Skip to content

incorrect french tokenization and tagging for c'est #119

@joprice

Description

@joprice

Language
French

Describe the bug

"C'est" is tagged by catalyst as (token, lemma, pos):

C None NOUN
' None PUNCT
est None AUX

and by spacy using fr_core_news_lg as:

C' ce PRON
est être AUX

Expected behavior

Contracted forms like C' should match the spacy output and common verbs like est should include their lemma.

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions