Lots of failing iquery pages

[mlissner]

Usually Sentry catches this and maybe it did, but I haven't seen an issue. On Friday we released a new version of RECAP that uploads the iquery page. As of this moment, we've gotten about 2k of them (kind of a surprising number!), and of those, about 700 didn't work:

https://www.courtlistener.com/api/rest/v3/recap/?upload_type=12&status!=2

(Status of 2 is "success", so these are ones that have status!=2, a clever trick our API supports.)

Not sure what's up with these, but we should investigate. @albertisfu, I'm going to put this on your queue, but if it winds up being the extension, let's kick it over to @ERosendo.

[albertisfu]

I've checked this, I got some random files from these unsuccessful CASE_QUERY_PAGE uploads.

Seems that the problem is partially related to #598

From my sample, some of the uploads belong to CASE_QUERY_RESULT_PAGE. So once we start getting these uploads using the new upload_type, this might be solved.

Some more are from a different page "Select a Person" (maybe this page also uses the iquery url), so I think the solution that @ERosendo proposed for the extension about not relying on the URL to identify Case Query pages and instead use the HTML might solve the problem.

Here are some screenshots from my sample:

I attach the Select a Person HTML if is useful for Eduardo to identify this page:
person_html.zip

[ERosendo]

@albertisfu I've just created a PR to address this issue related to the CASE_QUERY_PAGE uploads.

[albertisfu]

@ERosendo Great! So with this change in the extension, are also other pages like "Select a Person" going to be ignored, right?

[ERosendo]

@albertisfu Yes, you're right. The extension will ignore pages similar to the "Select a Person" page.

[albertisfu]

Perfect, thanks you!