💫Fix SP tag, tweak Vectors.__init__, fix Morphology #1442
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Renaming the tag with an underscore lets us add it to the tag map without worrying that we'll change the sequence of tags, which throws off the tag-to-ID mapping. For instance, if we inserted a 'SP' tag, the "VERB" tag is pushed to a different class ID, and the model is all messed up.
Previously the data and width options were one argument in Vectors, which meant you couldn't say vectors = Vectors(strings, width=300). It's better to have two keywords.
ines
changed the title
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Re #1052, #683
This patch mixes a few changes that aren't obviously related, but ended up having inter-dependencies. Not wonderful.
The main motivation was to fix the requirement that each tag map must specify the
SPtag. This tag is used internally by spaCy, so it's weird to make the data provide it. However, previous efforts to add it automatically threw the class ID mapping out for the tagger, becauseSPwas inserted somewhere in the middle of the tag list.To address this, I've renamed the
SPtag to_SP, to denote it's a hard-coded value. This also means it sorts to the end of the ordinary tags.While making this change I came across a bug in the way the
Vectorsclass would be created when there were strings, but no vocabulary items. In fixing this, I also fixed a wart in theVectors.__init__API: instead of having one argumentdata_or_width, there's no two keyword arguments.Finally, making changes to
morphology.pyxled me to an issue that caused very slow compilation of this module. The problem was due to acpdef enumdeclaration inmorphology.pxd, which causes a lot of code to be generated in recent versions of Cython.Clearly I should've made some of these patches on other branches. I'll improve my branch bureaucracy in future.
Types of changes
Checklist: