Handle libretexts.org glossary rewriting

[benoit74]

Fix #53

While a bit basic (as mentioned in the issue, we do not handle all properties and we do not handle nested glossary definitions), this is sufficient to provide the sheer added value of glossary pages.

[codecov]

Codecov Report

Attention: Patch coverage is 22.22222% with 21 lines in your changes missing coverage. Please review.

Project coverage is 44.43%. Comparing base (f0e6cfe) to head (477304f).
Report is 4 commits behind head on main.

Files with missing lines	Patch %	Lines
scraper/src/mindtouch2zim/libretexts/glossary.py	23.80%	16 Missing ⚠️
scraper/src/mindtouch2zim/processor.py	16.66%	5 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #66      +/-   ##
==========================================
- Coverage   45.08%   44.43%   -0.66%     
==========================================
  Files          13       14       +1     
  Lines         845      871      +26     
  Branches      111      116       +5     
==========================================
+ Hits          381      387       +6     
- Misses        451      471      +20     
  Partials       13       13              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[rgaudin]

LGTM but I dont see in the ticket nor in the PR why this is limited to English. Should probably be mentioned in the README and should it be a current shortcut, probably have a ticket to support other languages later

[benoit74]

You are right, the fact this is limited to English is a shortcut I took without even really noticing it. I checked the list of ZIMs we have to build and there is some international libraries, so we should support it.

I changed the way it works by:

• detecting we are on a glossary based on path pattern instead of title : /zz:_*/20:_* (seems to be stable across all libraries even in foreign languages)
• ignoring all tables on the page but the last one (seems to be stable as well)
• rewriting with regular logic when there is no table (happens a lot in the wild in fact)

Tested on:

• https://query.libretexts.org/Idioma_Portugues/Introdu%C3%A7%C3%A3o_%C3%A0_pol%C3%ADtica_e_governo_comparados_(Bozonelos_et_al.)/zz%3A_Mat%C3%A9ria_traseira/20%3A_%C3%8Dndice (which does not use the template)
• https://query.libretexts.org/Kiswahili/Akizungumza_Kutumia_Muhimu_Kufikiri_(Marteney)/zz%3A_Nyuma_jambo/20%3A_Kamusi (which does not properly render online, JS is not activated)
• https://query.libretexts.org/Francais/Livre_%3A_Microeconomics_(OpenStax)/zz%3A_Mati%C3%A8re_dorsale/20%3A_Lexique (which does not display the glossary online because it has been set in example table instead of creating a new table, but I think it is way better to display the hidden one)
• https://geo.libretexts.org/Courses/Sierra_College/Physical_Geology_(Sierra_College_Edition)/zz%3A_Back_Matter/20%3A_Glossary

[rgaudin]

Great ; please re-request when ready

[benoit74]

Ready, first commit is unchanged (just rebased), please look only at second commit for last changes

[rgaudin]

👍 please add suggested comment