2020-02-17 03:59:33 +08:00
|
|
|
/* Copyright 2003-2018, University Corporation for Atmospheric
|
|
|
|
* Research. See the COPYRIGHT file for copying and redistribution
|
|
|
|
* conditions.
|
|
|
|
*/
|
|
|
|
/**
|
|
|
|
* @file @internal Internal netcdf-4 functions for filters
|
|
|
|
*
|
|
|
|
* This file contains functions internal to the netcdf4 library. None of
|
|
|
|
* the functions in this file are exposed in the exetnal API. These
|
|
|
|
* functions all relate to the manipulation of netcdf-4 filters
|
|
|
|
*
|
|
|
|
* @author Dennis Heimbigner
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include "config.h"
|
|
|
|
#include <stdlib.h>
|
|
|
|
#include "hdf5internal.h"
|
|
|
|
#include "hdf5debug.h"
|
2020-09-28 02:43:46 +08:00
|
|
|
#include "netcdf.h"
|
|
|
|
#include "netcdf_filter.h"
|
2020-02-17 03:59:33 +08:00
|
|
|
|
2021-09-03 07:04:26 +08:00
|
|
|
#ifdef ENABLE_BLOSC
|
|
|
|
#include <blosc.h>
|
|
|
|
#endif
|
|
|
|
|
2020-09-28 02:43:46 +08:00
|
|
|
#undef TFILTERS
|
2020-02-17 03:59:33 +08:00
|
|
|
|
2021-09-03 07:04:26 +08:00
|
|
|
/* Forward */
|
2020-09-28 02:43:46 +08:00
|
|
|
static int NC4_hdf5_filter_free(struct NC_HDF5_Filter* spec);
|
2020-02-17 03:59:33 +08:00
|
|
|
|
2020-05-11 23:42:31 +08:00
|
|
|
/**************************************************/
|
|
|
|
/* Filter registration support */
|
2020-02-17 03:59:33 +08:00
|
|
|
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
#ifdef ENABLE_CLIENTSIDE_FILTERS
|
2020-05-11 23:42:31 +08:00
|
|
|
|
2020-09-28 02:43:46 +08:00
|
|
|
/* Mnemonic */
|
|
|
|
#define FILTERACTIVE 1
|
|
|
|
|
2020-05-11 23:42:31 +08:00
|
|
|
/* WARNING: GLOBAL VARIABLE */
|
2020-02-17 03:59:33 +08:00
|
|
|
/* Define list of registered filters */
|
|
|
|
static NClist* NC4_registeredfilters = NULL; /** List<NC_FILTER_CLIENT_HDF5*> */
|
|
|
|
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
/**************************************************/
|
|
|
|
/* Filter registration support */
|
|
|
|
|
2020-02-17 03:59:33 +08:00
|
|
|
static int
|
2020-05-11 23:42:31 +08:00
|
|
|
clientfilterlookup(unsigned int id)
|
2020-02-17 03:59:33 +08:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
if(NC4_registeredfilters == NULL)
|
|
|
|
NC4_registeredfilters = nclistnew();
|
|
|
|
for(i=0;i<nclistlength(NC4_registeredfilters);i++) {
|
|
|
|
NC_FILTER_CLIENT_HDF5* x = nclistget(NC4_registeredfilters,i);
|
2020-05-11 23:42:31 +08:00
|
|
|
if(x != NULL && x->id == id) {
|
|
|
|
return i; /* return position */
|
|
|
|
}
|
2020-02-17 03:59:33 +08:00
|
|
|
}
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
reclaiminfo(NC_FILTER_CLIENT_HDF5* info)
|
|
|
|
{
|
|
|
|
nullfree(info);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
filterremove(int pos)
|
|
|
|
{
|
|
|
|
NC_FILTER_CLIENT_HDF5* info = NULL;
|
|
|
|
if(NC4_registeredfilters == NULL)
|
|
|
|
return THROW(NC_EINVAL);
|
|
|
|
if(pos < 0 || pos >= nclistlength(NC4_registeredfilters))
|
|
|
|
return THROW(NC_EINVAL);
|
|
|
|
info = nclistget(NC4_registeredfilters,pos);
|
|
|
|
reclaiminfo(info);
|
|
|
|
nclistremove(NC4_registeredfilters,pos);
|
|
|
|
return NC_NOERR;
|
|
|
|
}
|
|
|
|
|
|
|
|
static NC_FILTER_CLIENT_HDF5*
|
|
|
|
dupfilterinfo(NC_FILTER_CLIENT_HDF5* info)
|
|
|
|
{
|
|
|
|
NC_FILTER_CLIENT_HDF5* dup = NULL;
|
|
|
|
if(info == NULL) goto fail;
|
|
|
|
if((dup = calloc(1,sizeof(NC_FILTER_CLIENT_HDF5))) == NULL) goto fail;
|
|
|
|
*dup = *info;
|
|
|
|
return dup;
|
|
|
|
fail:
|
|
|
|
reclaiminfo(dup);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
nc4_global_filter_action(int op, unsigned int id, NC_FILTER_OBJ_HDF5* infop)
|
|
|
|
{
|
|
|
|
int stat = NC_NOERR;
|
|
|
|
H5Z_class2_t* h5filterinfo = NULL;
|
|
|
|
herr_t herr;
|
|
|
|
int pos = -1;
|
|
|
|
NC_FILTER_CLIENT_HDF5* dup = NULL;
|
|
|
|
NC_FILTER_CLIENT_HDF5* elem = NULL;
|
|
|
|
NC_FILTER_CLIENT_HDF5 ncf;
|
|
|
|
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
NC_UNUSED(format);
|
|
|
|
|
2020-02-17 03:59:33 +08:00
|
|
|
switch (op) {
|
|
|
|
case NCFILTER_CLIENT_REG: /* Ignore id argument */
|
|
|
|
if(infop == NULL) {stat = NC_EINVAL; goto done;}
|
|
|
|
assert(NC_FILTER_FORMAT_HDF5 == infop->hdr.format);
|
|
|
|
assert(NC_FILTER_SORT_CLIENT == infop->sort);
|
|
|
|
elem = (NC_FILTER_CLIENT_HDF5*)&infop->u.client;
|
|
|
|
h5filterinfo = elem->info;
|
|
|
|
/* Another sanity check */
|
|
|
|
if(id != h5filterinfo->id)
|
|
|
|
{stat = NC_EINVAL; goto done;}
|
|
|
|
/* See if this filter is already defined */
|
2020-05-11 23:42:31 +08:00
|
|
|
if((pos = clientfilterlookup(id)) >= 0)
|
2020-02-17 03:59:33 +08:00
|
|
|
{stat = NC_ENAMEINUSE; goto done;} /* Already defined */
|
|
|
|
if((herr = H5Zregister(h5filterinfo)) < 0)
|
|
|
|
{stat = NC_EFILTER; goto done;}
|
|
|
|
/* Save a copy of the passed in info */
|
|
|
|
ncf.id = id;
|
|
|
|
ncf.info = elem->info;
|
|
|
|
if((dup=dupfilterinfo(&ncf)) == NULL)
|
|
|
|
{stat = NC_ENOMEM; goto done;}
|
|
|
|
nclistpush(NC4_registeredfilters,dup);
|
|
|
|
break;
|
|
|
|
case NCFILTER_CLIENT_UNREG:
|
|
|
|
if(id <= 0)
|
|
|
|
{stat = NC_ENOTNC4; goto done;}
|
|
|
|
/* See if this filter is already defined */
|
2020-05-11 23:42:31 +08:00
|
|
|
if((pos = clientfilterlookup(id)) < 0)
|
2020-02-17 03:59:33 +08:00
|
|
|
{stat = NC_ENOFILTER; goto done;} /* Not defined */
|
|
|
|
if((herr = H5Zunregister(id)) < 0)
|
|
|
|
{stat = NC_EFILTER; goto done;}
|
|
|
|
if((stat=filterremove(pos))) goto done;
|
|
|
|
break;
|
|
|
|
case NCFILTER_CLIENT_INQ:
|
|
|
|
if(infop == NULL) goto done;
|
|
|
|
/* Look up the id in our local table */
|
2020-05-11 23:42:31 +08:00
|
|
|
if((pos = clientfilterlookup(id)) < 0)
|
2020-02-17 03:59:33 +08:00
|
|
|
{stat = NC_ENOFILTER; goto done;} /* Not defined */
|
|
|
|
elem = (NC_FILTER_CLIENT_HDF5*)nclistget(NC4_registeredfilters,pos);
|
|
|
|
if(elem == NULL) {stat = NC_EINTERNAL; goto done;}
|
|
|
|
if(infop != NULL) {
|
|
|
|
infop->u.client = *elem;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
{stat = NC_EINTERNAL; goto done;}
|
|
|
|
}
|
|
|
|
done:
|
|
|
|
return THROW(stat);
|
|
|
|
}
|
|
|
|
|
2020-05-11 23:42:31 +08:00
|
|
|
#endif /*ENABLE_CLIENTSIDE_FILTERS*/
|
|
|
|
|
|
|
|
/**************************************************/
|
2020-09-28 02:43:46 +08:00
|
|
|
/**************************************************/
|
|
|
|
/**
|
|
|
|
* @file
|
|
|
|
* @internal
|
|
|
|
* Internal netcdf hdf5 filter functions.
|
|
|
|
*
|
|
|
|
* This file contains functions internal to the libhdf5 library.
|
|
|
|
* None of the functions in this file are exposed in the exetnal API. These
|
|
|
|
* functions all relate to the manipulation of netcdf-4's var->filters list.
|
|
|
|
*
|
|
|
|
* @author Dennis Heimbigner
|
|
|
|
*/
|
|
|
|
#ifdef TFILTERS
|
|
|
|
static void
|
|
|
|
printfilter1(struct NC_HDF5_Filter* nfs)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
if(nfs == NULL) {
|
|
|
|
fprintf(stderr,"{null}");
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
fprintf(stderr,"{%u,(%u)",nfs->filterid,(int)nfs->nparams);
|
|
|
|
for(i=0;i<nfs->nparams;i++) {
|
|
|
|
fprintf(stderr," %s",nfs->params[i]);
|
|
|
|
}
|
|
|
|
fprintf(stderr,"}");
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
printfilter(struct NC_HDF5_Filter* nfs, const char* tag, int line)
|
|
|
|
{
|
|
|
|
fprintf(stderr,"%s: line=%d: ",tag,line);
|
|
|
|
printfilter1(nfs);
|
|
|
|
fprintf(stderr,"\n");
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
printfilterlist(NC_VAR_INFO_T* var, const char* tag, int line)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
const char* name;
|
|
|
|
if(var == NULL) name = "null";
|
|
|
|
else if(var->hdr.name == NULL) name = "?";
|
|
|
|
else name = var->hdr.name;
|
|
|
|
fprintf(stderr,"%s: line=%d: var=%s filters=",tag,line,name);
|
|
|
|
if(var != NULL) {
|
|
|
|
NClist* filters = (NClist*)var->filters;
|
|
|
|
for(i=0;i<nclistlength(filters);i++) {
|
|
|
|
struct NC_HDF5_Filter* nfs = (struct NC_HDF5_Filter*)nclistget(filters,i);
|
|
|
|
fprintf(stderr,"[%d]",i);
|
|
|
|
printfilter1(nfs);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
fprintf(stderr,"\n");
|
|
|
|
}
|
|
|
|
|
|
|
|
#define PRINTFILTER(nfs, tag) printfilter(nfs,tag,__LINE__)
|
|
|
|
#define PRINTFILTERLIST(var,tag) printfilterlist(var,tag,__LINE__)
|
|
|
|
#else
|
|
|
|
#define PRINTFILTER(nfs, tag)
|
|
|
|
#define PRINTFILTERLIST(var,tag)
|
|
|
|
#endif
|
2020-05-11 23:42:31 +08:00
|
|
|
|
|
|
|
int
|
2020-09-28 02:43:46 +08:00
|
|
|
NC4_hdf5_filter_freelist(NC_VAR_INFO_T* var)
|
|
|
|
{
|
|
|
|
int i, stat=NC_NOERR;
|
|
|
|
NClist* filters = (NClist*)var->filters;
|
|
|
|
|
|
|
|
if(filters == NULL) goto done;
|
|
|
|
PRINTFILTERLIST(var,"free: before");
|
|
|
|
/* Free the filter list backward */
|
|
|
|
for(i=nclistlength(filters)-1;i>=0;i--) {
|
|
|
|
struct NC_HDF5_Filter* spec = (struct NC_HDF5_Filter*)nclistremove(filters,i);
|
|
|
|
if(spec->nparams > 0) nullfree(spec->params);
|
|
|
|
nullfree(spec);
|
|
|
|
}
|
|
|
|
PRINTFILTERLIST(var,"free: after");
|
|
|
|
nclistfree(filters);
|
|
|
|
var->filters = NULL;
|
|
|
|
done:
|
|
|
|
return stat;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
NC4_hdf5_filter_free(struct NC_HDF5_Filter* spec)
|
|
|
|
{
|
|
|
|
if(spec == NULL) goto done;
|
|
|
|
PRINTFILTER(spec,"free");
|
|
|
|
if(spec->nparams > 0) nullfree(spec->params)
|
|
|
|
free(spec);
|
|
|
|
done:
|
|
|
|
return NC_NOERR;
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
2021-05-18 09:49:58 +08:00
|
|
|
NC4_hdf5_addfilter(NC_VAR_INFO_T* var, unsigned int id, size_t nparams, const unsigned int* params, int flags)
|
2020-05-11 23:42:31 +08:00
|
|
|
{
|
|
|
|
int stat = NC_NOERR;
|
2020-09-28 02:43:46 +08:00
|
|
|
struct NC_HDF5_Filter* fi = NULL;
|
|
|
|
int olddef = 0; /* 1=>already defined */
|
|
|
|
NClist* flist = (NClist*)var->filters;
|
|
|
|
|
|
|
|
if(nparams > 0 && params == NULL)
|
|
|
|
{stat = NC_EINVAL; goto done;}
|
|
|
|
|
|
|
|
if((stat=NC4_hdf5_filter_lookup(var,id,&fi))==NC_NOERR) {
|
|
|
|
assert(fi != NULL);
|
|
|
|
/* already exists */
|
|
|
|
olddef = 1;
|
|
|
|
} else {
|
|
|
|
stat = NC_NOERR;
|
|
|
|
if((fi = calloc(1,sizeof(struct NC_HDF5_Filter))) == NULL)
|
|
|
|
{stat = NC_ENOMEM; goto done;}
|
|
|
|
fi->filterid = id;
|
|
|
|
olddef = 0;
|
|
|
|
}
|
|
|
|
fi->nparams = nparams;
|
|
|
|
if(fi->params != NULL) {
|
|
|
|
nullfree(fi->params);
|
|
|
|
fi->params = NULL;
|
|
|
|
}
|
|
|
|
assert(fi->params == NULL);
|
|
|
|
if(fi->nparams > 0) {
|
|
|
|
if((fi->params = (unsigned int*)malloc(sizeof(unsigned int)*fi->nparams)) == NULL)
|
|
|
|
{stat = NC_ENOMEM; goto done;}
|
|
|
|
memcpy(fi->params,params,sizeof(unsigned int)*fi->nparams);
|
2020-05-11 23:42:31 +08:00
|
|
|
}
|
2021-05-18 09:49:58 +08:00
|
|
|
fi->flags = flags;
|
2020-09-28 02:43:46 +08:00
|
|
|
if(!olddef) {
|
|
|
|
nclistpush(flist,fi);
|
|
|
|
PRINTFILTERLIST(var,"add");
|
|
|
|
}
|
|
|
|
fi = NULL; /* either way,its in the var->filters list */
|
|
|
|
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
done:
|
2020-09-28 02:43:46 +08:00
|
|
|
if(fi) NC4_hdf5_filter_free(fi);
|
2020-05-11 23:42:31 +08:00
|
|
|
return THROW(stat);
|
|
|
|
}
|
|
|
|
|
2020-02-17 03:59:33 +08:00
|
|
|
int
|
2020-09-28 02:43:46 +08:00
|
|
|
NC4_hdf5_filter_remove(NC_VAR_INFO_T* var, unsigned int id)
|
|
|
|
{
|
|
|
|
int k;
|
|
|
|
NClist* flist = (NClist*)var->filters;
|
|
|
|
|
|
|
|
/* Walk backwards */
|
|
|
|
for(k=nclistlength(flist)-1;k>=0;k--) {
|
|
|
|
struct NC_HDF5_Filter* f = (struct NC_HDF5_Filter*)nclistget(flist,k);
|
|
|
|
if(f->filterid == id) {
|
|
|
|
/* Remove from variable */
|
|
|
|
nclistremove(flist,k);
|
|
|
|
#ifdef TFILTERS
|
|
|
|
PRINTFILTERLIST(var,"remove");
|
|
|
|
fprintf(stderr,"\tid=%s\n",id);
|
|
|
|
#endif
|
|
|
|
/* Reclaim */
|
|
|
|
NC4_hdf5_filter_free(f);
|
|
|
|
return NC_NOERR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return NC_ENOFILTER;
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
NC4_hdf5_filter_lookup(NC_VAR_INFO_T* var, unsigned int id, struct NC_HDF5_Filter** specp)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
NClist* flist = (NClist*)var->filters;
|
|
|
|
|
|
|
|
if(flist == NULL) {
|
|
|
|
if((flist = nclistnew())==NULL)
|
|
|
|
return NC_ENOMEM;
|
|
|
|
var->filters = (void*)flist;
|
|
|
|
}
|
|
|
|
for(i=0;i<nclistlength(flist);i++) {
|
|
|
|
struct NC_HDF5_Filter* spec = (struct NC_HDF5_Filter*)nclistget(flist,i);
|
|
|
|
if(id == spec->filterid) {
|
|
|
|
if(specp) *specp = spec;
|
|
|
|
return NC_NOERR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return NC_ENOFILTER;
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
NC4_hdf5_def_var_filter(int ncid, int varid, unsigned int id, size_t nparams,
|
|
|
|
const unsigned int* params)
|
2020-02-17 03:59:33 +08:00
|
|
|
{
|
|
|
|
int stat = NC_NOERR;
|
2020-09-28 02:43:46 +08:00
|
|
|
NC *nc;
|
|
|
|
NC_FILE_INFO_T* h5 = NULL;
|
|
|
|
NC_GRP_INFO_T* grp = NULL;
|
|
|
|
NC_VAR_INFO_T* var = NULL;
|
|
|
|
struct NC_HDF5_Filter* oldspec = NULL;
|
2021-05-18 09:49:58 +08:00
|
|
|
int flags = 0;
|
|
|
|
htri_t avail = -1;
|
2021-01-07 04:35:59 +08:00
|
|
|
#ifdef HAVE_H5Z_SZIP
|
2020-09-28 02:43:46 +08:00
|
|
|
int havedeflate = 0;
|
|
|
|
int haveszip = 0;
|
2021-01-07 04:35:59 +08:00
|
|
|
#endif
|
2020-02-17 03:59:33 +08:00
|
|
|
|
2020-09-28 02:43:46 +08:00
|
|
|
LOG((2, "%s: ncid 0x%x varid %d", __func__, ncid, varid));
|
2020-02-17 03:59:33 +08:00
|
|
|
|
2020-09-28 02:43:46 +08:00
|
|
|
if((stat = NC_check_id(ncid,&nc))) return stat;
|
|
|
|
assert(nc);
|
2020-02-17 03:59:33 +08:00
|
|
|
|
|
|
|
/* Find info for this file and group and var, and set pointer to each. */
|
|
|
|
if ((stat = nc4_hdf5_find_grp_h5_var(ncid, varid, &h5, &grp, &var)))
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
{stat = THROW(stat); goto done;}
|
2020-02-17 03:59:33 +08:00
|
|
|
|
|
|
|
assert(h5 && var && var->hdr.id == varid);
|
|
|
|
|
2020-09-28 02:43:46 +08:00
|
|
|
/* If the HDF5 dataset has already been created, then it is too
|
|
|
|
* late to set all the extra stuff. */
|
|
|
|
if (!(h5->flags & NC_INDEF))
|
|
|
|
{stat = THROW(NC_EINDEFINE); goto done;}
|
|
|
|
if (!var->ndims)
|
|
|
|
{stat = NC_EINVAL; goto done;} /* For scalars, complain */
|
|
|
|
if (var->created)
|
|
|
|
{stat = THROW(NC_ELATEDEF); goto done;}
|
|
|
|
/* Can't turn on parallel and szip before HDF5 1.10.2. */
|
2020-02-17 03:59:33 +08:00
|
|
|
#ifdef USE_PARALLEL
|
|
|
|
#ifndef HDF5_SUPPORTS_PAR_FILTERS
|
2020-09-28 02:43:46 +08:00
|
|
|
if (h5->parallel == NC_TRUE)
|
|
|
|
{stat = THROW(NC_EINVAL); goto done;}
|
2020-02-17 03:59:33 +08:00
|
|
|
#endif /* HDF5_SUPPORTS_PAR_FILTERS */
|
|
|
|
#endif /* USE_PARALLEL */
|
2020-09-28 02:43:46 +08:00
|
|
|
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
/* Lookup incoming id to see if already defined */
|
2020-09-28 02:43:46 +08:00
|
|
|
switch((stat=NC4_hdf5_filter_lookup(var,id,&oldspec))) {
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
case NC_NOERR: break; /* already defined */
|
|
|
|
case NC_ENOFILTER: break; /*not defined*/
|
|
|
|
default: goto done;
|
|
|
|
}
|
2021-01-07 04:35:59 +08:00
|
|
|
#ifdef HAVE_H5Z_SZIP
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
/* See if deflate &/or szip is defined */
|
2020-09-28 02:43:46 +08:00
|
|
|
switch ((stat = NC4_hdf5_filter_lookup(var,H5Z_FILTER_DEFLATE,NULL))) {
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
case NC_NOERR: havedeflate = 1; break;
|
|
|
|
case NC_ENOFILTER: havedeflate = 0; break;
|
|
|
|
default: goto done;
|
|
|
|
}
|
2020-09-28 02:43:46 +08:00
|
|
|
switch ((stat = NC4_hdf5_filter_lookup(var,H5Z_FILTER_SZIP,NULL))) {
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
case NC_NOERR: haveszip = 1; break;
|
|
|
|
case NC_ENOFILTER: haveszip = 0; break;
|
|
|
|
default: goto done;
|
|
|
|
}
|
2020-09-28 02:43:46 +08:00
|
|
|
#endif /* HAVE_H5Z_SZIP */
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
|
2021-05-18 09:49:58 +08:00
|
|
|
/* See if this filter is missing or not */
|
|
|
|
if((avail = H5Zfilter_avail(id)) < 0)
|
|
|
|
{stat = NC_EHDFERR; goto done;} /* Something in HDF5 went wrong */
|
|
|
|
if(!avail) {
|
|
|
|
NC_HDF5_VAR_INFO_T* hdf5_var = (NC_HDF5_VAR_INFO_T *)var->format_var_info;
|
|
|
|
flags |= NC_HDF5_FILTER_MISSING;
|
|
|
|
/* mark variable as unreadable */
|
|
|
|
hdf5_var->flags |= NC_HDF5_VAR_FILTER_MISSING;
|
|
|
|
}
|
|
|
|
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
/* If incoming filter not already defined, then check for conflicts */
|
|
|
|
if(oldspec == NULL) {
|
|
|
|
if(id == H5Z_FILTER_DEFLATE) {
|
|
|
|
int level;
|
|
|
|
if(nparams != 1)
|
|
|
|
{stat = THROW(NC_EFILTER); goto done;}/* incorrect no. of parameters */
|
|
|
|
level = (int)params[0];
|
|
|
|
if (level < NC_MIN_DEFLATE_LEVEL || level > NC_MAX_DEFLATE_LEVEL)
|
|
|
|
{stat = THROW(NC_EINVAL); goto done;}
|
2020-09-28 02:43:46 +08:00
|
|
|
#ifdef HAVE_H5Z_SZIP
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
/* If szip compression is already applied, return error. */
|
|
|
|
if(haveszip) {stat = THROW(NC_EINVAL); goto done;}
|
2020-02-17 03:59:33 +08:00
|
|
|
#endif
|
2020-09-28 02:43:46 +08:00
|
|
|
}
|
2020-02-17 03:59:33 +08:00
|
|
|
#ifdef HAVE_H5Z_SZIP
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
if(id == H5Z_FILTER_SZIP) { /* Do error checking */
|
|
|
|
if(nparams != 2)
|
|
|
|
{stat = THROW(NC_EFILTER); goto done;}/* incorrect no. of parameters */
|
|
|
|
/* Pixels per block must be an even number, < 32. */
|
|
|
|
if (params[1] % 2 || params[1] > NC_MAX_PIXELS_PER_BLOCK)
|
|
|
|
{stat = THROW(NC_EINVAL); goto done;}
|
|
|
|
/* If zlib compression is already applied, return error. */
|
|
|
|
if(havedeflate) {stat = THROW(NC_EINVAL); goto done;}
|
|
|
|
}
|
2020-02-17 03:59:33 +08:00
|
|
|
#else /*!HAVE_H5Z_SZIP*/
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
if(id == H5Z_FILTER_SZIP)
|
|
|
|
{stat = THROW(NC_EFILTER); goto done;} /* Not allowed */
|
2020-02-17 03:59:33 +08:00
|
|
|
#endif
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
/* Filter => chunking */
|
|
|
|
var->storage = NC_CHUNKED;
|
|
|
|
/* Determine default chunksizes for this variable unless already specified */
|
|
|
|
if(var->chunksizes && !var->chunksizes[0]) {
|
|
|
|
/* Should this throw error? */
|
|
|
|
if((stat = nc4_find_default_chunksizes2(grp, var)))
|
|
|
|
goto done;
|
|
|
|
/* Adjust the cache. */
|
|
|
|
if ((stat = nc4_adjust_var_cache(grp, var)))
|
|
|
|
goto done;
|
|
|
|
}
|
|
|
|
}
|
2020-02-17 03:59:33 +08:00
|
|
|
#ifdef HAVE_H5Z_SZIP
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
/* More error checking */
|
2020-02-17 03:59:33 +08:00
|
|
|
if(id == H5Z_FILTER_SZIP) { /* szip X chunking error checking */
|
|
|
|
/* For szip, the pixels_per_block parameter must not be greater
|
|
|
|
* than the number of elements in a chunk of data. */
|
|
|
|
size_t num_elem = 1;
|
|
|
|
int d;
|
|
|
|
for (d = 0; d < var->ndims; d++)
|
2020-07-03 00:55:34 +08:00
|
|
|
if (var->dim[d]->len)
|
|
|
|
num_elem *= var->dim[d]->len;
|
2020-02-17 03:59:33 +08:00
|
|
|
/* Pixels per block must be <= number of elements. */
|
|
|
|
if (params[1] > num_elem)
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
{stat = THROW(NC_EINVAL); goto done;}
|
2020-02-17 03:59:33 +08:00
|
|
|
}
|
|
|
|
#endif
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
/* addfilter can handle case where filter is already defined, and will just replace parameters */
|
2021-05-18 09:49:58 +08:00
|
|
|
if((stat = NC4_hdf5_addfilter(var,id,nparams,params,flags)))
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
goto done;
|
2020-02-17 03:59:33 +08:00
|
|
|
#ifdef USE_PARALLEL
|
2020-02-29 02:36:58 +08:00
|
|
|
#ifdef HDF5_SUPPORTS_PAR_FILTERS
|
|
|
|
/* Switch to collective access. HDF5 requires collevtive access
|
2020-02-17 03:59:33 +08:00
|
|
|
* for filter use with parallel I/O. */
|
|
|
|
if (h5->parallel)
|
|
|
|
var->parallel_access = NC_COLLECTIVE;
|
2020-02-29 02:36:58 +08:00
|
|
|
#else
|
|
|
|
if (h5->parallel)
|
This PR adds EXPERIMENTAL support for accessing data in the
cloud using a variant of the Zarr protocol and storage
format. This enhancement is generically referred to as "NCZarr".
The data model supported by NCZarr is netcdf-4 minus the user-defined
types and the String type. In this sense it is similar to the CDF-5
data model.
More detailed information about enabling and using NCZarr is
described in the document NUG/nczarr.md and in a
[Unidata Developer's blog entry](https://www.unidata.ucar.edu/blogs/developer/en/entry/overview-of-zarr-support-in).
WARNING: this code has had limited testing, so do use this version
for production work. Also, performance improvements are ongoing.
Note especially the following platform matrix of successful tests:
Platform | Build System | S3 support
------------------------------------
Linux+gcc | Automake | yes
Linux+gcc | CMake | yes
Visual Studio | CMake | no
Additionally, and as a consequence of the addition of NCZarr,
major changes have been made to the Filter API. NOTE: NCZarr
does not yet support filters, but these changes are enablers for
that support in the future. Note that it is possible
(probable?) that there will be some accidental reversions if the
changes here did not correctly mimic the existing filter testing.
In any case, previously filter ids and parameters were of type
unsigned int. In order to support the more general zarr filter
model, this was all converted to char*. The old HDF5-specific,
unsigned int operations are still supported but they are
wrappers around the new, char* based nc_filterx_XXX functions.
This entailed at least the following changes:
1. Added the files libdispatch/dfilterx.c and include/ncfilter.h
2. Some filterx utilities have been moved to libdispatch/daux.c
3. A new entry, "filter_actions" was added to the NCDispatch table
and the version bumped.
4. An overly complex set of structs was created to support funnelling
all of the filterx operations thru a single dispatch
"filter_actions" entry.
5. Move common code to from libhdf5 to libsrc4 so that it is accessible
to nczarr.
Changes directly related to Zarr:
1. Modified CMakeList.txt and configure.ac to support both C and C++
-- this is in support of S3 support via the awd-sdk libraries.
2. Define a size64_t type to support nczarr.
3. More reworking of libdispatch/dinfermodel.c to
support zarr and to regularize the structure of the fragments
section of a URL.
Changes not directly related to Zarr:
1. Make client-side filter registration be conditional, with default off.
2. Hack include/nc4internal.h to make some flags added by Ed be unique:
e.g. NC_CREAT, NC_INDEF, etc.
3. cleanup include/nchttp.h and libdispatch/dhttp.c.
4. Misc. changes to support compiling under Visual Studio including:
* Better testing under windows for dirent.h and opendir and closedir.
5. Misc. changes to the oc2 code to support various libcurl CURLOPT flags
and to centralize error reporting.
6. By default, suppress the vlen tests that have unfixed memory leaks; add option to enable them.
7. Make part of the nc_test/test_byterange.sh test be contingent on remotetest.unidata.ucar.edu being accessible.
Changes Left TO-DO:
1. fix provenance code, it is too HDF5 specific.
2020-06-29 08:02:47 +08:00
|
|
|
{stat = THROW(NC_EINVAL); goto done;}
|
2020-02-17 03:59:33 +08:00
|
|
|
#endif /* HDF5_SUPPORTS_PAR_FILTERS */
|
|
|
|
#endif /* USE_PARALLEL */
|
2020-09-28 02:43:46 +08:00
|
|
|
|
|
|
|
done:
|
|
|
|
return stat;
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
NC4_hdf5_inq_var_filter_ids(int ncid, int varid, size_t* nfiltersp, unsigned int* ids)
|
|
|
|
{
|
|
|
|
int stat = NC_NOERR;
|
|
|
|
NC *nc;
|
|
|
|
NC_FILE_INFO_T* h5 = NULL;
|
|
|
|
NC_GRP_INFO_T* grp = NULL;
|
|
|
|
NC_VAR_INFO_T* var = NULL;
|
|
|
|
NClist* flist = NULL;
|
|
|
|
size_t nfilters;
|
|
|
|
|
|
|
|
LOG((2, "%s: ncid 0x%x varid %d", __func__, ncid, varid));
|
|
|
|
|
|
|
|
if((stat = NC_check_id(ncid,&nc))) return stat;
|
|
|
|
assert(nc);
|
|
|
|
|
|
|
|
/* Find info for this file and group and var, and set pointer to each. */
|
|
|
|
if ((stat = nc4_hdf5_find_grp_h5_var(ncid, varid, &h5, &grp, &var)))
|
|
|
|
{stat = THROW(stat); goto done;}
|
|
|
|
|
|
|
|
assert(h5 && var && var->hdr.id == varid);
|
|
|
|
|
|
|
|
flist = var->filters;
|
|
|
|
|
|
|
|
nfilters = nclistlength(flist);
|
|
|
|
if(nfilters > 0 && ids != NULL) {
|
|
|
|
int k;
|
|
|
|
for(k=0;k<nfilters;k++) {
|
|
|
|
struct NC_HDF5_Filter* f = (struct NC_HDF5_Filter*)nclistget(flist,k);
|
|
|
|
ids[k] = f->filterid;
|
2020-02-17 03:59:33 +08:00
|
|
|
}
|
|
|
|
}
|
2020-09-28 02:43:46 +08:00
|
|
|
if(nfiltersp) *nfiltersp = nfilters;
|
|
|
|
|
2020-02-17 03:59:33 +08:00
|
|
|
done:
|
2020-09-28 02:43:46 +08:00
|
|
|
return stat;
|
|
|
|
|
2020-02-17 03:59:33 +08:00
|
|
|
}
|
|
|
|
|
2020-09-28 02:43:46 +08:00
|
|
|
int
|
|
|
|
NC4_hdf5_inq_var_filter_info(int ncid, int varid, unsigned int id, size_t* nparamsp, unsigned int* params)
|
2020-02-17 03:59:33 +08:00
|
|
|
{
|
2020-09-28 02:43:46 +08:00
|
|
|
int stat = NC_NOERR;
|
|
|
|
NC *nc;
|
|
|
|
NC_FILE_INFO_T* h5 = NULL;
|
|
|
|
NC_GRP_INFO_T* grp = NULL;
|
|
|
|
NC_VAR_INFO_T* var = NULL;
|
|
|
|
struct NC_HDF5_Filter* spec = NULL;
|
|
|
|
|
|
|
|
LOG((2, "%s: ncid 0x%x varid %d", __func__, ncid, varid));
|
|
|
|
|
|
|
|
if((stat = NC_check_id(ncid,&nc))) return stat;
|
|
|
|
assert(nc);
|
|
|
|
|
|
|
|
/* Find info for this file and group and var, and set pointer to each. */
|
|
|
|
if ((stat = nc4_hdf5_find_grp_h5_var(ncid, varid, &h5, &grp, &var)))
|
|
|
|
{stat = THROW(stat); goto done;}
|
|
|
|
|
|
|
|
assert(h5 && var && var->hdr.id == varid);
|
|
|
|
|
|
|
|
if((stat = NC4_hdf5_filter_lookup(var,id,&spec))) goto done;
|
|
|
|
if(nparamsp) *nparamsp = spec->nparams;
|
|
|
|
if(params && spec->nparams > 0) {
|
|
|
|
memcpy(params,spec->params,sizeof(unsigned int)*spec->nparams);
|
2020-02-17 03:59:33 +08:00
|
|
|
}
|
2020-09-28 02:43:46 +08:00
|
|
|
|
|
|
|
done:
|
|
|
|
return stat;
|
|
|
|
|
2020-02-17 03:59:33 +08:00
|
|
|
}
|
2021-05-18 09:49:58 +08:00
|
|
|
|
|
|
|
/* Return ID for the first missing filter; 0 if no missing filters */
|
|
|
|
int
|
|
|
|
NC4_hdf5_find_missing_filter(NC_VAR_INFO_T* var, unsigned int* idp)
|
|
|
|
{
|
|
|
|
int i,stat = NC_NOERR;
|
|
|
|
NClist* flist = (NClist*)var->filters;
|
|
|
|
int id = 0;
|
|
|
|
|
|
|
|
for(i=0;i<nclistlength(flist);i++) {
|
|
|
|
struct NC_HDF5_Filter* spec = (struct NC_HDF5_Filter*)nclistget(flist,i);
|
|
|
|
if(spec->flags & NC_HDF5_FILTER_MISSING) {id = spec->filterid; break;}
|
|
|
|
}
|
|
|
|
if(idp) *idp = id;
|
|
|
|
return stat;
|
|
|
|
}
|
2021-09-03 07:04:26 +08:00
|
|
|
|
|
|
|
int
|
|
|
|
NC4_hdf5_filter_initialize(void)
|
|
|
|
{
|
|
|
|
return NC_NOERR;
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
NC4_hdf5_filter_finalize(void)
|
|
|
|
{
|
|
|
|
return NC_NOERR;
|
|
|
|
}
|