match file version 1.0.0 voice parsing bug

[sildater]

in the newest match file version voice information is stored as scoreattribute number with a prefix "v". this is not correctly parsed in load_match, where the voice information is cast to integers. The loading doesn't break but all voice information is lost.

[sildater]

I'll add it here: the match files for chopin excerpts in 4x22 contain note IDs without dashes, e.g., "n1", the others contain repeats and are identified with dashes, e.g., "n1-1". It would probably be good to have this consistent and append a suffix for the ones without repeats.

[manoskary]

Is this linked issue to any PR? Should it be scheduled for Release 1.3.2 ?

[sildater]

@huispaty are you looking into this for the next release?

some more things I just checked:

• voice information as part of the ScoreAttributesList is not documented in the match specification: https://cpjku.github.io/matchfile/
• I think the problem stems from number_pattern = re.compile(r"\d+") and

            note_attributes["voice"] = next(
                (int(a) for a in note.ScoreAttributesList if number_pattern.match(a)),
                None,
            )

in importmatch.py

• I'm not sure whether this regex for voice number is used anywhere else
• I'm not sure whether earlier match files encoded voice information beyond the "s" tag.
• my other comment about consistency in 4x22 might be moot because of the unfolding required foir Schubert and Mozart, just ignore it

[huispaty]

Hello! Yes I pushed a fix for this a few days ago into branch match_import -- I'll check again if number_pattern is used somewhere else and adapt my fix if necessary.
There was another fix for importing match (trill notes parsing for match version v0) -- when both are fixed I'll create a PR and link you!