legend-pydataobj 1.11.12__tar.gz → 1.11.13__tar.gz

{legend_pydataobj-1.11.12 → legend_pydataobj-1.11.13}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: legend_pydataobj
-Version: 1.11.12
+Version: 1.11.13
 Summary: LEGEND Python Data Objects
 Author: The LEGEND Collaboration
 Maintainer: The LEGEND Collaboration

{legend_pydataobj-1.11.12 → legend_pydataobj-1.11.13}/src/legend_pydataobj.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: legend_pydataobj
-Version: 1.11.12
+Version: 1.11.13
 Summary: LEGEND Python Data Objects
 Author: The LEGEND Collaboration
 Maintainer: The LEGEND Collaboration

{legend_pydataobj-1.11.12 → legend_pydataobj-1.11.13}/src/legend_pydataobj.egg-info/SOURCES.txt

@@ -28,6 +28,7 @@ src/lgdo/lh5/core.py
 src/lgdo/lh5/datatype.py
 src/lgdo/lh5/exceptions.py
 src/lgdo/lh5/iterator.py
+src/lgdo/lh5/settings.py
 src/lgdo/lh5/store.py
 src/lgdo/lh5/tools.py
 src/lgdo/lh5/utils.py

{legend_pydataobj-1.11.12 → legend_pydataobj-1.11.13}/src/lgdo/_version.py

@@ -17,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '1.11.12'
-__version_tuple__ = version_tuple = (1, 11, 12)
+__version__ = version = '1.11.13'
+__version_tuple__ = version_tuple = (1, 11, 13)

{legend_pydataobj-1.11.12 → legend_pydataobj-1.11.13}/src/lgdo/lh5/__init__.py

@@ -7,7 +7,6 @@ browsed easily in python like any `HDF5 <https://www.hdfgroup.org>`_ file using
 
 from __future__ import annotations
 
-from ._serializers.write.array import DEFAULT_HDF5_SETTINGS
 from .core import read, read_as, write
 from .iterator import LH5Iterator
 from .store import LH5Store
@@ -15,16 +14,17 @@ from .tools import load_dfs, load_nda, ls, show
 from .utils import read_n_rows
 
 __all__ = [
-    "DEFAULT_HDF5_SETTINGS",
     "LH5Iterator",
     "LH5Store",
     "concat",
+    "default_hdf5_settings",
     "load_dfs",
     "load_nda",
     "ls",
     "read",
     "read_as",
     "read_n_rows",
+    "reset_default_hdf5_settings",
     "show",
     "write",
 ]

{legend_pydataobj-1.11.12 → legend_pydataobj-1.11.13}/src/lgdo/lh5/_serializers/write/array.py

@@ -6,12 +6,11 @@ import h5py
 import numpy as np
 
 from .... import types
+from ... import settings
 from ...exceptions import LH5EncodeError
 
 log = logging.getLogger(__name__)
 
-DEFAULT_HDF5_SETTINGS: dict[str, ...] = {"shuffle": True, "compression": "gzip"}
-
 
 def _h5_write_array(
     obj,
@@ -49,7 +48,7 @@ def _h5_write_array(
             del group[name]
 
         # set default compression options
-        for k, v in DEFAULT_HDF5_SETTINGS.items():
+        for k, v in settings.DEFAULT_HDF5_SETTINGS.items():
             h5py_kwargs.setdefault(k, v)
 
         # compress using the 'compression' LGDO attribute, if available

legend_pydataobj-1.11.13/src/lgdo/lh5/settings.py

@@ -0,0 +1,34 @@
+from __future__ import annotations
+
+from typing import Any
+
+
+def default_hdf5_settings() -> dict[str, Any]:
+    """Returns the HDF5 settings for writing data to disk to the pydataobj defaults.
+
+    Examples
+    --------
+    >>> from lgdo import lh5
+    >>> lh5.DEFAULT_HDF5_SETTINGS["compression"] = "lzf"
+    >>> lh5.write(data, "data", "file.lh5")  # compressed with LZF
+    >>> lh5.DEFAULT_HDF5_SETTINGS = lh5.default_hdf5_settings()
+    >>> lh5.write(data, "data", "file.lh5", "of")  # compressed with default settings (GZIP)
+    """
+
+    return {
+        "shuffle": True,
+        "compression": "gzip",
+    }
+
+
+DEFAULT_HDF5_SETTINGS: dict[str, ...] = default_hdf5_settings()
+"""Global dictionary storing the default HDF5 settings for writing data to disk.
+
+Modify this global variable before writing data to disk with this package.
+
+Examples
+--------
+>>> from lgdo import lh5
+>>> lh5.DEFAULT_HDF5_SETTINGS["compression"] = "lzf"
+>>> lh5.write(data, "data", "file.lh5")  # compressed with LZF
+"""

{legend_pydataobj-1.11.12 → legend_pydataobj-1.11.13}/tests/lh5/test_lh5_store.py

@@ -9,7 +9,6 @@ import pytest
 
 import lgdo
 from lgdo import lh5, types
-from lgdo.lh5 import DEFAULT_HDF5_SETTINGS
 
 
 def test_init():
@@ -97,7 +96,7 @@ def test_read_array(lh5_file):
     with h5py.File(lh5_file) as h5f:
         assert (
             h5f["/data/struct/array"].compression
-            is DEFAULT_HDF5_SETTINGS["compression"]
+            is lh5.settings.DEFAULT_HDF5_SETTINGS["compression"]
         )
 
     lh5_obj, n_rows = store.read("/data/struct_full/array2d", lh5_file)
@@ -168,11 +167,11 @@ def test_read_vov(lh5_file):
     with h5py.File(lh5_file) as h5f:
         assert (
             h5f["/data/struct/vov/cumulative_length"].compression
-            is DEFAULT_HDF5_SETTINGS["compression"]
+            is lh5.settings.DEFAULT_HDF5_SETTINGS["compression"]
         )
         assert (
             h5f["/data/struct/vov/flattened_data"].compression
-            is DEFAULT_HDF5_SETTINGS["compression"]
+            is lh5.settings.DEFAULT_HDF5_SETTINGS["compression"]
         )
 
     lh5_obj, n_rows = store.read("/data/struct/vov3d", lh5_file)
@@ -229,11 +228,11 @@ def test_read_voev(lh5_file):
         assert h5f["/data/struct/voev/encoded_data/flattened_data"].compression is None
         assert (
             h5f["/data/struct/voev/encoded_data/cumulative_length"].compression
-            is DEFAULT_HDF5_SETTINGS["compression"]
+            is lh5.settings.DEFAULT_HDF5_SETTINGS["compression"]
         )
         assert (
             h5f["/data/struct/voev/decoded_size"].compression
-            is DEFAULT_HDF5_SETTINGS["compression"]
+            is lh5.settings.DEFAULT_HDF5_SETTINGS["compression"]
         )
 
 
@@ -312,7 +311,7 @@ def test_read_hdf5_compressed_data(lh5_file):
     with h5py.File(lh5_file) as h5f:
         assert (
             h5f["/data/struct/table/a"].compression
-            is DEFAULT_HDF5_SETTINGS["compression"]
+            is lh5.settings.DEFAULT_HDF5_SETTINGS["compression"]
         )
         assert h5f["/data/struct/table/b"].compression == "gzip"
         assert h5f["/data/struct/table/c"].compression == "gzip"
@@ -333,15 +332,15 @@ def test_read_wftable(lh5_file):
     with h5py.File(lh5_file) as h5f:
         assert (
             h5f["/data/struct/wftable/values"].compression
-            is DEFAULT_HDF5_SETTINGS["compression"]
+            is lh5.settings.DEFAULT_HDF5_SETTINGS["compression"]
         )
         assert (
             h5f["/data/struct/wftable/t0"].compression
-            is DEFAULT_HDF5_SETTINGS["compression"]
+            is lh5.settings.DEFAULT_HDF5_SETTINGS["compression"]
         )
         assert (
             h5f["/data/struct/wftable/dt"].compression
-            is DEFAULT_HDF5_SETTINGS["compression"]
+            is lh5.settings.DEFAULT_HDF5_SETTINGS["compression"]
         )
 
 
@@ -387,11 +386,11 @@ def test_read_wftable_encoded(lh5_file):
         assert h5f["/data/struct/wftable_enc/values/decoded_size"].compression is None
         assert (
             h5f["/data/struct/wftable_enc/t0"].compression
-            is DEFAULT_HDF5_SETTINGS["compression"]
+            is lh5.settings.DEFAULT_HDF5_SETTINGS["compression"]
         )
         assert (
             h5f["/data/struct/wftable_enc/dt"].compression
-            is DEFAULT_HDF5_SETTINGS["compression"]
+            is lh5.settings.DEFAULT_HDF5_SETTINGS["compression"]
         )
 
 

{legend_pydataobj-1.11.12 → legend_pydataobj-1.11.13}/tests/lh5/test_lh5_write.py

@@ -15,6 +15,30 @@ def test_write_compressed_lgnd_waveform_table(enc_lgnd_file):
     pass
 
 
+def test_write_with_hdf5_compression_global(tmptestdir):
+    data = types.Table(
+        size=1000,
+        col_dict={
+            "col1": types.Array(np.arange(0, 100, 0.1)),
+            "col2": types.Array(np.random.default_rng().random(1000)),
+        },
+    )
+    outfile = f"{tmptestdir}/write_hdf5_data_global_var.lh5"
+
+    lh5.settings.DEFAULT_HDF5_SETTINGS["shuffle"] = False
+    lh5.settings.DEFAULT_HDF5_SETTINGS["compression"] = "lzf"
+
+    lh5.write(data, "data", outfile, wo_mode="of")
+
+    with h5py.File(outfile) as h5f:
+        assert h5f["/data/col1"].shuffle is False
+        assert h5f["/data/col1"].compression == "lzf"
+
+    lh5.settings.DEFAULT_HDF5_SETTINGS = lh5.settings.default_hdf5_settings()
+    assert lh5.settings.DEFAULT_HDF5_SETTINGS["shuffle"] is True
+    assert lh5.settings.DEFAULT_HDF5_SETTINGS["compression"] == "gzip"
+
+
 def test_write_with_hdf5_compression(lgnd_file, tmptestdir):
     store = lh5.LH5Store()
     wft, n_rows = store.read("/geds/raw/waveform", lgnd_file)