freelawproject / courtlistener

A fully-searchable and accessible archive of court data including growing repositories of opinions, oral arguments, judges, judicial financial records, and federal filings.
https://www.courtlistener.com
Other
544 stars 150 forks source link

Parties API Returning Attorneys and Party Types for Other Cases #3858

Open a-drumm opened 8 months ago

a-drumm commented 8 months ago

Examples:

https://www.courtlistener.com/api/rest/v3/parties/?docket=4508405 https://www.courtlistener.com/api/rest/v3/parties/?docket=6283353 https://www.courtlistener.com/api/rest/v3/parties/?docket=6280169 https://www.courtlistener.com/api/rest/v3/parties/?docket=4263779

Several are the City of Chicago, and the other attorney/party_types are from other cases where the City of Chicago is also a party. Some though have attorney/party_types for cases that do not appear to even be the same party.

Only one of these I was able to find the page for on courtlistener.com (sans API) and none of those extra attorney/party_types show up.

I don't know the system well enough to have any idea what is going on, but it does not seem intentional.

mlissner commented 8 months ago

I need a bit more to work with here, @a-drumm. You can find the docket page for any of these by doing:

https://www.courtlistener.com/docket/$id/parties/anything-you-want/

Can you give me a specific example from one of the cases where you see a party in the API that shouldn't be there?

a-drumm commented 8 months ago

Right sorry. I couldn't find some of the docket pages through searching the case name + district court. Shows up if I go right to the page. Thank you for that direction.

Moving on to the actual issue:

I apologize, I was being lazy and I came at this the wrong way before.

Specific example:

I ran a query for parties in Rodriguez v. Chicago: https://www.courtlistener.com/api/rest/v3/parties/?docket=4263779

The 13th result is “Chad Smith” and gives the resource uri https://www.courtlistener.com/api/rest/v3/parties/68837/

That party id has party types/associated attorneys for 7 different cases (type of party and court location are provided for indication that these are almost certainly not the same person):

It appears like the same party ID is being given in cases where not only is it not the same case, but in this instance not even the same person.

Sometimes it is the same party, just not the same case. e.g. party id 150980 (City of Chicago) connects to 219 distinct cases.

If you need more please let me know. Again, apologize for the initial lack of specificity.

mlissner commented 8 months ago

Thanks for those details. Yeah, it looks like we're using the same "Chad Smith" (https://www.courtlistener.com/api/rest/v3/parties/68837/) in multiple entirely unrelated dockets in a way that is wrong. It's pretty obvious that the Chad Smith in these two cases is not the same person:

This isn't how the system is supposed to work and I don't know of a bug that would have caused this. I do notice though that these cases are all pretty old. It's possible this problem arose some time ago during an import.

I'm going to go purchase the party info and see if it gets automatically fixed.

mlissner commented 8 months ago

Hm, I bought and RECAP'ed both of these cases and it didn't fix the issue. This is worth more investigation to see if it's happening now (I doubt it), or if it's "just" the case that older cases don't get updated when new data comes in.

Another thing that surprises me is that the Chad Smith in the Chicago case has a "Start number" in PACER, but that never seems to go into RECAP:

image

a-drumm commented 8 months ago

I believe that's the Officer Star #, like a badge ID.

And this doesn't affect my ability to do anything, I just have to dump all the case info I don't want. Did confuse the hell out of me at first though.

mlissner commented 8 months ago

Yeah, I guess the fix would be to split up these kinds of parties, but I actually thought that was how it already worked, which is why I think it's only (mainly?) affecting older cases?

a-drumm commented 7 months ago

I found this after matching data from the City on cases finalized w/ settlements from 2019-2022. No newer cases had anything like this, so the only affecting older cases thing seems reasonable. All the same kind of cases, same place, etc. Only difference is when they were filed.

mlissner commented 7 months ago

Sorry, when you say "older" can you clarify what you mean? I think I mean before 2016 or so? Are you saying you see this in your 2019-2022 cases?

a-drumm commented 7 months ago

I have cases that terminated in 2019-2022, by older I'm referring to ones that were filed about 2016 or earlier.

mlissner commented 7 months ago

OK. So I think the first task here, in that case, is just to make sure new additions are working properly. If so, we can investigate how big the problem is and think through how to solve it. Thanks @a-drumm.