udapi.block.read.conllup module

Conllup is a reader block for the CoNLL-UPlus format.

Columns which don’t have standardize attributes in Udapi/CoNLL-U are stored in MISC (as key=value pairs).

This code has been only tested on Hungarian KorKor files for CorefUD so far. However, in the end, it is not used there (xtsv files are used instead conllup).

class udapi.block.read.conllup.Conllup(attributes='autodetect', save_global_columns=False, **kwargs)

Bases: Conll

A reader of the CoNLL-UPlus files.

parse_comment_line(line, root)

Parse one line of CoNLL-U and fill sent_id, text, newpar, newdoc in root.

parse_node_line(line, root, nodes, parents, mwts)