Open ghedlund opened 3 years ago
Related to #164
Wouldn't the second option above create issues in Phones queries (for example a query on initial consonants)? Could we instead flag all cases of unaligned pauses, as these should truly trickle down across data tiers? (i.e. Orthography, IPA Target, IPA Actual, and all other tiers that may be queried for Aligned Words data) Session Check is the best place we have to 'enforce' our coding standards, so I see this as an opportunity.
During the query: we need to make a decision which makes sense for the alignment of a pause to another IPA transcript. The job of the query is not to check data.
Session checks are a different topic - the issue we have already discussed about word alignment checks will cover your concerns.
Updated original comment
Pauses are not included with alignment data so Phon reports the result as '(.) <-> null'.
Modify the 'getAlignedPhones' method of the Phones query so that pauses are treated as a special case and are aligned with:
Note - this isn't a change to how Phon views the data but rather a fix to how the aligned phones come back as null on the IPA Actual side of the results when it shouldn't.
(edited to fix how alignment to another word should work)