[python-hdf5storage] annotated tag 0.1.2 created (now 6c78963)
Ghislain Vaillant
ghisvail-guest at moszumanska.debian.org
Mon Feb 29 08:25:29 UTC 2016
This is an automated email from the git hooks/post-receive script.
ghisvail-guest pushed a change to annotated tag 0.1.2
in repository python-hdf5storage.
at 6c78963 (tag)
tagging 03f0a58af5ad115546e5df3da39df11ad217bbc0 (commit)
replaces 0.1.1
tagged by Freja Nordsiek
on Thu Aug 14 18:50:04 2014 -0400
- Log -----------------------------------------------------------------
bugfix release
Freja Nordsiek (18):
Made it so that MATLAB compatible userblock is put in HDF5 file before data is written as opposed to after (fixes bug where the file becomes not MATLAB compatible if an exception occurs).
Added MATLAB_fields Attribute support to the dict marshaller.
Added MATLAB_fields Attribute write support to the NumpyScalarArrayMarshaller.
Fixed bug where Attribute MATLAB_fields is not deleted for non-structured ndarrays.
Added MATLAB_fields Attribute read support for empty structure arrays to the NumpyScalarArrayMarshaller.
Fixed bug in MATLAB_fields Attribute generation for NumpyScalarArrayMarshaller.
Fixed bug in MATLAB_fields Attribute generation for NumpyScalarArrayMarshaller.
Fixed bug in reading empty MATLAB structures.
Fixed bug in reading empty MATLAB structures.
Fixed bug in reading empty MATLAB structures.
Fixed writing of empty matlab structs to use a Dataset with the appropriate fields attributes set and several reading errors with the new format.
Made it so that the MATLAB_fields Attribute is properly used on structures to order the fields in the right order.
Fixed bug where structured numpy.ndarrays with a field having an 'object' type is written as an HDF5 COMPOUND when the structured_numpy_ndarray_as_struct option is disabled. Is now written as a GROUP instead as if the option was set.
Added check for non-unicode keys in PythonDictMarshaller right to the beginning of the writing process so that nothing is changed if there is a non-unicode key.
Removed numpy.float16 support for when the h5py version is less than 2.2 since that was when support for it was introduced.
Updated documentation to reflect code changes and new version.
Incremented package version to 0.1.2
Made it so that ez_setup.py is for sure included in the manifest.
-----------------------------------------------------------------------
This annotated tag includes the following new commits:
new 8167dd7 Initial commit
new 3e70816 Added some needed files including license.
new e6d9c0c Initial set of python codes to write only (still need polishing)
new d22803c Changed line endings to be native.
new 73450df Added missing import.
new 8fd3b65 Fixed method call typo.
new 2ffa814 Fixed error in detection of whether an overwrite is needed based on shape and type information.
new bbb0482 Fixed bug in writting MATLAB_int_decode attribute for strings.
new 94eb039 Fleshed out core.Options (properties and docstrings) and added core module docstring.
new e042e6f Fleshed out the rest of Marshallers.TypeMarshaller to support reading.
new e779992 Added docstrings to Marshallers.TypeMarshaller.
new 69f745a Added minimal docstring for Marshallers module.
new 6ba7878 Small, mostly docstring fixes in core.Options.
new 177f9e6 More docstring work and adding a method in the core module.
new d3a5afa API changes in preparation for read support, especially in regard to MATLAB types.
new 00d15ff Added a function to get an Attribute.
new 7ebe9aa Added low level, somewhat buggy for strings, read support for Numpy types.
new 33124c8 Added utility function to encode complex types to be written to HDF5 file with the right complex field names.
new 254420b Changed writing of complex types to set complex field names manually.
new 7631790 Fixed but in complex type encoding where type was not kept constant.
new 3f75a8d Made utilities.decode_complex more general (can set one or both field names manually).
new 7b8c734 Added read support for Python scaler types and None.
new 649b3d8 Changed the way marshallers for container objects recursively write their elements.
new 6b008c4 Removed partial code for the currently impossible to set MATLAB_fields attribute.
new d507cd7 Added read support for dicts.
new 2089de7 Added utility to convert strings to str, and changed other string conversion utilities to use it.
new 553f044 Added Python string read support and changed it so bytearrays are written as numpy.string_ instead of numpy.uint8.
new f04e283 Added high level read function.
new 1a72e51 Fixed reading bugs.
new 844b751 Fixed bugs with reading/writing scalar types and reading complex types.
new 71a2c88 Fixed bug on writing empty datasets so that Matlab can read them.
new 3b2e602 Fixed bugs in writing boolean types in a Matlab compatible way.
new e497e0f Changed reading to have the Matlab compatiblity option set to False by default.
new 1dd5482 Removed quad support as h5py doesn't support them.
new 54e13e8 Added package installer (distutils) and wrote Readme.
new ef8501e Escaped underscores in README.rst.
new e3e2889 Added what Python types are converted to in numpy in README.rst
new a43da17 Changed the way that None is written.
new d997d00 Added and fixed documentation.
new b401071 Added initial Sphinx based documentation.
new 9209376 Added ability to pass an Options object to the high level functions directly.
new a7aa29d Changed default names to read and write data to in the high level functions.
new 374e62f Gathered all options keywords in high level read/write functions into a **keywords.
new 9e75476 Fixed error in README.rst about package functionality.
new 4d24528 Changed calling arguments for core.read and core.write and added their doc strings.
new 7b12ec5 Added Sphixnx API documentation structure for Marshallers.
new 28ed1f0 Fixed documentation and changed references to the aliases np.string_ and np.unicode to their targets np.bytes_ and np.str_.
new d2f0a32 Changed manifest to include documentation.
new 1284b76 Moved items in core module to main module (__init__.py), added examples, and fixed Sphinx based documentation.
new 56cdd73 Added initial tests.
new cee850c Added more tests for write and readback when not writing MATLAB metadata.
new 2c8716e Added tests to the manifest.
new 0a8db47 Fixed bugs with writing and reading empty strings, in addition to a small file closing bug.
new 1083c70 Added read/write support for numpy.object_ types (become HDF5 references).
new bc9c847 Added read/write support for lists.
new 9d8a8e9 Added read/write support for tuple, set, frozenset, and collections.deque.
new 5bc85e0 Dropped the C in all the CPython parts of attribute names.
new bb8b912 Added write readback tests for the numerical numpy types (not strings).
new 927e5b2 Fixed issue in test where NaN's were causing equality tests to fail.
new 0923f91 Fixed bug of scalar numpy complex types being turned into arrays when doing write readback.
new 25682d3 Changed option name from MATLAB_comaptible to matlab_compatible.
new 9e4bc8c Fixed name of the PyTables package and group.
new 62b39fd Added file truncation options for writing.
new 452ba79 Fixed bugs in numpy bools not being converted back to book from uint8 correctly.
new 3246897 Added numpy bool read writeback tests.
new 54d2b0f Fixed a file clean up bug on error.
new 38dd35d Changed tests over to nose.
new d726cab Fixed and rearranged tests.
new 988b052 Added tests for numpy bytes_ and str_ types.
new 6b2dbc3 Fixed bugs in preserving the shape when reading back arrays of strings (all kinds).
new 5c9b649 Switched from distutils to setuptools for packaging so unit tests could be run from setup.py.
new b078faa Changes to setup.py for using setuptools (forgot to add in the last commit).
new f261dbf Added write readback tests for numpy.object_ type.
new a0a9730 Fixed a bug with deleting the H5PATH attribute on numpy.object_ arrays when not doint matlab compatibility.
new 7fc2be1 Split previously 2d array write readback test into 1d, 2d, and 3d cases.
new 1b3f53e Fixed bug in decoding 3d and higher dimensionality arrays of numpy string types when encoded as uints.
new 9e7b15a Added write readback unit tests for list, tuple, set, frozenset, and collection.deque.
new 6d5a8e7 Consolidated the assertion functions for python and numpy types together.
new bbb8638 Rearranged write readback test code and added some comments.
new cadb9eb Fixed string encoding and decoding bugs.
new d46aae8 Added write readback tests to file with no python type information stored (either no information or just matlab information).
new 126e8e4 Added tests for bug where a complex number with one field as NaN but the other not is read back as NaN in both fields.
new 470bd87 Parameterized the sizes of the random datasets used in the tests.
new 349b676 Renamed the convert_strings_to_utf16 option to convert_numpy_bytes_to_utf16.
new ec11484 Added an option to optionally convert numpy.str_ to np.uint16 in UTF-16 format and changed it so that str is converted to numpy.str_ instead of numpy.bytes_.
new 7526cdc Renamed decoders for numpy bytes_ and str_ in utilities.
new 192f430 Updated documentation on the utilities module.
new bf3702e Added custom assert functions for the tests into their own module.
new 048e24a Moved the none format (not type or matlab information) and the matlab format assertion functions to tests/asserts.py as well as consolidating the assertions for different types together and adding in assertial code for dictionaries.
new 3cb8514 Added write readback tests for dicts.
new 73c79d1 Adjuste appearance of the first table in the README.
new 4c75ef1 Added oned_as option to control whether 1D arrays become row or column vectors when converted to 2D.
new e7ca164 Renamed the convert_scalars_to_arrays option to make_atleast_2d.
new bc4f382 Added a savemat function modelled after scipy.io.savemat.
new 47f2cf4 Added a loadmat function modelled after scipy.io.loadmat.
new a6472a2 Added existing file truncation options to savemat.
new b5a0169 Added main information page to documentation (replicate of README.rst) and adjusted titles.
new bb295d3 Created a skeleton introduction in the documentation and then moved the examples from __init__.py to it and updated them.
new 444dc47 Fixed error in README where the other numpy integer types were not in the table of supported types.
new 76ab931 Reorganized table in README.
new 3fe9b62 Renamed name arguments in read and write functions to path.
new 12a6eec Finished rest of introduction.rst in the documentation.
new 1b70ac1 Added documentation on the storage format.
new cd4309a Fixed bugs where the name keyword argument was not updated to path in savemat and loadmat.
new 5bd9cfe Renamed the store_type_information option to store_python_metadata.
new 6924018 Added more documentation regarding the matlab_compatible option and how types are read back from MAT files.
new a0f1a5b Added licence header to test files.
new 85cd471 Added more information to setup.py
new 9b5c74c Added numpy.chararray support.
new bc4bab8 Added numpy.void support.
new 742f1eb Added ability to discard or throw an error when writing a matlab incompatible type when doing matlab compatibility.
new 10067f0 Fixed bugs in tests resulting from the changes in how matlab incompatible types are written.
new 82774db Fixed bugs in documentation.
new 760d2e7 Documentation and docstring updates and fixes.
new f381b8d Added write support for numpy.ndarrays that have fields in their dtype as Groups/matlab structures.
new de1364b Added the use of the canonical empty inside the group for references.
new 3b355b6 Updated documentation.
new 9194e81 Fixed missing information about exceptions in docstrings.
new 40c4b68 Moved the code to read an h5py.Reference array and make a numpy.object_ array from it to a function.
new 84dc2ff Added read support for structured numpy.ndarrays as they are, as opposed to reading them as dicts.
new 9608d03 Removed commented out code.
new 05aac0c Made it so that structured numpy.ndarrays with one element have their fields written as is as opposed to packing them into HDF5 Reference arrays, and made it so they can be read back correctly.
new 7e261a8 Changed the default marshaller for HDF5 Groups to be NumpyScalarArrayMarshaller as opposed to PythonDictMarshaller.
new 39a97c7 Renamed the fielded_numpy_ndarray_as_struct option to structured_numpy_ndarray_as_struct.
new 917d99b Fixed tests to work with how dicts are read back as structured np.ndarrays if python metadata isn't stored.
new cf753c7 Updated documentation with regards to reading/writing structured numpy.ndarrays and fixed errors.
new 8df8175 Changed the setting of string array attributes to use str instead of np.bytes_ (unicode instead of ASCII).
new 18210e6 Renamed the Attribute Python.numpy.Fields to Python.Fields and made it so that if it wasn't given, fields are sorted when making a structured ndarray.
new 03c0586 Updated documentation.
new 7cebd72 Added read/write numpy.recarray support.
new 2c059a0 Fixed a bug where a dict with no members was marked with metadata as a zero sized struct, instead of one with no fields.
new b585cc0 Added/changed code to handle empty structured ndarrays.
new e3fa8a1 Renamed all the different string converters in utilities from starting with decode_to to convert_to
new 7c4d561 Added license info in README.rst
new 8145a42 Added a development information page in the documentation.
new 6b71a13 Moved from using list.copy to copy.deepcopy(list) for Python 2 compatibility.
new f64d310 Added Python 2.7 support.
new 1e6a69e Replaced dictionary comprehensions to make code Python 2.6 compatible.
new 3166122 Changed all sets to be constructed through set([...]) instead of {...}
new f5b59b5 Fixed bug on Python 2.6 where sys.version_info is just a tuple unlike in later Python versions.
new 7aa9e8a Fixed bug on Python 2.6 where byterray.fromhex was given a non-unicode argument.
new c0eaddf Fixed bug on Python 2.6 where the encoding to bytes.decode needs to be given as the first positional argument as opposed to as a keyword argument.
new 90b4296 Formally added Python 2.6 support in the README.rst and the installer.
new 20b5798 Fixed bug where loadmat was not returning a dict of the variables and not putting it in mdict if it was given.
new b338a43 Added tests for reading data written by MATLAB.
new 9418e50 Fixed bugs in installer.
new 0c5e8c6 Added additional tests to take the data read from MATLAB, send it back, read it back, and compare.
new 77b9800 Renamed MATLAB compatibility tests.
new 0a74e01 Fixed bug where subprocess.check_output is not available on Python 2.6.
new c9f675f Modified MATLAB test code to reduce output.
new 95fa1ea Last changes to README.rst before releasing package.
new 8fe2b6c Changed description.
new c62b3c3 Fixed bug where str was converted to numpy.bytes_ instead of numpy.str_ before writing.
new dc3657b Changed tests to work after fixing bug where str was written like numpy.bytes_ instead of like numpy.str_ and added test for writing and reading back unicode characters.
new 0189bcd Fixed bug when reading complex numbers where one part is NaN but the other is not.
new 4fd8430 Modified tests to work with real/complex numbers that have NaN part/s.
new 1762f64 Fixed string conversion bug in Python 2.
new 946f513 Fixed bug where MATLAB structures could not be read back correctly in Python 2.
new 3037724 Fixed bug in Python2 where str.decode does not use a keyword for the encoding.
new 756f869 Removed replace keyword in str.decode for Python 2.6
new 94f10e2 Changed a reference to numpy.str_ to numpy.unicode_ so that they will be written as numpy.uin16 if possible on Python 2.
new 031d234 Fixed bug in tests.
new 286d6ea Incremented version to 0.1.1
new e2ef6f3 Fixed version in documentation.
new 9a271e8 Made it so that MATLAB compatible userblock is put in HDF5 file before data is written as opposed to after (fixes bug where the file becomes not MATLAB compatible if an exception occurs).
new 5a1bf8a Added MATLAB_fields Attribute support to the dict marshaller.
new 29fcef7 Added MATLAB_fields Attribute write support to the NumpyScalarArrayMarshaller.
new 72d97a1 Fixed bug where Attribute MATLAB_fields is not deleted for non-structured ndarrays.
new 9118a9a Added MATLAB_fields Attribute read support for empty structure arrays to the NumpyScalarArrayMarshaller.
new 6e31b58 Fixed bug in MATLAB_fields Attribute generation for NumpyScalarArrayMarshaller.
new 1973e2b Fixed bug in MATLAB_fields Attribute generation for NumpyScalarArrayMarshaller.
new 713276d Fixed bug in reading empty MATLAB structures.
new 5ff55a5 Fixed bug in reading empty MATLAB structures.
new 1a1f2d7 Fixed bug in reading empty MATLAB structures.
new 99d573c Fixed writing of empty matlab structs to use a Dataset with the appropriate fields attributes set and several reading errors with the new format.
new 41731cf Made it so that the MATLAB_fields Attribute is properly used on structures to order the fields in the right order.
new d6c1a7d Fixed bug where structured numpy.ndarrays with a field having an 'object' type is written as an HDF5 COMPOUND when the structured_numpy_ndarray_as_struct option is disabled. Is now written as a GROUP instead as if the option was set.
new cc763ff Added check for non-unicode keys in PythonDictMarshaller right to the beginning of the writing process so that nothing is changed if there is a non-unicode key.
new 392909a Removed numpy.float16 support for when the h5py version is less than 2.2 since that was when support for it was introduced.
new a15966c Updated documentation to reflect code changes and new version.
new 017d36c Incremented package version to 0.1.2
new 03f0a58 Made it so that ez_setup.py is for sure included in the manifest.
The 182 revisions listed above as "new" are entirely new to this
repository and will be described in separate emails. The revisions
listed as "adds" were already present in the repository and have only
been added to this reference.
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/debian-science/packages/python-hdf5storage.git
More information about the debian-science-commits
mailing list