search

freelawproject / recap

This repository is for filing issues on any RECAP-related effort.
https://free.law/recap/
12 stars 4 forks source link

A Large Docket Not Working #303

Closed brianwc closed 2 years ago

brianwc commented 2 years ago

I don't know if this is related to freelawproject/recap#218 or freelawproject/recap#261 but we only get a partial segment of this docket: https://www.courtlistener.com/docket/38277337/united-states-v-pacific-gas-and-electric-company/

On the website we show just entries 1213-1280 (with no pagination to let you go back or forward) but we are definitely getting individual documents outside that range as we have the scathing filing from Judge Alsup yesterday, docket number 1559.

Since this is a docket with journalists interested in it, it would be nice for it to perform better.

flooie commented 2 years ago

@brianwc

This looks like a good news bad news (but mostly good news situation). It looks like we have a duplicate of the case on CL but the one you found is incomplete. You can find the full docket with pagination (including the portion of the docket on the dupe) here.

https://www.courtlistener.com/docket/4180753/united-states-v-pacific-gas-and-electric-company/

flooie commented 2 years ago

@mlissner the duplicate has been inactive for a year and has no PDFs linked to it. I would think it would be safe to simply delete the mistake. Would this be appropriate?

johnhawkinson commented 2 years ago

Guys, this is a criminal case. Therefore it is subject to the doppelganger docket problem freelawproject/courtlistener#2185, as every single criminal case is. There are at least 2 sub-cases to retrieve any document in a criminal case (n+1 where n is the number of defendants), and depending on which you use, some will have some docs and some will have others.

There's something to be said for prioritizing fixing this problem, or maybe adding some mitigation for it since apparently we cannot allocate the resources to fix it properly :(

mlissner commented 2 years ago

Yep, closing this as a dup, unfortunately. I honestly don't understand what it will take to fix this, but if we can break that dam, that'll go a long way. Somewhere we need to update our data model to handle this, but I don't understand the PACER data well enough to understand how the model needs to be updated.

Mitigations welcome too (but also over at freelawproject/courtlistener#2185). It is indeed one of our worse bugs.

flooie commented 2 years ago

Wait, Can't we just do something like this

Docket number and court

https://www.courtlistener.com/api/rest/v3/dockets/?docket_number=3:14-cr-00175&court=cand

and add them as tabs for associated cases if it comes back as > 1

@mlissner @johnhawkinson

johnhawkinson commented 2 years ago

So, it turns out I was a little too summary in my review and disposition. This is even more broken than the usual kind of doppelganger situation, because the CL page Brian cited is associated with a now-invalid pacer_case_id in N.D. Cal. I would imagine this might have happened when the case was transferred from Judge Henderson to Judge Alsup, but maybe not. So in any event, there exist:

https://www.courtlistener.com/docket/38277337/united-states-v-pacific-gas-and-electric-company/ that Brian noted, which links that mostly don't work, but kind of if you push them hard enough https://www.courtlistener.com/docket/4180753/united-states-v-pacific-gas-and-electric-company/ with 9 pages of docket entries for id 276096 https://www.courtlistener.com/docket/4180752/united-states-v-pacific-gas-and-electric-company/ with 8 pages of docket entries for id 276095

Note that this situation where the CL docket id numbers (4180752, 4180753) are adjacent is rare and cannot be relied upon. Whereas the adjacent pacer_case_id numbers (276095, 276096) is the normal situation (but is not always true, e.g. if a new defendant is added after the initial case opening).