When running check_versification() in extract_corpora on a project with a custom versification, the error message can be misleading. In a recent example reported by Michael, it logged:
WARNING - Project versification setting 0 does not match detected versification(s) 4. The detected versification(s) were based on Daniel 3:30, Acts 19:41 being the last verse of their respective chapters.
even though the base versification of the project was versification 1. This is because custom versifications always have a type 0 (Unknown).
We should:
- Consider whether this error message even makes sense with custom versifications - i.e., since it is custom, it wouldn't be unreasonable that it looks like the wrong versification.
- If we do keep this for custom versifications, we should list the
versification.base_versification.type in the error message rather than the versification.type for custom versifications. In this example, that would mean logging 1 rather than 0 in the error message.
- We should use the versification name rather than the versification type (or perhaps in addition to it). This is generally more interpretable and also would be more helpful for describing custom versifications. For example,
versification.name for a custom versification yields the <base-versification-name>-<project-guid> - something like English-1234567890abcdef. If we do this, we may not need to do item 2.
When running
check_versification()inextract_corporaon a project with a custom versification, the error message can be misleading. In a recent example reported by Michael, it logged:even though the base versification of the project was versification 1. This is because custom versifications always have a type 0 (Unknown).
We should:
versification.base_versification.typein the error message rather than theversification.typefor custom versifications. In this example, that would mean logging 1 rather than 0 in the error message.versification.namefor a custom versification yields the<base-versification-name>-<project-guid>- something likeEnglish-1234567890abcdef. If we do this, we may not need to do item 2.