Releases: G-Node/nixpy
Releases · G-Node/nixpy
Version 1.4.8
Changes
- Support for data compression: When creating a file, Block, or DataArray, compression can be enabled using the
compression
argument with the valueDeflateNormal
. - Metadata printing:
pprint()
method for Sections and Properties prints the metadata tree in human readable form.
Bugfixes
- Out-of-bounds indexing for data is caught early, making the source of the exception easier to find and the message more useful.
- Other minor bug fixes (see relevant PRs)
Relevant PRs
- #302 Data compression
- #303 Catch out of bounds indexing errors in datasets
- #304 Slice data by index or dimensional data
- #309 Add metadata printing functionality
- #329 Move current_depth to end of param list
- #339 Tiny fixes to SampledDimension util functions
- #343 Bindings fixes (for indexOf overloads)
- #344 Change collections ABC imports to new module
Version 1.5.0 beta 2
Bug fixes and minor improvements
- Metadata Sections can now be created without specifying a
type
parameter. This will create the Section with the default type of"undefined"
. This is the same behaviour that is found in odML. - Fixed bugs in the way the Compression flag is handled and inherited from File to Block to DataArray.
Version 1.5.0 beta 1
Beta version of 1.5.0
Major changes
- Bindings to the C++ NIX library have been removed
- Major performance improvements
- Compatibility with odML v1.4 file format
Version 1.4.7
Changes
- Files can now be opened using the
with
statement for context management.
Relevant PRs
- #316: Open files using context manager
Version 1.4.6
Changes
- New feature: Pretty printing of metadata tree through
Section.pprint()
andProperty.pprint()
.
Relevant PRs
- #309: Add metadata printing functionality.
Version 1.4.5
Changes
- Allow non-SI units in Tags and MultiTags: This was already supported in DataArray and Property objects. The same rules and restrictions now apply to Tags and MultiTags. Any string can be set as a unit, but non-SI units will disable the ability to retrieve data through the (Multi)Tag.
Bugfixes
- Fixed an indexing and slicing bug in the h5py backend.
Relevant PRs
Version 1.4.4
Changes
- DataArrays can now be indexed using any numerical type (e.g., numpy scalars), not just
int
orslice
.
Relevant PRs
Version 1.4.3
Optionally requires NIX libraries with version 1.4.x.
Changes
- Feature and Reference retrieval
- Tag and MultiTag Features can now be retrieved by DataArray ID or name
- Tag
retrieve_data
function can now be called with reference ID or name
- Arbitrary unit strings
- Property units can now be set to any arbitrary string
- DataArray units can now be set to any arbitrary string, unless the DataArray has an AliasRangeDimension
- Empty strings (after removing spaces) in unit attributes are set as
None
.
Relevant PRs
Version 1.4.2
Optionally requires NIX libraries with version 1.4.x.
Changes:
- Fixed bug where integer data arrays with polynomial coefficients would result in truncated data values after applying polynomial.
Stable 1.4.1
Optionally requires NIX libraries with version 1.4.x.
Changes:
- Fix bug in the creation and loading of AliasRangeDimensions when using the 'h5py' backend.
- Added flush() method to File.