Skip to content

Enforcing POS comes first #5

@yuvalpinter

Description

@yuvalpinter

Hi!

As far as I can tell, the UniMorph convention for attribute ordering is that the POS is the first entry before the first ;. Nothing in the code I could see enforces this, and indeed when operating on a conllu file a lot of the POSs appear in random locations of the string (when it appears at all), making the tag difficult to process downstream. Can this be fixed?

Sample Input (en_gum treebank, l.152):
9 years year NOUN NNS Number=Plur 7 nmod 7:nmod:to Entity=23)|SpaceAfter=No

Output:
9 years year NOUN NNS PL;N 7 nmod 7:nmod:to Entity=23)|SpaceAfter=No

Thanks,

  • Yuval

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions