Skip to content

Releases: explosion/spacy-stanza

v1.0.4

09 Oct 08:43
b57f348
Compare
Choose a tag to compare
  • Extend stanza support to stanza v1.6.

v1.0.3

14 Mar 08:22
1361747
Compare
Choose a tag to compare
  • Extend stanza support to stanza v1.5.

v1.0.2

27 May 14:44
7bff03f
Compare
Choose a tag to compare
  • Extend stanza support to stanza v1.4.

v1.0.1

09 Nov 12:41
Compare
Choose a tag to compare
  • Extend stanza support to stanza v1.3.
  • Modify fallback for token.tag if XPOS is missing from FEATS to UPOS. (Thanks @BramVanroy!)

v1.0.0: Update for spaCy v3

04 Mar 14:52
a87c723
Compare
Choose a tag to compare

This release requires spaCy v3. See the updated usage documentation for more details: https://github.com/explosion/spacy-stanza/tree/v1.0.0#-usage--examples

✨ New features and improvements

  • Refactor library for spaCy v3.0
  • Config system integration
  • New API for loading a Stanza model following Stanza's Pipeline API
  • Include morphological analysis under token.morph

🔴 Bug fixes

  • Fix issue #59: Fix trailing whitespace token handling

v0.2.5: Bug fixes for trailing whitespace and entry points

10 Mar 14:54
2128d5c
Compare
Choose a tag to compare

🔴 Bug fixes

  • Fix issue #56: Add entry points for all supported languages
  • Fix issue #59: Fix trailing whitespace token handling

v0.2.4

10 Oct 11:17
71a3483
Compare
Choose a tag to compare
v0.2.4 Pre-release
Pre-release

Fix issue #43: fix alignment when tokenizing with spaCy
Fix issue #47: update dependencies for compatibility with stanza v1.1

v0.2.3

26 Jun 08:55
Compare
Choose a tag to compare
v0.2.3 Pre-release
Pre-release
  • Fix issue #30, #33: Rewrite alignment to preserve whitespace tokens.

v0.2.2

25 Jun 09:28
Compare
Choose a tag to compare
v0.2.2 Pre-release
Pre-release
  • Improve handling of texts with misaligned tokens.
  • Add the ability to use token.feats as tags.
  • Get language attribute directly from Stanza pipeline object.

Thanks to @yuhaozhang and @SergeyShk for the pull requests.

v0.2.1

19 Mar 10:28
Compare
Choose a tag to compare
v0.2.1 Pre-release
Pre-release
  • Add warning and don't populate doc.ents if entities returned by the Stanza NER model don't map to tokens produced by the Stanza tokenizer.