List view
Also see #3052. We didn't want to cram this all into the initial v2.1 release, since it'd introduce another layer of complexity. But now that the new models are out, we can focus on cleaning up the vocab and vectors, and resolving various issues around the model data.
No due date•2/2 issues closed- No due date•3/3 issues closed
There are a few issues that may indicate bugs in the `Parser` transition system, including how spaces and punctuation are attached and how sentences are broken.
No due date•7/7 issues closedThe current `Lemmatizer` currently only works for English and assumes WordNet-formatted input files. The `Morphology` class does not currently add attributes correctly. We also need a simple lookup-based `Lemmatizer` class.
No due date•8/8 issues closedFix and extend NER, tagging and dependency parsing training examples. Improve model saving and loading process, especially for vocabulary.
No due date•3/3 issues closedAs more languages are added, we need a better way to share information between them. For instance, most languages will need the same set of emoticons, some abbreviations etc. We also need to refactor the `Defaults` class, which provides access to the data to the `Language` class.
No due date•5/5 issues closedWrite usage workflows for new and improved docs.
No due date•14/14 issues closedBug fixes, enhancements, documentation for spaCy 1.0.
Due by October 19, 2016•28/28 issues closed