blosc2 2.3.0__tar.gz → 2.3.2__tar.gz

{blosc2-2.3.0 → blosc2-2.3.2}/.github/workflows/cibuildwheels.yml

@@ -68,6 +68,9 @@ jobs:
       - name: Build wheels
         run: |
           python -m cibuildwheel --output-dir wheelhouse
+        # Starting with C-Blosc2 2.11.3, we should not need to deactivate AVX512
+        #env:
+          # CIBW_CONFIG_SETTINGS: "--build-option=-DDEACTIVATE_AVX512:BOOL=YES"
 
       - name: Upload wheels
         uses: actions/upload-artifact@v3
@@ -116,6 +119,8 @@ jobs:
           cd ./dist
           tar -xzf blosc2-*.tar.gz
           cd ./blosc2-*/
+          # Starting with C-Blosc2 2.11.3, we should not need to deactivate AVX512
+          # python setup.py build_ext --inplace -DDEACTIVATE_AVX512:BOOL=YES
           python setup.py build_ext --inplace
 
       - name: Test sdist package with pytest

{blosc2-2.3.0 → blosc2-2.3.2}/ANNOUNCE.rst

@@ -1,12 +1,13 @@
-Announcing Python-Blosc2 2.3.0
+Announcing Python-Blosc2 2.3.1
 ==============================
 
-This new minor release adds methods to handle NDArrays serialized into memory
-frames, and allows opening super-chunks stored in container files (like HDF5)
-by their offset.  C-Blosc2 is updated to the latest stable version, which
-provides support for AVX512 and fixes compatibility with ARM platforms
-(e.g. Raspberry Pi).  Cython 3 is now required for building, and other minor
-fixes and additions are included as well.
+The new minor release 2.3 adds methods to handle NDArrays serialized into
+memory frames, and allows opening super-chunks stored in container files (like
+HDF5) by their offset.  C-Blosc2 is updated to the latest stable version,
+which provides support for AVX512 (temporarily disabled in wheels) and fixes
+compatibility with ARM platforms (e.g. Raspberry Pi).  Cython 3 is now
+required for building, and other minor fixes and additions are included as
+well.
 
 For more info, you can have a look at the release notes in:
 

{blosc2-2.3.0/blosc2.egg-info → blosc2-2.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: blosc2
-Version: 2.3.0
+Version: 2.3.2
 Summary: Python wrapper for the C-Blosc2 library
 Author-email: Blosc Development Team <blosc@blosc.org>
 Maintainer-email: Blosc Development Team <blosc@blosc.org>

{blosc2-2.3.0 → blosc2-2.3.2}/RELEASE_NOTES.md

@@ -1,5 +1,21 @@
 # Release notes
 
+## Changes from 2.3.1 to 2.3.2
+
+* Support for `pathlib.Path` objects in all the places where `urlpath` is
+  used (e.g. `blosc2.open()`). Thanks to Marta Iborra.
+
+* Included docs for `SChunk.fill_special()` and `NDArray.dtype`. Thanks
+  to Francesc Alted.
+
+* Upgrade to latest C-Blosc2 2.11.3. It fixes a bug preventing the use of
+  typesize > 255 in frames.  Now you can use a typesize up to 2**31-1.
+
+## Changes from 2.3.0 to 2.3.1
+
+* Temporarily disable AVX512 support in C-Blosc2 for wheels built by CI until
+  run-time detection works properly.
+
 ## Changes from 2.2.9 to 2.3.0
 
 * Require at least Cython 3 for building. Using previous versions worked but

{blosc2-2.3.0 → blosc2-2.3.2}/RELEASING.rst

@@ -4,10 +4,10 @@ python-blosc2 release procedure
 Preliminaries
 -------------
 
-* Check that `VERSION` file contains the correct number.
+* Check that ``VERSION`` file contains the correct number.
 
 * Make sure that the c-blosc2 submodule is updated to the latest version (or a specific
-  version that will be documented in the `RELEASE_NOTES.md`)::
+  version that will be documented in the ``RELEASE_NOTES.md``)::
 
     cd blosc2/c-blosc2
     git pull
@@ -22,9 +22,9 @@ Preliminaries
 
     PYTHONPATH=. python -c "import blosc2; blosc2.print_versions()"
 
-is printing the correct versions.
+  is printing the correct versions.
 
-* Make sure that `RELEASE_NOTES.md` and `ANNOUNCE.rst` are up to date with the latest news
+* Make sure that ``RELEASE_NOTES.md`` and ``ANNOUNCE.rst`` are up to date with the latest news
   in the release.
 
 * Commit the changes::
@@ -33,12 +33,12 @@ is printing the correct versions.
     git push
 
 * Double check that the supported Python versions for the wheels are the correct ones
-  (`.github/workflows/cibuildwheels.yml`).  Add/remove Python version if needed.
-  Also, update the `classifiers` field for the supported Python versions.
+  (``.github/workflows/cibuildwheels.yml``).  Add/remove Python version if needed.
+  Also, update the ``classifiers`` field for the supported Python versions.
 
 * Check that the metainfo for the package is correct::
 
-    python -m build --sdist 
+    python -m build --sdist
     twine check dist/*
 
 
@@ -47,24 +47,24 @@ Tagging
 
 * Create a (signed, if possible) tag ``X.Y.Z`` from ``main``.  Use the next message::
 
-    git tag -a vX.Y.Z -m "Tagging version X.Y.Z"
+    git tag -a vX.Y.Z -m "Tagging python-blosc2 version X.Y.Z"
 
 * Push the tag to the github repo::
 
     git push --tags
 
-* Go to Blosc/blogsite repo and click on the `Re-run all jobs` button to regenerate the
-  documentation and check that it has been correctly updated in https://www.blosc.org.
+* Make sure that the tag is passing the tests in continuous integration (this
+  may take more than an hour).
+
+* Go to ``Blosc/blogsite`` repo, then to "Actions", click on the most recent
+  workflow run (at the top of the list), and then click on the "Re-run all
+  jobs" button to regenerate the documentation and check that it has been
+  correctly updated in https://www.blosc.org.
 
 
 Checking packaging
 ------------------
 
-* Do an actual release in github by visiting:
-  https://github.com/Blosc/python-blosc2/releases/new
-  Add the notes specific for this release.
-  This will upload the wheels to PyPI too.
-
 * Check that the package (and wheels!) have been uploaded to PyPI
   (they should have been created when GHA would finish the tag trigger):
   https://pypi.org/project/blosc2/
@@ -72,16 +72,23 @@ Checking packaging
 * Check that the packages and wheels are sane::
 
     python -m pip install blosc2 -U
+    python -m pip install -r requirements-test-wheels.txt
     cd tests
-    python -c"import blosc2; blosc2.print_versions()"
+    python -c "import blosc2; blosc2.print_versions()"
     pytest
 
+* Do an actual release in github by visiting:
+  https://github.com/Blosc/python-blosc2/releases/new
+  Add the notes specific for this release
+  (you may want to remove newlines for aesthetics).
+  This will upload the wheels to PyPI too.
+
 
 Announcing
 ----------
 
 * Send an announcement to the Blosc and PyData lists.  Use the ``ANNOUNCE.rst`` file as
-skeleton (or possibly as the definitive version). Start the subject with ANN:.
+  skeleton (or possibly as the definitive version). Start the subject with ANN:.
 
 * Announce in Mastodon via https://fosstodon.org/@Blosc2 account and rejoice.
 
@@ -93,13 +100,15 @@ Post-release actions
 
     cd $HOME/blosc/python-blosc2
 
-* Create new headers for adding new features in ``RELEASE_NOTES.md``
-  add this place-holder:
+* Create a new header for adding new features in ``RELEASE_NOTES.md``
+  with a placeholder text::
+
+    ## Changes from X.Y.Z to X.Y.(Z+1)
 
-  XXX version-specific blurb XXX
+    XXX version-specific blurb XXX
 
 * Edit ``VERSION`` in master to increment the version to the next
-  minor one (i.e. X.Y.Z --> X.Y.(Z+1).dev0).
+  patch one (i.e. X.Y.Z --> X.Y.(Z+1).dev0).
 
 * Commit your changes with::
 

blosc2-2.3.2/VERSION

@@ -0,0 +1 @@
+2.3.2

{blosc2-2.3.0 → blosc2-2.3.2}/bench/compress_numpy.py

@@ -38,12 +38,16 @@ t0 = time.time()
 for i in range(NREP):
     np.copyto(out_, in_)
 tcpy = (time.time() - t0) / NREP
-print("  *** np.copyto() *** Time for memcpy():\t{:.3f} s\t({:.2f} GB/s)".format(tcpy, (N * 8 / tcpy) / 2**30))
+print(
+    "  *** np.copyto() *** Time for memcpy():\t{:.3f} s\t({:.2f} GB/s)".format(
+        tcpy, (N * 8 / tcpy) / 2**30
+    )
+)
 
 print("\nTimes for compressing/decompressing:")
 for in_, label in arrays:
     print(f"\n*** {label} ***")
-    for codec in blosc2.Codec:
+    for codec in blosc2.compressor_list():
         for filter in (
             blosc2.Filter.NOFILTER,
             blosc2.Filter.SHUFFLE,
@@ -61,8 +65,7 @@ for in_, label in arrays:
             td = (time.time() - t0) / NREP
             assert np.array_equal(in_, out)
             print(
-                "  *** {:-7s}, {:-10s} *** {:6.3f} s ({:.2f} GB/s) / {:5.3f} s ({:.2f} GB/s)"
-                .format(
+                "  *** {:15s}, {:20s} *** {:6.3f} s ({:.2f} GB/s) / {:5.3f} s ({:.2f} GB/s)".format(
                     codec,
                     filter,
                     tc,

{blosc2-2.3.0 → blosc2-2.3.2}/bench/ndarray/compare_getslice.py

@@ -89,8 +89,8 @@ if persistent:
     blosc2.remove_urlpath(fname_h5py)
 
 # Create datasets in different formats
-# content = np.random.normal(0, 1, int(np.prod(shape))).reshape(shape)
-content = np.linspace(0, 1, int(np.prod(shape))).reshape(shape)
+# content = np.random.normal(0, 1, int(np.prod(shape)), dtype=dtype).reshape(shape)
+content = np.linspace(0, 1, int(np.prod(shape)), dtype=dtype).reshape(shape)
 
 print("\nCreating datasets...")
 # Create and fill a NDArray
@@ -108,7 +108,7 @@ print(f"Time for filling array (blosc2): {t:.3f} s ({speed:.2f} GB/s) ; cratio:
 t0 = time()
 compressor = numcodecs.Blosc(cname=cname, clevel=clevel, shuffle=zfilter, blocksize=blocksize)
 numcodecs.blosc.set_nthreads(nthreads)
-z = zarr.open(fname_zarr, shape=shape, chunks=chunks, dtype=dtype, compressor=compressor)
+z = zarr.open(fname_zarr, shape=shape, chunks=chunks, dtype=content.dtype, compressor=compressor)
 z[:] = content
 t = time() - t0
 speed = dset_size / (t * 2**30)
@@ -136,7 +136,7 @@ if persistent:
     h5pyf = h5py.File(fname_h5py, "w")
 else:
     h5pyf = h5py.File(fname_h5py, "w", driver="core", backing_store=False)
-h5d = h5pyf.create_dataset("dataset", dtype=dtype, data=content, chunks=chunks, **filters)
+h5d = h5pyf.create_dataset("dataset", dtype=content.dtype, data=content, chunks=chunks, **filters)
 t = time() - t0
 speed = dset_size / (t * 2**30)
 if persistent:

{blosc2-2.3.0 → blosc2-2.3.2}/blosc2/blosc2_ext.pyx

@@ -10,6 +10,7 @@
 
 import ast
 import atexit
+import pathlib
 
 import _ctypes
 
@@ -856,6 +857,8 @@ cdef class SChunk:
         # because its value is referenced via a C-pointer
         urlpath = kwargs.get("urlpath", None)
         if urlpath is not None:
+            if isinstance(urlpath, pathlib.PurePath):
+                urlpath = str(urlpath)
             self._urlpath = urlpath.encode() if isinstance(urlpath, str) else urlpath
             kwargs["urlpath"] = self._urlpath
         self.mode = mode = kwargs.get("mode", "a")
@@ -1255,9 +1258,9 @@ cdef class SChunk:
 
     def get_slice(self, start=0, stop=None, out=None):
         cdef int64_t nitems = self.schunk.nbytes // self.schunk.typesize
-        start, stop = self._massage_key(start, stop, nitems)
-        if start >= nitems:
-            raise ValueError("`start` cannot be greater or equal than the SChunk nitems")
+        start, stop, _ = slice(start, stop, 1).indices(nitems)
+        if start >= stop:
+            return b''
 
         cdef Py_ssize_t nbytes = (stop - start) * self.schunk.typesize
         cdef Py_buffer *buf
@@ -2007,7 +2010,6 @@ cdef class NDArray:
     def dtype(self):
         """
         Data-type of the array’s elements.
-        If it is an old caterva array, it will be a bytes string of `typesize` length.
         """
         if self._dtype is not None:
             return self._dtype
@@ -2185,6 +2187,8 @@ cdef b2nd_context_t* create_b2nd_context(shape, chunks, blocks, dtype, kwargs):
         kwargs['contiguous'] = False if urlpath is None else True
 
     if urlpath is not None:
+        if isinstance(urlpath, pathlib.PurePath):
+            urlpath = str(urlpath)
         _urlpath = urlpath.encode() if isinstance(urlpath, str) else urlpath
         kwargs["urlpath"] = _urlpath
 

{blosc2-2.3.0 → blosc2-2.3.2}/blosc2/c-blosc2/ANNOUNCE.md

@@ -1,10 +1,10 @@
-# Announcing C-Blosc2 2.11.2
+# Announcing C-Blosc2 2.11.3
 A fast, compressed and persistent binary data store library for C.
 
 ## What is new?
 
-This is a maintenance release for fixing compilation on Raspberry Pi (4).
-Only affects to ARM non-64bit builds.
+This is a maintenance release for allowing typesizes > 255 in frames
+and other small fixes.
 
 For more info, please see the release notes in:
 

{blosc2-2.3.0 → blosc2-2.3.2}/blosc2/c-blosc2/CMakeLists.txt

@@ -285,7 +285,10 @@ if(CMAKE_SYSTEM_PROCESSOR STREQUAL i386 OR
         else()
             set(COMPILER_SUPPORT_AVX2 FALSE)
         endif()
-        if(CMAKE_C_COMPILER_VERSION VERSION_GREATER 4.9 OR CMAKE_C_COMPILER_VERSION VERSION_EQUAL 4.9)
+        # GCC 10.3.2 (the version in manylinux_2014) seems to have issues supporting dynamic dispatching
+        # of AVX512.  GCC 11.4 is the first minimal version that works well here.
+        # That means that Linux wheels will have AVX512 disabled, but that's life.
+        if(CMAKE_C_COMPILER_VERSION VERSION_GREATER 11.4 OR CMAKE_C_COMPILER_VERSION VERSION_EQUAL 11.4)
             set(COMPILER_SUPPORT_AVX512 TRUE)
         else()
             set(COMPILER_SUPPORT_AVX512 FALSE)
@@ -297,7 +300,9 @@ if(CMAKE_SYSTEM_PROCESSOR STREQUAL i386 OR
         else()
             set(COMPILER_SUPPORT_AVX2 FALSE)
         endif()
-        if(CMAKE_C_COMPILER_VERSION VERSION_GREATER 10.0 OR CMAKE_C_COMPILER_VERSION VERSION_EQUAL 10.0)
+        # Clang 13 is the minimum version that we know that works with AVX512 dynamic dispatch.
+        # Perhaps lesser versions work too, better to err on the safe side.
+        if(CMAKE_C_COMPILER_VERSION VERSION_GREATER 13.0 OR CMAKE_C_COMPILER_VERSION VERSION_EQUAL 13.0)
             set(COMPILER_SUPPORT_AVX512 TRUE)
         else()
             set(COMPILER_SUPPORT_AVX512 FALSE)

{blosc2-2.3.0 → blosc2-2.3.2}/blosc2/c-blosc2/RELEASE_NOTES.md

@@ -1,6 +1,19 @@
 Release notes for C-Blosc2
 ==========================
 
+Changes from 2.11.2 to 2.11.3
+=============================
+
+* Frames accept now typesizes that are larger than 255 (and up to 2**31).
+  See https://github.com/PyTables/PyTables/issues/1086.  Thanks to
+  @chris-allan for the report.
+
+* AVX512 runtime dispatching has been fixed (basically disabled) for GCC
+  versions <= 10.
+
+* Use typedef for blosc_timestamp_t.  Thanks to Magnus Ulimoen.
+
+
 Changes from 2.11.1 to 2.11.2
 =============================
 

{blosc2-2.3.0 → blosc2-2.3.2}/blosc2/c-blosc2/blosc/frame.c

@@ -428,8 +428,8 @@ int get_header_info(blosc2_frame_s *frame, int32_t *header_len, int64_t *frame_l
   }
   if (typesize != NULL) {
     from_big(typesize, framep + FRAME_TYPESIZE, sizeof(*typesize));
-    if (*typesize <= 0 || *typesize > BLOSC_MAX_TYPESIZE) {
-      BLOSC_TRACE_ERROR("`typesize` is zero or greater than max allowed.");
+    if (*typesize <= 0) {
+      BLOSC_TRACE_ERROR("`typesize` cannot be zero or negative.");
       return BLOSC2_ERROR_INVALID_HEADER;
     }
   }

{blosc2-2.3.0 → blosc2-2.3.2}/blosc2/c-blosc2/include/blosc2.h

@@ -83,10 +83,10 @@ extern "C" {
 /* Version numbers */
 #define BLOSC2_VERSION_MAJOR    2    /* for major interface/format changes  */
 #define BLOSC2_VERSION_MINOR    11   /* for minor interface/format changes  */
-#define BLOSC2_VERSION_RELEASE  2    /* for tweaks, bug-fixes, or development */
+#define BLOSC2_VERSION_RELEASE  3    /* for tweaks, bug-fixes, or development */
 
-#define BLOSC2_VERSION_STRING   "2.11.2"  /* string version.  Sync with above! */
-#define BLOSC2_VERSION_DATE     "$Date:: 2023-11-07 #$"    /* date version */
+#define BLOSC2_VERSION_STRING   "2.11.3"  /* string version.  Sync with above! */
+#define BLOSC2_VERSION_DATE     "$Date:: 2023-11-30 #$"    /* date version */
 
 
 /* The maximum number of dimensions for Blosc2 NDim arrays */
@@ -2306,9 +2306,9 @@ BLOSC_EXPORT int blosc2_vlmeta_get_names(blosc2_schunk *schunk, char **names);
 
 /* The type of timestamp used on this system. */
 #if defined(_WIN32)
-#define blosc_timestamp_t LARGE_INTEGER
+typedef LARGE_INTEGER blosc_timestamp_t;
 #else
-#define blosc_timestamp_t struct timespec
+typedef struct timespec blosc_timestamp_t;
 #endif
 
 /*

{blosc2-2.3.0 → blosc2-2.3.2}/blosc2/core.py

@@ -8,6 +8,7 @@
 import copy
 import math
 import os
+import pathlib
 import pickle
 import sys
 
@@ -1370,6 +1371,8 @@ def remove_urlpath(path):
     None
     """
     if path is not None:
+        if isinstance(path, pathlib.PurePath):
+            path = str(path)
         path = path.encode("utf-8") if isinstance(path, str) else path
         blosc2_ext.remove_urlpath(path)
 

{blosc2-2.3.0 → blosc2-2.3.2}/blosc2/ndarray.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import math
 from collections import namedtuple
 from typing import Sequence
 
@@ -19,19 +20,10 @@ def process_key(key, shape):
     return key, mask
 
 
-def prod(list):
-    prod = 1
-    for li in list:
-        prod *= li
-    return prod
-
-
 def get_ndarray_start_stop(ndim, key, shape):
     start = tuple(s.start if s.start is not None else 0 for s in key)
     stop = tuple(s.stop if s.stop is not None else sh for s, sh in zip(key, shape))
-
-    size = prod([stop[i] - start[i] for i in range(ndim)])
-
+    size = math.prod([stop[i] - start[i] for i in range(ndim)])
     return start, stop, size
 
 

{blosc2-2.3.0 → blosc2-2.3.2}/blosc2/schunk.py

@@ -7,6 +7,7 @@
 #######################################################################
 
 import os
+import pathlib
 from collections import namedtuple
 from collections.abc import Mapping, MutableMapping
 
@@ -981,6 +982,8 @@ def open(urlpath, mode="a", offset=0, **kwargs):
     True
 
     """
+    if isinstance(urlpath, pathlib.PurePath):
+        urlpath = str(urlpath)
     if not os.path.exists(urlpath):
         raise FileNotFoundError(f"No such file or directory: {urlpath}")
     return blosc2_ext.open(urlpath, mode, offset, **kwargs)

blosc2-2.3.2/blosc2/version.py

@@ -0,0 +1 @@
+__version__ = "2.3.2"

{blosc2-2.3.0 → blosc2-2.3.2/blosc2.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: blosc2
-Version: 2.3.0
+Version: 2.3.2
 Summary: Python wrapper for the C-Blosc2 library
 Author-email: Blosc Development Team <blosc@blosc.org>
 Maintainer-email: Blosc Development Team <blosc@blosc.org>

{blosc2-2.3.0 → blosc2-2.3.2}/blosc2.egg-info/SOURCES.txt

@@ -925,6 +925,7 @@ doc/reference/autofiles/schunk/blosc2.schunk.SChunk.__setitem__.rst
 doc/reference/autofiles/schunk/blosc2.schunk.SChunk.append_data.rst
 doc/reference/autofiles/schunk/blosc2.schunk.SChunk.decompress_chunk.rst
 doc/reference/autofiles/schunk/blosc2.schunk.SChunk.delete_chunk.rst
+doc/reference/autofiles/schunk/blosc2.schunk.SChunk.fill_special.rst
 doc/reference/autofiles/schunk/blosc2.schunk.SChunk.filler.rst
 doc/reference/autofiles/schunk/blosc2.schunk.SChunk.get_chunk.rst
 doc/reference/autofiles/schunk/blosc2.schunk.SChunk.get_slice.rst
@@ -1006,6 +1007,7 @@ tests/test_compressors.py
 tests/test_decompress.py
 tests/test_iterchunks.py
 tests/test_open.py
+tests/test_pathlib.py
 tests/test_postfilters.py
 tests/test_prefilters.py
 tests/test_python_blosc.py

{blosc2-2.3.0 → blosc2-2.3.2}/doc/python-blosc2.rst

@@ -1,8 +1,10 @@
 Python-Blosc2 documentation
 ============================
 
-Python-Blosc2 is a Python package
-that wraps C-Blosc2, the newest version of the Blosc compressor.
+Python-Blosc2 is a Python package for handling large NDim and plain datasets stored in a compressed state either in-memory or on-disk.
+Under the hood, it uses [C-Blosc2](https://www.blosc.org/c-blosc2/c-blosc2.html), a fast, compressed and persistent data store library
+for C.
+
 
 .. panels::
     :card: + intro-card text-center

blosc2-2.3.2/doc/reference/autofiles/schunk/blosc2.schunk.SChunk.fill_special.rst

@@ -0,0 +1,6 @@
+SChunk.fill\_special
+=======================
+
+.. currentmodule:: blosc2.schunk
+
+.. automethod:: SChunk.fill_special

{blosc2-2.3.0 → blosc2-2.3.2}/doc/reference/ndarray_api.rst

@@ -37,6 +37,7 @@ Attributes
     blocks
     blocksize
     chunksize
+    dtype
     schunk
     size
 

{blosc2-2.3.0 → blosc2-2.3.2}/doc/reference/schunk_api.rst

@@ -21,6 +21,7 @@ Methods
     SChunk.insert_data
     SChunk.iterchunks
     SChunk.iterchunks_info
+    SChunk.fill_special
     SChunk.update_chunk
     SChunk.update_data
     SChunk.get_slice

{blosc2-2.3.0 → blosc2-2.3.2}/examples/filler.py

@@ -6,6 +6,8 @@
 # LICENSE file in the root directory of this source tree)
 #######################################################################
 
+# Fill an SChunk with a filler decorator
+
 import numpy as np
 
 import blosc2
@@ -14,15 +16,11 @@ nchunks = 3
 chunk_len = 200 * 1000
 schunk_dtype = np.dtype(np.float64)
 
-# Set the compression and decompression parameters
-cparams = {"codec": blosc2.Codec.LZ4, "typesize": schunk_dtype.itemsize, "nthreads": 1}
-dparams = {"nthreads": 4}
-contiguous = True
-urlpath = None
-storage = {"contiguous": contiguous, "urlpath": urlpath, "cparams": cparams, "dparams": dparams}
+# Set the compression parameters. We need nthreads=1 for this example.
+cparams = {"typesize": schunk_dtype.itemsize, "nthreads": 1}
 
 # Create empty SChunk
-schunk = blosc2.SChunk(chunksize=chunk_len * schunk_dtype.itemsize, **storage)
+schunk = blosc2.SChunk(chunksize=chunk_len * schunk_dtype.itemsize, cparams=cparams)
 
 # Create operands (can be a SChunk, numpy.ndarray or Python scalar)
 op_dtype = np.dtype(np.int32)

{blosc2-2.3.0 → blosc2-2.3.2}/examples/prefilter.py

@@ -6,6 +6,8 @@
 # LICENSE file in the root directory of this source tree)
 #######################################################################
 
+# Example of prefiltering data before compression
+
 import numpy as np
 
 import blosc2
@@ -15,12 +17,9 @@ input_dtype = np.dtype(np.int32)
 output_dtype = np.dtype(np.float32)
 
 # Set the compression and decompression parameters
-cparams = {"codec": blosc2.Codec.LZ4, "typesize": 4, "nthreads": 1}
+cparams = {"typesize": 4, "nthreads": 1}
 dparams = {"nthreads": 4}
-contiguous = True
-urlpath = None
-storage = {"contiguous": contiguous, "urlpath": urlpath, "cparams": cparams, "dparams": dparams}
-
+storage = {"cparams": cparams, "dparams": dparams}
 # Create empty schunk
 schunk = blosc2.SChunk(chunksize=200 * 1000 * input_dtype.itemsize, **storage)