimited dimension, it is implicitly extended for all other arrays that reference that dimension. This is the standard semantics for unlimited dimensions. Adding unlimited dimensions required a number of other changes to the NCZarr code-base. These included the following. * Did a partial refactor of the slice handling code in zwalk.c to clean it up. * Added a number of tests for unlimited dimensions derived from the same test in nc_test4. * Added several NCZarr specific unlimited tests; more are needed. * Add test of endianness. ## Misc. Other Changes * Modify libdispatch/ncs3sdk_aws.cpp to optionally support use of the AWS Transfer Utility mechanism. This is controlled by the ```#define TRANSFER```` command in that file. It defaults to being disabled. * Parameterize both the standard Unidata S3 bucket (S3TESTBUCKET) and the netcdf-c test data prefix (S3TESTSUBTREE). * Fixed an obscure memory leak in ncdump. * Removed some obsolete unit testing code and test cases. * Uncovered a bug in the netcdf-c handling of big-endian floats and doubles. Have not fixed yet. See tst_h5_endians.c. * Renamed some nczarr_tests testcases to avoid name conflicts with nc_test4. * Modify the semantics of zmap\#ncsmap_write to only allow total rewrite of objects. * Modify the semantics of zodom to properly handle stride > 1. * Add a truncate operation to the libnczarr zmap code.
2023-09-27 06:56:48 +08:00
/* the hash key */
hkey = ncxcachekey(indices,sizeof(size64_t)*cache->ndims);
/* See if already in cache */
Mitigate S3 test interference + Unlimited Dimensions in NCZarr This PR started as an attempt to add unlimited dimensions to NCZarr. It did that, but this exposed significant problems with test interference. So this PR is mostly about fixing -- well mitigating anyway -- test interference. The problem of test interference is now documented in the document docs/internal.md. The solutions implemented here are also describe in that document. The solution is somewhat fragile but multiple cleanup mechanisms are provided. Note that this feature requires that the AWS command line utility must be installed. ## Unlimited Dimensions. The existing NCZarr extensions to Zarr are modified to support unlimited dimensions. NCzarr extends the Zarr meta-data for the ".zgroup" object to include netcdf-4 model extensions. This information is stored in ".zgroup" as dictionary named "_nczarr_group". Inside "_nczarr_group", there is a key named "dims" that stores information about netcdf-4 named dimensions. The value of "dims" is a dictionary whose keys are the named dimensions. The value associated with each dimension name has one of two forms Form 1 is a special case of form 2, and is kept for backward compatibility. Whenever a new file is written, it uses format 1 if possible, otherwise format 2. * Form 1: An integer representing the size of the dimension, which is used for simple named dimensions. * Form 2: A dictionary with the following keys and values" - "size" with an integer value representing the (current) size of the dimension. - "unlimited" with a value of either "1" or "0" to indicate if this dimension is an unlimited dimension. For Unlimited dimensions, the size is initially zero, and as variables extend the length of that dimension, the size value for the dimension increases. That dimension size is shared by all arrays referencing that dimension, so if one array extends an unlimited dimension, it is implicitly extended for all other arrays that reference that dimension. This is the standard semantics for unlimited dimensions. Adding unlimited dimensions required a number of other changes to the NCZarr code-base. These included the following. * Did a partial refactor of the slice handling code in zwalk.c to clean it up. * Added a number of tests for unlimited dimensions derived from the same test in nc_test4. * Added several NCZarr specific unlimited tests; more are needed. * Add test of endianness. ## Misc. Other Changes * Modify libdispatch/ncs3sdk_aws.cpp to optionally support use of the AWS Transfer Utility mechanism. This is controlled by the ```#define TRANSFER```` command in that file. It defaults to being disabled. * Parameterize both the standard Unidata S3 bucket (S3TESTBUCKET) and the netcdf-c test data prefix (S3TESTSUBTREE). * Fixed an obscure memory leak in ncdump. * Removed some obsolete unit testing code and test cases. * Uncovered a bug in the netcdf-c handling of big-endian floats and doubles. Have not fixed yet. See tst_h5_endians.c. * Renamed some nczarr_tests testcases to avoid name conflicts with nc_test4. * Modify the semantics of zmap\#ncsmap_write to only allow total rewrite of objects. * Modify the semantics of zodom to properly handle stride > 1. * Add a truncate operation to the libnczarr zmap code.
2023-09-27 06:56:48 +08:00
if((stat=ncxcachelookup(cache->xcache, hkey, (void**)&entry))) {stat = NC_EINTERNAL; goto done;}
setmodified(entry,1);
done:
return THROW(stat);
}
/**************************************************/
/*
From Zarr V2 Specification:
"The compressed sequence of bytes for each chunk is stored under
a key formed from the index of the chunk within the grid of
chunks representing the array. To form a string key for a
chunk, the indices are converted to strings and concatenated
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
with the dimension_separator character ('.' or '/') separating
each index. For example, given an array with shape (10000,
10000) and chunk shape (1000, 1000) there will be 100 chunks
laid out in a 10 by 10 grid. The chunk with indices (0, 0)
provides data for rows 0-1000 and columns 0-1000 and is stored
under the key "0.0"; the chunk with indices (2, 4) provides data
for rows 2000-3000 and columns 4000-5000 and is stored under the
key "2.4"; etc."
*/
/**
* @param R Rank
* @param chunkindices The chunk indices
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
* @param dimsep the dimension separator
* @param keyp Return the chunk key string
*/
int
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
NCZ_buildchunkkey(size_t R, const size64_t* chunkindices, char dimsep, char** keyp)
{
int stat = NC_NOERR;
int r;
NCbytes* key = ncbytesnew();
if(keyp) *keyp = NULL;
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
assert(islegaldimsep(dimsep));
for(r=0;r<R;r++) {
char sindex[64];
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
if(r > 0) ncbytesappend(key,dimsep);
/* Print as decimal with no leading zeros */
snprintf(sindex,sizeof(sindex),"%lu",(unsigned long)chunkindices[r]);
ncbytescat(key,sindex);
}
ncbytesnull(key);
if(keyp) *keyp = ncbytesextract(key);
ncbytesfree(key);
return THROW(stat);
}
/**
* @internal Push data to chunk of a file.
* If chunk does not exist, create it
*
* @param file Pointer to file info struct.
* @param proj Chunk projection
* @param datalen size of data
* @param data Buffer containing the chunk data to write
*
* @return ::NC_NOERR No error.
* @author Dennis Heimbigner
*/
static int
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
put_chunk(NCZChunkCache* cache, NCZCacheEntry* entry)
{
int stat = NC_NOERR;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
NC_FILE_INFO_T* file = NULL;
NCZ_FILE_INFO_T* zfile = NULL;
NCZMAP* map = NULL;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
char* path = NULL;
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
nc_type tid = NC_NAT;
void* strchunk = NULL;
ZTRACE(5,"cache.var=%s entry.key=%s",cache->var->hdr.name,entry->key);
LOG((3, "%s: var: %p", __func__, cache->var));
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
file = (cache->var->container)->nc4_info;
zfile = file->format_file_info;
map = zfile->map;
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
/* Collect some info */
tid = cache->var->type_info->hdr.id;
if(tid == NC_STRING && !entry->isfixedstring) {
/* Convert from char* to char[strlen] format */
int maxstrlen = NCZ_get_maxstrlen((NC_OBJ*)cache->var);
assert(maxstrlen > 0);
if((strchunk = malloc((size_t)cache->chunkcount * (size_t)maxstrlen))==NULL) {stat = NC_ENOMEM; goto done;}
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
/* copy char* to char[] format */
if((stat = NCZ_char2fixed((const char**)entry->data,strchunk,cache->chunkcount,maxstrlen))) goto done;
/* Reclaim the old chunk */
Improve performance of the nc_reclaim_data and nc_copy_data functions. re: Issue https://github.com/Unidata/netcdf-c/issues/2685 re: PR https://github.com/Unidata/netcdf-c/pull/2179 As noted in PR https://github.com/Unidata/netcdf-c/pull/2179, the old code did not allow for reclaiming instances of types, nor for properly copying them. That PR provided new functions capable of reclaiming/copying instances of arbitrary types. However, as noted by Issue https://github.com/Unidata/netcdf-c/issues/2685, using these most general functions resulted in a significant performance degradation, even for common cases. This PR attempts to mitigate the cost of using the general reclaim/copy functions in two ways. First, the previous functions operating at the top level by using ncid and typeid arguments. These functions were augmented with equivalent versions that used the netcdf-c library internal data structures to allow direct access to needed information. These new functions are used internally to the library. The second mitigation involves optimizing the internal functions by providing early tests for common cases. This avoids unnecessary recursive function calls. The overall result is a significant improvement in speed by a factor of roughly twenty -- your mileage may vary. These optimized functions are still not as fast as the original (more limited) functions, but they are getting close. Additional optimizations are possible. But the cost is a significant "uglification" of the code that I deemed a step too far, at least for now. ## Misc. Changes 1. Added a test case to check the proper reclamation/copy of complex types. 2. Found and fixed some places where nc_reclaim/copy should have been used. 3. Replaced, in the netcdf-c library, (almost all) occurrences of nc_reclaim_copy with calls to NC_reclaim/copy. This plus the optimizations is the primary speed-up mechanism. 4. In DAP4, the metadata is held in a substrate in-memory file; this required some changes so that the reclaim/copy code accessed that substrate dispatcher rather than the DAP4 dispatcher. 5. Re-factored and isolated the code that computes if a type is (transitively) variable-sized or not. 6. Clean up the reclamation code in ncgen; adding the use of nc_reclaim exposed some memory problems.
2023-05-21 07:11:25 +08:00
if((stat = NC_reclaim_data_all(file->controller,tid,entry->data,cache->chunkcount))) goto done;
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
entry->data = NULL;
entry->data = strchunk; strchunk = NULL;
entry->size = cache->chunkcount * maxstrlen;
entry->isfixedstring = 1;
}
Support MSYS2/Mingw platform re: The current netcdf-c release has some problems with the mingw platform on windows. Mostly they are path issues. Changes to support mingw+msys2: ------------------------------- * Enable option of looking into the windows registry to find the mingw root path. In aid of proper path handling. * Add mingw+msys as a specific platform in configure.ac and move testing of the platform to the front so it is available early. * Handle mingw X libncpoco (dynamic loader) properly even though mingw does not yet support it. * Handle mingw X plugins properly even though mingw does not yet support it. * Alias pwd='pwd -W' to better handle paths in shell scripts. * Plus a number of other minor compile irritations. * Disallow the use of multiple nc_open's on the same file for windows (and mingw) because windows does not seem to handle these properly. Not sure why we did not catch this earlier. * Add mountpoint info to dpathmgr.c to help support mingw. * Cleanup dpathmgr conversions. Known problems: --------------- * I have not been able to get shared libraries to work, so plugins/filters must be disabled. * There is some kind of problem with libcurl that I have not solved, so all uses of libcurl (currently DAP+Byterange) must be disabled. Misc. other fixes: ------------------ * Cleanup the relationship between ENABLE_PLUGINS and various other flags in CMakeLists.txt and configure.ac. * Re-arrange the TESTDIRS order in Makefile.am. * Add pseudo-breakpoint to nclog.[ch] for debugging. * Improve the documentation of the path manager code in ncpathmgr.h * Add better support for relative paths in dpathmgr.c * Default the mode args to NCfopen to include "b" (binary) for windows. * Add optional debugging output in various places. * Make sure that everything builds with plugins disabled. * Fix numerous (s)printf inconsistencies betweenb the format spec and the arguments.
2021-12-24 13:18:56 +08:00
#ifdef ENABLE_NCZARR_FILTERS
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* Make sure the entry is in filtered state */
if(!entry->isfiltered) {
NC_VAR_INFO_T* var = cache->var;
void* filtered = NULL; /* pointer to the filtered data */
size_t flen; /* length of filtered data */
/* Get the filter chain to apply */
NClist* filterchain = (NClist*)var->filters;
if(nclistlength(filterchain) > 0) {
/* Apply the filter chain to get the filtered data; will reclaim entry->data */
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
if((stat = NCZ_applyfilterchain(file,var,filterchain,entry->size,entry->data,&flen,&filtered,ENCODING))) goto done;
/* Fix up the cache entry */
/* Note that if filtered is different from entry->data, then entry->data will have been freed */
entry->data = filtered;
entry->size = flen;
entry->isfiltered = 1;
}
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
}
Support MSYS2/Mingw platform re: The current netcdf-c release has some problems with the mingw platform on windows. Mostly they are path issues. Changes to support mingw+msys2: ------------------------------- * Enable option of looking into the windows registry to find the mingw root path. In aid of proper path handling. * Add mingw+msys as a specific platform in configure.ac and move testing of the platform to the front so it is available early. * Handle mingw X libncpoco (dynamic loader) properly even though mingw does not yet support it. * Handle mingw X plugins properly even though mingw does not yet support it. * Alias pwd='pwd -W' to better handle paths in shell scripts. * Plus a number of other minor compile irritations. * Disallow the use of multiple nc_open's on the same file for windows (and mingw) because windows does not seem to handle these properly. Not sure why we did not catch this earlier. * Add mountpoint info to dpathmgr.c to help support mingw. * Cleanup dpathmgr conversions. Known problems: --------------- * I have not been able to get shared libraries to work, so plugins/filters must be disabled. * There is some kind of problem with libcurl that I have not solved, so all uses of libcurl (currently DAP+Byterange) must be disabled. Misc. other fixes: ------------------ * Cleanup the relationship between ENABLE_PLUGINS and various other flags in CMakeLists.txt and configure.ac. * Re-arrange the TESTDIRS order in Makefile.am. * Add pseudo-breakpoint to nclog.[ch] for debugging. * Improve the documentation of the path manager code in ncpathmgr.h * Add better support for relative paths in dpathmgr.c * Default the mode args to NCfopen to include "b" (binary) for windows. * Add optional debugging output in various places. * Make sure that everything builds with plugins disabled. * Fix numerous (s)printf inconsistencies betweenb the format spec and the arguments.
2021-12-24 13:18:56 +08:00
#endif
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
path = NCZ_chunkpath(entry->key);
Mitigate S3 test interference + Unlimited Dimensions in NCZarr This PR started as an attempt to add unlimited dimensions to NCZarr. It did that, but this exposed significant problems with test interference. So this PR is mostly about fixing -- well mitigating anyway -- test interference. The problem of test interference is now documented in the document docs/internal.md. The solutions implemented here are also describe in that document. The solution is somewhat fragile but multiple cleanup mechanisms are provided. Note that this feature requires that the AWS command line utility must be installed. ## Unlimited Dimensions. The existing NCZarr extensions to Zarr are modified to support unlimited dimensions. NCzarr extends the Zarr meta-data for the ".zgroup" object to include netcdf-4 model extensions. This information is stored in ".zgroup" as dictionary named "_nczarr_group". Inside "_nczarr_group", there is a key named "dims" that stores information about netcdf-4 named dimensions. The value of "dims" is a dictionary whose keys are the named dimensions. The value associated with each dimension name has one of two forms Form 1 is a special case of form 2, and is kept for backward compatibility. Whenever a new file is written, it uses format 1 if possible, otherwise format 2. * Form 1: An integer representing the size of the dimension, which is used for simple named dimensions. * Form 2: A dictionary with the following keys and values" - "size" with an integer value representing the (current) size of the dimension. - "unlimited" with a value of either "1" or "0" to indicate if this dimension is an unlimited dimension. For Unlimited dimensions, the size is initially zero, and as variables extend the length of that dimension, the size value for the dimension increases. That dimension size is shared by all arrays referencing that dimension, so if one array extends an unlimited dimension, it is implicitly extended for all other arrays that reference that dimension. This is the standard semantics for unlimited dimensions. Adding unlimited dimensions required a number of other changes to the NCZarr code-base. These included the following. * Did a partial refactor of the slice handling code in zwalk.c to clean it up. * Added a number of tests for unlimited dimensions derived from the same test in nc_test4. * Added several NCZarr specific unlimited tests; more are needed. * Add test of endianness. ## Misc. Other Changes * Modify libdispatch/ncs3sdk_aws.cpp to optionally support use of the AWS Transfer Utility mechanism. This is controlled by the ```#define TRANSFER```` command in that file. It defaults to being disabled. * Parameterize both the standard Unidata S3 bucket (S3TESTBUCKET) and the netcdf-c test data prefix (S3TESTSUBTREE). * Fixed an obscure memory leak in ncdump. * Removed some obsolete unit testing code and test cases. * Uncovered a bug in the netcdf-c handling of big-endian floats and doubles. Have not fixed yet. See tst_h5_endians.c. * Renamed some nczarr_tests testcases to avoid name conflicts with nc_test4. * Modify the semantics of zmap\#ncsmap_write to only allow total rewrite of objects. * Modify the semantics of zodom to properly handle stride > 1. * Add a truncate operation to the libnczarr zmap code.
2023-09-27 06:56:48 +08:00
stat = nczmap_write(map,path,entry->size,entry->data);
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
nullfree(path); path = NULL;
switch(stat) {
case NC_NOERR:
break;
case NC_EEMPTY:
default: goto done;
}
done:
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
nullfree(strchunk);
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
nullfree(path);
return ZUNTRACE(stat);
}
/**
* @internal Push data from memory to file.
*
* @param cache Pointer to parent cache
* @param key chunk key
* @param entry cache entry to read into
*
* @return ::NC_NOERR No error.
* @author Dennis Heimbigner
*/
static int
get_chunk(NCZChunkCache* cache, NCZCacheEntry* entry)
{
int stat = NC_NOERR;
NCZMAP* map = NULL;
NC_FILE_INFO_T* file = NULL;
NCZ_FILE_INFO_T* zfile = NULL;
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
NC_TYPE_INFO_T* xtype = NULL;
char** strchunk = NULL;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
size64_t size;
int empty = 0;
char* path = NULL;
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
int tid;
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
ZTRACE(5,"cache.var=%s entry.key=%s sep=%d",cache->var->hdr.name,entry->key,cache->dimension_separator);
LOG((3, "%s: file: %p", __func__, file));
file = (cache->var->container)->nc4_info;
zfile = file->format_file_info;
map = zfile->map;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
assert(map);
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
/* Collect some info */
xtype = cache->var->type_info;
tid = xtype->hdr.id;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* get size of the "raw" data on "disk" */
path = NCZ_chunkpath(entry->key);
stat = nczmap_len(map,path,&size);
nullfree(path); path = NULL;
switch(stat) {
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
case NC_NOERR: entry->size = size; break;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
case NC_EEMPTY: empty = 1; stat = NC_NOERR; break;
default: goto done;
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
}
/* make room in the cache */
if((stat = constraincache(cache,size))) goto done;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
if(!empty) {
/* Make sure we have a place to read it */
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
if((entry->data = (void*)calloc(1,entry->size)) == NULL)
{stat = NC_ENOMEM; goto done;}
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* Read the raw data */
path = NCZ_chunkpath(entry->key);
stat = nczmap_read(map,path,0,entry->size,(char*)entry->data);
nullfree(path); path = NULL;
switch (stat) {
case NC_NOERR: break;
case NC_EEMPTY: empty = 1; stat = NC_NOERR;break;
default: goto done;
}
entry->isfiltered = (int)FILTERED(cache); /* Is the data being read filtered? */
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
if(tid == NC_STRING)
entry->isfixedstring = 1; /* fill cache is in char[maxstrlen] format */
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
}
if(empty) {
/* fake the chunk */
Mitigate S3 test interference + Unlimited Dimensions in NCZarr This PR started as an attempt to add unlimited dimensions to NCZarr. It did that, but this exposed significant problems with test interference. So this PR is mostly about fixing -- well mitigating anyway -- test interference. The problem of test interference is now documented in the document docs/internal.md. The solutions implemented here are also describe in that document. The solution is somewhat fragile but multiple cleanup mechanisms are provided. Note that this feature requires that the AWS command line utility must be installed. ## Unlimited Dimensions. The existing NCZarr extensions to Zarr are modified to support unlimited dimensions. NCzarr extends the Zarr meta-data for the ".zgroup" object to include netcdf-4 model extensions. This information is stored in ".zgroup" as dictionary named "_nczarr_group". Inside "_nczarr_group", there is a key named "dims" that stores information about netcdf-4 named dimensions. The value of "dims" is a dictionary whose keys are the named dimensions. The value associated with each dimension name has one of two forms Form 1 is a special case of form 2, and is kept for backward compatibility. Whenever a new file is written, it uses format 1 if possible, otherwise format 2. * Form 1: An integer representing the size of the dimension, which is used for simple named dimensions. * Form 2: A dictionary with the following keys and values" - "size" with an integer value representing the (current) size of the dimension. - "unlimited" with a value of either "1" or "0" to indicate if this dimension is an unlimited dimension. For Unlimited dimensions, the size is initially zero, and as variables extend the length of that dimension, the size value for the dimension increases. That dimension size is shared by all arrays referencing that dimension, so if one array extends an unlimited dimension, it is implicitly extended for all other arrays that reference that dimension. This is the standard semantics for unlimited dimensions. Adding unlimited dimensions required a number of other changes to the NCZarr code-base. These included the following. * Did a partial refactor of the slice handling code in zwalk.c to clean it up. * Added a number of tests for unlimited dimensions derived from the same test in nc_test4. * Added several NCZarr specific unlimited tests; more are needed. * Add test of endianness. ## Misc. Other Changes * Modify libdispatch/ncs3sdk_aws.cpp to optionally support use of the AWS Transfer Utility mechanism. This is controlled by the ```#define TRANSFER```` command in that file. It defaults to being disabled. * Parameterize both the standard Unidata S3 bucket (S3TESTBUCKET) and the netcdf-c test data prefix (S3TESTSUBTREE). * Fixed an obscure memory leak in ncdump. * Removed some obsolete unit testing code and test cases. * Uncovered a bug in the netcdf-c handling of big-endian floats and doubles. Have not fixed yet. See tst_h5_endians.c. * Renamed some nczarr_tests testcases to avoid name conflicts with nc_test4. * Modify the semantics of zmap\#ncsmap_write to only allow total rewrite of objects. * Modify the semantics of zodom to properly handle stride > 1. * Add a truncate operation to the libnczarr zmap code.
2023-09-27 06:56:48 +08:00
setmodified(entry,(file->no_write?0:1));
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
entry->size = cache->chunksize;
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
entry->data = NULL;
entry->isfixedstring = 0;
entry->isfiltered = 0;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* apply fill value */
if(cache->fillchunk == NULL)
{if((stat = NCZ_ensure_fill_chunk(cache))) goto done;}
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
if((entry->data = calloc(1,entry->size))==NULL) {stat = NC_ENOMEM; goto done;}
Mitigate S3 test interference + Unlimited Dimensions in NCZarr This PR started as an attempt to add unlimited dimensions to NCZarr. It did that, but this exposed significant problems with test interference. So this PR is mostly about fixing -- well mitigating anyway -- test interference. The problem of test interference is now documented in the document docs/internal.md. The solutions implemented here are also describe in that document. The solution is somewhat fragile but multiple cleanup mechanisms are provided. Note that this feature requires that the AWS command line utility must be installed. ## Unlimited Dimensions. The existing NCZarr extensions to Zarr are modified to support unlimited dimensions. NCzarr extends the Zarr meta-data for the ".zgroup" object to include netcdf-4 model extensions. This information is stored in ".zgroup" as dictionary named "_nczarr_group". Inside "_nczarr_group", there is a key named "dims" that stores information about netcdf-4 named dimensions. The value of "dims" is a dictionary whose keys are the named dimensions. The value associated with each dimension name has one of two forms Form 1 is a special case of form 2, and is kept for backward compatibility. Whenever a new file is written, it uses format 1 if possible, otherwise format 2. * Form 1: An integer representing the size of the dimension, which is used for simple named dimensions. * Form 2: A dictionary with the following keys and values" - "size" with an integer value representing the (current) size of the dimension. - "unlimited" with a value of either "1" or "0" to indicate if this dimension is an unlimited dimension. For Unlimited dimensions, the size is initially zero, and as variables extend the length of that dimension, the size value for the dimension increases. That dimension size is shared by all arrays referencing that dimension, so if one array extends an unlimited dimension, it is implicitly extended for all other arrays that reference that dimension. This is the standard semantics for unlimited dimensions. Adding unlimited dimensions required a number of other changes to the NCZarr code-base. These included the following. * Did a partial refactor of the slice handling code in zwalk.c to clean it up. * Added a number of tests for unlimited dimensions derived from the same test in nc_test4. * Added several NCZarr specific unlimited tests; more are needed. * Add test of endianness. ## Misc. Other Changes * Modify libdispatch/ncs3sdk_aws.cpp to optionally support use of the AWS Transfer Utility mechanism. This is controlled by the ```#define TRANSFER```` command in that file. It defaults to being disabled. * Parameterize both the standard Unidata S3 bucket (S3TESTBUCKET) and the netcdf-c test data prefix (S3TESTSUBTREE). * Fixed an obscure memory leak in ncdump. * Removed some obsolete unit testing code and test cases. * Uncovered a bug in the netcdf-c handling of big-endian floats and doubles. Have not fixed yet. See tst_h5_endians.c. * Renamed some nczarr_tests testcases to avoid name conflicts with nc_test4. * Modify the semantics of zmap\#ncsmap_write to only allow total rewrite of objects. * Modify the semantics of zodom to properly handle stride > 1. * Add a truncate operation to the libnczarr zmap code.
2023-09-27 06:56:48 +08:00
if((stat = NCZ_copy_data(file,cache->var,cache->fillchunk,cache->chunkcount,ZREADING,entry->data))) goto done;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
stat = NC_NOERR;
}
Support MSYS2/Mingw platform re: The current netcdf-c release has some problems with the mingw platform on windows. Mostly they are path issues. Changes to support mingw+msys2: ------------------------------- * Enable option of looking into the windows registry to find the mingw root path. In aid of proper path handling. * Add mingw+msys as a specific platform in configure.ac and move testing of the platform to the front so it is available early. * Handle mingw X libncpoco (dynamic loader) properly even though mingw does not yet support it. * Handle mingw X plugins properly even though mingw does not yet support it. * Alias pwd='pwd -W' to better handle paths in shell scripts. * Plus a number of other minor compile irritations. * Disallow the use of multiple nc_open's on the same file for windows (and mingw) because windows does not seem to handle these properly. Not sure why we did not catch this earlier. * Add mountpoint info to dpathmgr.c to help support mingw. * Cleanup dpathmgr conversions. Known problems: --------------- * I have not been able to get shared libraries to work, so plugins/filters must be disabled. * There is some kind of problem with libcurl that I have not solved, so all uses of libcurl (currently DAP+Byterange) must be disabled. Misc. other fixes: ------------------ * Cleanup the relationship between ENABLE_PLUGINS and various other flags in CMakeLists.txt and configure.ac. * Re-arrange the TESTDIRS order in Makefile.am. * Add pseudo-breakpoint to nclog.[ch] for debugging. * Improve the documentation of the path manager code in ncpathmgr.h * Add better support for relative paths in dpathmgr.c * Default the mode args to NCfopen to include "b" (binary) for windows. * Add optional debugging output in various places. * Make sure that everything builds with plugins disabled. * Fix numerous (s)printf inconsistencies betweenb the format spec and the arguments.
2021-12-24 13:18:56 +08:00
#ifdef ENABLE_NCZARR_FILTERS
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* Make sure the entry is in unfiltered state */
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
if(!empty && entry->isfiltered) {
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
NC_VAR_INFO_T* var = cache->var;
void* unfiltered = NULL; /* pointer to the unfiltered data */
void* filtered = NULL; /* pointer to the filtered data */
size_t unflen; /* length of unfiltered data */
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
assert(tid != NC_STRING || entry->isfixedstring);
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
/* Get the filter chain to apply */
NClist* filterchain = (NClist*)var->filters;
if(nclistlength(filterchain) == 0) {stat = NC_EFILTER; goto done;}
/* Apply the filter chain to get the unfiltered data */
filtered = entry->data;
entry->data = NULL;
if((stat = NCZ_applyfilterchain(file,var,filterchain,entry->size,filtered,&unflen,&unfiltered,!ENCODING))) goto done;
/* Fix up the cache entry */
entry->data = unfiltered;
entry->size = unflen;
entry->isfiltered = 0;
}
Support MSYS2/Mingw platform re: The current netcdf-c release has some problems with the mingw platform on windows. Mostly they are path issues. Changes to support mingw+msys2: ------------------------------- * Enable option of looking into the windows registry to find the mingw root path. In aid of proper path handling. * Add mingw+msys as a specific platform in configure.ac and move testing of the platform to the front so it is available early. * Handle mingw X libncpoco (dynamic loader) properly even though mingw does not yet support it. * Handle mingw X plugins properly even though mingw does not yet support it. * Alias pwd='pwd -W' to better handle paths in shell scripts. * Plus a number of other minor compile irritations. * Disallow the use of multiple nc_open's on the same file for windows (and mingw) because windows does not seem to handle these properly. Not sure why we did not catch this earlier. * Add mountpoint info to dpathmgr.c to help support mingw. * Cleanup dpathmgr conversions. Known problems: --------------- * I have not been able to get shared libraries to work, so plugins/filters must be disabled. * There is some kind of problem with libcurl that I have not solved, so all uses of libcurl (currently DAP+Byterange) must be disabled. Misc. other fixes: ------------------ * Cleanup the relationship between ENABLE_PLUGINS and various other flags in CMakeLists.txt and configure.ac. * Re-arrange the TESTDIRS order in Makefile.am. * Add pseudo-breakpoint to nclog.[ch] for debugging. * Improve the documentation of the path manager code in ncpathmgr.h * Add better support for relative paths in dpathmgr.c * Default the mode args to NCfopen to include "b" (binary) for windows. * Add optional debugging output in various places. * Make sure that everything builds with plugins disabled. * Fix numerous (s)printf inconsistencies betweenb the format spec and the arguments.
2021-12-24 13:18:56 +08:00
#endif
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
if(tid == NC_STRING && entry->isfixedstring) {
/* Convert from char[strlen] to char* format */
int maxstrlen = NCZ_get_maxstrlen((NC_OBJ*)cache->var);
assert(maxstrlen > 0);
/* copy char[] to char* format */
if((strchunk = (char**)malloc(sizeof(char*)*cache->chunkcount))==NULL)
{stat = NC_ENOMEM; goto done;}
if((stat = NCZ_fixed2char(entry->data,strchunk,cache->chunkcount,maxstrlen))) goto done;
/* Reclaim the old chunk */
nullfree(entry->data);
entry->data = NULL;
entry->data = strchunk; strchunk = NULL;
entry->size = cache->chunkcount * sizeof(char*);
entry->isfixedstring = 0;
}
/* track new chunk */
cache->used += entry->size;
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
done:
Add support for Zarr string type to NCZarr * re: https://github.com/Unidata/netcdf-c/pull/2278 * re: https://github.com/Unidata/netcdf-c/issues/2485 * re: https://github.com/Unidata/netcdf-c/issues/2474 This PR subsumes PR https://github.com/Unidata/netcdf-c/pull/2278. Actually is a bit an omnibus covering several issues. ## PR https://github.com/Unidata/netcdf-c/pull/2278 Add support for the Zarr string type. Zarr strings are restricted currently to be of fixed size. The primary issue to be addressed is to provide a way for user to specify the size of the fixed length strings. This is handled by providing the following new attributes special: 1. **_nczarr_default_maxstrlen** &mdash; This is an attribute of the root group. It specifies the default maximum string length for string types. If not specified, then it has the value of 64 characters. 2. **_nczarr_maxstrlen** &mdash; This is a per-variable attribute. It specifies the maximum string length for the string type associated with the variable. If not specified, then it is assigned the value of **_nczarr_default_maxstrlen**. This PR also requires some hacking to handle the existing netcdf-c NC_CHAR type, which does not exist in zarr. The goal was to choose numpy types for both the netcdf-c NC_STRING type and the netcdf-c NC_CHAR type such that if a pure zarr implementation read them, it would still work and an NC_CHAR type would be handled by zarr as a string of length 1. For writing variables and NCZarr attributes, the type mapping is as follows: * "|S1" for NC_CHAR. * ">S1" for NC_STRING && MAXSTRLEN==1 * ">Sn" for NC_STRING && MAXSTRLEN==n Note that it is a bit of a hack to use endianness, but it should be ok since for string/char, the endianness has no meaning. For reading attributes with pure zarr (i.e. with no nczarr atribute types defined), they will always be interpreted as of type NC_CHAR. ## Issue: https://github.com/Unidata/netcdf-c/issues/2474 This PR partly fixes this issue because it provided more comprehensive support for Zarr attributes that are JSON valued expressions. This PR still does not address the problem in that issue where the _ARRAY_DIMENSION attribute is incorrectly set. Than can only be fixed by the creator of the datasets. ## Issue: https://github.com/Unidata/netcdf-c/issues/2485 This PR also fixes the scalar failure shown in this issue. It generally cleans up scalar handling. It also adds a note to the documentation describing that NCZarr supports scalars while Zarr does not and also how scalar interoperability is achieved. ## Misc. Other Changes 1. Convert the nczarr special attributes and keys to be all lower case. So "_NCZARR_ATTR" now used "_nczarr_attr. Support back compatibility for the upper case names. 2. Cleanup my too-clever-by-half handling of scalars in libnczarr.
2022-08-28 10:21:13 +08:00
nullfree(strchunk);
Add filter support to NCZarr Filter support has three goals: 1. Use the existing HDF5 filter implementations, 2. Allow filter metadata to be stored in the NumCodecs metadata format used by Zarr, 3. Allow filters to be used even when HDF5 is disabled Detailed usage directions are define in docs/filters.md. For now, the existing filter API is left in place. So filters are defined using ''nc_def_var_filter'' using the HDF5 style where the id and parameters are unsigned integers. This is a big change since filters affect many parts of the code. In the following, the terms "compressor" and "filter" and "codec" are generally used synonomously. ### Filter-Related Changes: * In order to support dynamic loading of shared filter libraries, a new library was added in the libncpoco directory; it helps to isolate dynamic loading across multiple platforms. * Provide a json parsing library for use by plugins; this is created by merging libdispatch/ncjson.c with include/ncjson.h. * Add a new _Codecs attribute to allow clients to see what codecs are being used; let ncdump -s print it out. * Provide special headers to help support compilation of HDF5 filters when HDF5 is not enabled: netcdf_filter_hdf5_build.h and netcdf_filter_build.h. * Add a number of new test to test the new nczarr filters. * Let ncgen parse _Codecs attribute, although it is ignored. ### Plugin directory changes: * Add support for the Blosc compressor; this is essential because it is the most common compressor used in Zarr datasets. This also necessitated adding a CMake FindBlosc.cmake file * Add NCZarr support for the big-four filters provided by HDF5: shuffle, fletcher32, deflate (zlib), and szip * Add a Codec defaulter (see docs/filters.md) for the big four filters. * Make plugins work with windows by properly adding __declspec declaration. ### Misc. Non-Filter Changes * Replace most uses of USE_NETCDF4 (deprecated) with USE_HDF5. * Improve support for caching * More fixes for path conversion code * Fix misc. memory leaks * Add new utility -- ncdump/ncpathcvt -- that does more or less the same thing as cygpath. * Add a number of new test to test the non-filter fixes. * Update the parsers * Convert most instances of '#ifdef _MSC_VER' to '#ifdef _WIN32'
2021-09-03 07:04:26 +08:00
nullfree(path);
return ZUNTRACE(stat);
}
int
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
NCZ_buildchunkpath(NCZChunkCache* cache, const size64_t* chunkindices, struct ChunkKey* key)
{
int stat = NC_NOERR;
char* chunkname = NULL;
char* varkey = NULL;
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
assert(key != NULL);
/* Get the chunk object name */
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
if((stat = NCZ_buildchunkkey(cache->ndims, chunkindices, cache->dimension_separator, &chunkname))) goto done;
/* Get the var object key */
if((stat = NCZ_varkey(cache->var,&varkey))) goto done;
Upgrade the nczarr code to match Zarr V2 Re: https://github.com/zarr-developers/zarr-python/pull/716 The Zarr version 2 spec has been extended to include the ability to choose the dimension separator in chunk name keys. The legal separators has been extended from {'.'} to {'.' '/'}. So now it is possible to use a key like "0/1/2/0" for chunk names. This PR implements this for NCZarr. The V2 spec now says that this separator can be set on a per-variable basis. For now, I have chosen to allow this be set only globally by adding a key named "ZARR.DIMENSION_SEPARATOR=<char>" in the .daprc/.dodsrc/ncrc file. Currently, the only legal separator characters are '.' (the default) and '/'. On writing, this key will only be written if its value is different than the default. This change caused problems because supporting a separator of '/' is difficult to parse when keys/paths use '/' as the path separator. A test case was added for this. Additionally, make nczarr be enabled default by default. This required some additional changes so that if zip and/or AWS S3 sdk are unavailable, then they are disabled for NCZarr. In addition the following unrelated changes were made. 1. Tested that pure-zarr mode could read an nczarr formatted store. 1. The .rc file handling now merges all known .rc files (.ncrc,.daprc, and .dodsrc) in that order and using those in HOME first, then in current directory. For duplicate entries, the later ones override the earlier ones. This change is to remove some of the conflicts inherent in the current .rc file load process. A set of test cases was also added. 1. Re-order tests in configure.ac and CMakeLists.txt so that if libcurl is not found then the other options that depend upon it properly are disabled. 1. I decided that xarray support should be enabled by default for pure zarr. In order to allow disabling, I added a new mode flag "noxarray". 1. Certain test in nczarr_test depend on use of .dodsrc. In order for these to work when testing in parallel, some inter-test dependencies needed to be added. 1. Improved authorization testing to use changes in thredds.ucar.edu
2021-04-25 09:48:15 +08:00
key->varkey = varkey; varkey = NULL;
key->chunkkey = chunkname; chunkname = NULL;
done:
nullfree(chunkname);
nullfree(varkey);
return THROW(stat);
}
void
NCZ_dumpxcacheentry(NCZChunkCache* cache, NCZCacheEntry* e, NCbytes* buf)
{
char s[8192];
char idx[64];
int i;
ncbytescat(buf,"{");
snprintf(s,sizeof(s),"modified=%u isfiltered=%u indices=",
(unsigned)e->modified,
(unsigned)e->isfiltered
);
ncbytescat(buf,s);
for(i=0;i<cache->ndims;i++) {
snprintf(idx,sizeof(idx),"%s%llu",(i==0?"":"."),e->indices[i]);
ncbytescat(buf,idx);
}
snprintf(s,sizeof(s),"size=%llu data=%p",
e->size,
e->data
);
ncbytescat(buf,s);
ncbytescat(buf,"}");
}
void
NCZ_printxcache(NCZChunkCache* cache)
{
static char xs[20000];
NCbytes* buf = ncbytesnew();
char s[8192];
size_t i;
ncbytescat(buf,"NCZChunkCache:\n");
snprintf(s,sizeof(s),"\tvar=%s\n\tndims=%u\n\tchunksize=%u\n\tchunkcount=%u\n\tfillchunk=%p\n",
cache->var->hdr.name,
(unsigned)cache->ndims,
(unsigned)cache->chunksize,
(unsigned)cache->chunkcount,
cache->fillchunk
);
ncbytescat(buf,s);
snprintf(s,sizeof(s),"\tmaxentries=%u\n\tmaxsize=%u\n\tused=%u\n\tdimsep='%c'\n",
(unsigned)cache->params.nelems,
(unsigned)cache->params.size,
(unsigned)cache->used,
cache->dimension_separator
);
ncbytescat(buf,s);
snprintf(s,sizeof(s),"\tmru: (%u)\n",(unsigned)nclistlength(cache->mru));
ncbytescat(buf,s);
if(nclistlength(cache->mru)==0)
ncbytescat(buf,"\t\t<empty>\n");
for(i=0;i<nclistlength(cache->mru);i++) {
NCZCacheEntry* e = (NCZCacheEntry*)nclistget(cache->mru,i);
snprintf(s,sizeof(s),"\t\t[%zu] ", i);
ncbytescat(buf,s);
if(e == NULL)
ncbytescat(buf,"<null>");
else
NCZ_dumpxcacheentry(cache, e, buf);
ncbytescat(buf,"\n");
}
xs[0] = '\0';
strlcat(xs,ncbytescontents(buf),sizeof(xs));
ncbytesfree(buf);
fprintf(stderr,"%s\n",xs);
}