Fix citation html for cites near page numbers.

[jcushman]

I think this avoids the HTML errors on pages like https://cite.case.law/ark/13/409/ where citation html gets mashed up with page number html:

This would require running sync_case_body_cache for all volumes, which is a big intervention, so TBD if we want to do that. Possibly there's some way to filter to just cases with html display errors if we can find those.

[bensteinberg]

This needs an update to a test fixture, looks like.

[mdellabitta]

FWIW, I did some forensics on the HTML of the smaller set of reporters to find cites that linked to other released reporters with Beautiful Soup and it wasn't too bad. Perhaps we could do something similar where we look for the string "href" in the text nodes of each document?

[jcushman]

Annoyingly, the tests had detected an actual typo in my PR; I fixed the broken line.

@mdellabitta good idea, it looks like there are about 1000 cases with "href" in them:

In [1]: CaseBodyCache.objects.filter(text__contains='href').count()
Out[1]: 907

I'm thinking as the next step we could push this change and try updating just those cases and spot check them to see if it worked.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (1f201a0) 63.62% compared to head (09dbcad) 63.68%.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #2200      +/-   ##
===========================================
+ Coverage    63.62%   63.68%   +0.05%     
===========================================
  Files          107      107              
  Lines        11803    11820      +17     
===========================================
+ Hits          7510     7527      +17     
  Misses        4293     4293              

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