korpling / annatto

Converts linguistic data formats based on the graphANNIS data model as intermediate representation and can apply consistency tests.
Apache License 2.0
1 stars 0 forks source link

Extend support of CoNLL sentence metadata / annotations #257

Closed MartinKl closed 3 months ago

MartinKl commented 3 months ago

In conllu treebanks there still are "metadata" annotations on sentence level not consisting of key and value separated by a "=", but rather comment style annotations that might, but not necessarily need to be introduced by a key followed by a :.

Sentence annotations that do not start with an annotation key and a "=" should be treated as a bare value (concatenation for multiple lines) that will be stored in a sentence level annotation conll::comment.

MartinKl commented 3 months ago

Once the PR is merged, this will be implemented and available in release 0.10.0.

The annotation qname for sentence comments can be adapted in the config entry of the import by using:

comment_anno = { ns = "comment_namespace", name = "comment_name"}

The attribute defaults to { ns = "conll", name = "comment" }.