netcdf-c/ncgen/ncdumpchunks.c

203 lines
5.0 KiB
C
Raw Normal View History

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
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <netcdf.h>
#include <hdf5.h>
#include <H5DSpublic.h>
#if 0
#define FILE_NAME "tst_vars.nc"
#define VAR "v1"
#define RANK 3
static const size_t dimlens[RANK] = {4,4,2};
static const size_t chunksize[RANK] = {2,2,2};
static const size_t chunkcount[RANK] = {2,2,1};
static const size_t CHUNKPROD = 8;
#endif
typedef struct Odometer {
size_t rank; /*rank */
size_t stop[NC_MAX_VAR_DIMS];
size_t max[NC_MAX_VAR_DIMS]; /* max size of ith index */
size_t index[NC_MAX_VAR_DIMS]; /* current value of the odometer*/
} Odometer;
#define floordiv(x,y) ((x) / (y))
#define ceildiv(x,y) (((x) % (y)) == 0 ? ((x) / (y)) : (((x) / (y)) + 1))
Odometer* odom_new(size_t rank, const size_t* stop, const size_t* max);
void odom_free(Odometer* odom);
int odom_more(Odometer* odom);
int odom_next(Odometer* odom);
size_t* odom_indices(Odometer* odom);
size_t odom_offset(Odometer* odom);
void setoffset(Odometer* odom, size_t* chunksizes, hsize_t* offset);
static void
usage(int err)
{
if(err != 0) {
fprintf(stderr,"Error: (%d) %s\n",err,nc_strerror(err));
}
fprintf(stderr,"usage: ncdumpchunks <file> <var>\n");
fflush(stderr);
exit(1);
}
int
main(int argc, char** argv)
{
int i,stat = NC_NOERR;
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
hid_t fileid, grpid, datasetid;
int* chunkdata = NULL; /*[CHUNKPROD];*/
#ifdef HDF5_SUPPORTS_PAR_FILTERS
int r;
hid_t dxpl_id = H5P_DEFAULT; /*data transfer property 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
unsigned int filter_mask = 0;
#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
const char* file_name = NULL;
const char* var_name = NULL;
int ncid, varid, dimids[NC_MAX_VAR_DIMS];
int rank, vtype, storage;
size_t dimlens[NC_MAX_VAR_DIMS];
size_t chunklens[NC_MAX_VAR_DIMS];
size_t chunkcounts[NC_MAX_VAR_DIMS];
size_t chunkprod;
Odometer* odom = NULL;
hsize_t offset[NC_MAX_VAR_DIMS];
if(argc < 3)
usage(0);
file_name = argv[1];
var_name = argv[2];
/* Get the info about the var */
if((stat=nc_open(file_name,0,&ncid))) usage(stat);
if((stat=nc_inq_varid(ncid,var_name,&varid))) usage(stat);
if((stat=nc_inq_var(ncid,varid,NULL,&vtype,&rank,dimids,NULL))) usage(stat);
if(rank == 0) usage(NC_EDIMSIZE);
if((stat=nc_inq_var_chunking(ncid,varid,&storage,chunklens))) usage(stat);
if(storage != NC_CHUNKED) usage(NC_EBADCHUNK);
chunkprod = 1;
for(i=0;i<rank;i++) {
if((stat=nc_inq_dimlen(ncid,dimids[i],&dimlens[i]))) usage(stat);
chunkcounts[i] = ceildiv(dimlens[i],chunklens[i]);
chunkprod *= chunklens[i];
}
if((stat=nc_close(ncid))) usage(stat);
if ((fileid = H5Fopen(file_name, H5F_ACC_RDONLY, H5P_DEFAULT)) < 0) usage(NC_EHDFERR);
if ((grpid = H5Gopen(fileid, "/", H5P_DEFAULT)) < 0) usage(NC_EHDFERR);
if ((datasetid = H5Dopen1(grpid, var_name)) < 0) usage(NC_EHDFERR);
if((odom = odom_new(rank,chunkcounts,dimlens))==NULL) usage(NC_ENOMEM);
if((chunkdata = calloc(sizeof(int),chunkprod))==NULL) usage(NC_ENOMEM);
while(odom_more(odom)) {
setoffset(odom,chunklens,offset);
#ifdef DEBUG
fprintf(stderr,"(");
for(i=0;i<rank;i++)
fprintf(stderr,"%s%lu",(i > 0 ? "," : ""),(unsigned long)offset[i]);
fprintf(stderr,")\n");
fflush(stderr);
#endif
memset(chunkdata,0,sizeof(int)*chunkprod);
#ifdef HDF5_SUPPORTS_PAR_FILTERS
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(H5Dread_chunk(datasetid, dxpl_id, offset, &filter_mask, chunkdata) < 0) abort();
for(r=0;r<rank;r++)
printf("[%lu/%lu]",(unsigned long)odom->index[r],(unsigned long)offset[r]);
printf(" =");
for(r=0;r<chunkprod;r++)
printf(" %02d", chunkdata[r]);
printf("\n");
fflush(stdout);
#endif
odom_next(odom);
}
/* Close up. */
if (H5Dclose(datasetid) < 0) abort();
if (H5Gclose(grpid) < 0) abort();
if (H5Fclose(fileid) < 0) abort();
/* Cleanup */
free(chunkdata);
odom_free(odom);
return 0;
}
Odometer*
odom_new(size_t rank, const size_t* stop, const size_t* max)
{
int i;
Odometer* odom = NULL;
if((odom = calloc(1,sizeof(Odometer))) == NULL)
return NULL;
odom->rank = rank;
for(i=0;i<rank;i++) {
odom->stop[i] = stop[i];
odom->max[i] = max[i];
odom->index[i] = 0;
}
return odom;
}
void
odom_free(Odometer* odom)
{
if(odom) free(odom);
}
int
odom_more(Odometer* odom)
{
return (odom->index[0] < odom->stop[0]);
}
int
odom_next(Odometer* odom)
{
size_t i;
for(i=odom->rank-1;i>=0;i--) {
odom->index[i]++;
if(odom->index[i] < odom->stop[i]) break;
if(i == 0) return 0; /* leave the 0th entry if it overflows */
odom->index[i] = 0; /* reset this position */
}
return 1;
}
/* Get the value of the odometer */
size_t*
odom_indices(Odometer* odom)
{
return odom->index;
}
size_t
odom_offset(Odometer* odom)
{
size_t offset;
int i;
offset = 0;
for(i=0;i<odom->rank;i++) {
offset *= odom->max[i];
offset += odom->index[i];
}
return offset;
}
void
setoffset(Odometer* odom, size_t* chunksizes, hsize_t* offset)
{
int i;
for(i=0;i<odom->rank;i++)
offset[i] = odom->index[i] * chunksizes[i];
}