Support parsing `short_description` for Bankruptcy Multi Docket NEF

[sentry-io[bot]]

Sentry Issue: COURTLISTENER-6HZ

Not parsing description for Bankruptcy Multi Docket NEF for court 'paeb'

Related to parsing docket_entry.short_description for bankruptcy courts #912 . We have no method for parsing short descriptions when we deal with multi docket NEFs. I wrote a logger.error call for this and now we have some examples.

Filed by @grossir

[sentry-io[bot]]

Another instance of this...

Sentry Issue: COURTLISTENER-6Q8

Filed by @mlissner

[sentry-io[bot]]

And another...

Sentry Issue: COURTLISTENER-6Q7

[sentry-io[bot]]

Sentry Issue: COURTLISTENER-6QV

[sentry-io[bot]]

Sentry Issue: COURTLISTENER-71E

For mdb Linked by @grossir

[grossir]

Taking advantage of the fact that we can upload images, I will document the problem with the multi dockets here Usually, we extract the short_description from a NEF by splitting the email Subject using the docket's case name and number in various combinations. However, when we have more than 1 docket, we may have to try out different splits:

Uses 2nd docket's case name and docket number:

Uses 1st docket's number, but a partial case name

Uses 1st docket's number and case name

Uses 2nd docket's number, and partial case name

[mlissner]

Looks like this might not be possible in a reliable way or like it just might not be worth it.

[grossir]

I think I found a good enough way, which is basically the old parsing inside a for loop and some extra checks. Doesn't break anything (from the tests). Can you review it? maybe it is good enough

[grossir]

@mlissner I created a new PR to address only multi dockets on njb.

Luckily, njb NEFs follow an easy format that only uses the case_name to split the short description from the subject

https://github.com/freelawproject/juriscraper/pull/1032

[sentry-io[bot]]

Sentry Issue: COURTLISTENER-7GB