[lat] Broken selector #514
Labels
Comments
kylebgorman
added
bug
This was referenced Feb 15, 2024
kylebgorman
added a commit
to kylebgorman/wikipron
that referenced
this issue
Feb 21, 2024
kylebgorman
added a commit
that referenced
this issue
Feb 21, 2024
* [mlt] Updates Maltese phonelist. Due either to bugs or changes in the upstream data, I noticed there was a very high rate of filtration on Maltese. It seems that [u] was not included, nor was one of the affricates. There are still some filtration for "archaic" pronunciations of [ɣ] for <għ>, which is WAI. * Changelog * Adds Python 3.12 support * project classifers * tests on CircleCI * Revert "Adds Python 3.12 support" This reverts commit e72bc3d. * Pauses Latin testing in lieu of #514. * Fixes typo in test_split * More explicit comments. * changelog * Reruns black * black
kylebgorman
added a commit
that referenced
this issue
Feb 21, 2024
* [mlt] Updates Maltese phonelist. Due either to bugs or changes in the upstream data, I noticed there was a very high rate of filtration on Maltese. It seems that [u] was not included, nor was one of the affricates. There are still some filtration for "archaic" pronunciations of [ɣ] for <għ>, which is WAI. * Changelog * Adds Python 3.12 support * project classifers * tests on CircleCI * Revert "Adds Python 3.12 support" This reverts commit e72bc3d. * Pauses Latin testing in lieu of #514. * Fixes typo in test_split * More explicit comments. * changelog * Reruns black * black * Adds Python 3.12 support * Changelog * updates flake8
|
Testing of Latin is (hackily) paused in #520. |
This was referenced May 1, 2024
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
As of at least #509 the custom selector for Latin has been broken.
Latin has a custom selector because the headwords lack macrons. Now the Romans of course didn't use macrons (and they did not consistently indicate vowel length) but just about every modern-era pedagogical resource does, so this was frankly a bizarre decision by the editors, and requires us to find the macronized forms somewhere else on the page (namely in the etymology subsection), then merge these together with the pronunciations. To debug, the obvious thing to do is to mock up one or the other stream: the etymological macronized headwords, or the pronunciations, and see which one isn't matching.
It should be possible to push through with a new version of #509 and just ignore Latin, which is probably a good idea given that it's been a while since a big scrape has been merged.
The text was updated successfully, but these errors were encountered: