Releases: mortonne/wikivector
Releases · mortonne/wikivector
wikivector 1.2.1
This version switches to a modern build system and adds a utility for stripping remaining tags that may be left by WikiExtractor for older Wikipedia dumps.
- The build system is defined by the
pyproject.toml
file. - Setup options are now mostly set by
setup.cfg
. - The
wiki_remove_tags
script can be used to strip out some common remaining tags from extracted Wikipedia text.
wikivector 1.2.0
This version fixes a bug and adds a script to help identify missing articles.
- Fixes a problem that can occur if creating a header file using an absolute path or from outside the text directory. Header files now save paths relative to the extracted Wikipedia text directory to avoid this problem.
- Adds a script,
wiki_check_map
, to make it easier to check for missing articles in a header file.
wikivector 1.1.0
This version fixes an issue when loading vectors from a file using h5py 3. Installation now requires h5py>=3.
wikivector 1.0.0
This version adds support for Zenodo archiving.
wikivector 0.1.1
Fixes an issue with embed_articles where vectors were not saved correctly.