Missing zero leading to broken links

[v-anne]

This issue is related to this post.

The URL in the post is dead, because of the missing 0 in the document number.
https://www.courtlistener.com/docket/69673016/1208714529/aids-vaccine-advocacy-coalition-v-dos/?redirect_or_modal=True
vs.
https://www.courtlistener.com/docket/69673016/01208714529/aids-vaccine-advocacy-coalition-v-dos/

I am willing to fix it, if someone can point out where the issue may be.

[v-anne]

My suspicion is this is caused by the storage of document numbers as integers in the db, because they will drop leading 0s.

See bc/subscription/models.py, namely the definition of document_number in FilingWebhookEvent. I don't know how many courts are affected by this problem, or whether changing the type from int to str in the DB would break other things. Might not be worth fixing.

[mlissner]

Probably this is the issue that needs to be fixed upstream to fix this issue here: freelawproject/courtlistener#3385. I think there's a whole stream of fixes that are needed to get this fixed — or we can probably do some workaround here, but that's a bummer and defeats the purpose of this project (to dogfood our systems and flush out bugs).

So I'll bump the priority of that issue and hopefully we can get the fix here eventually.

[v-anne]

Yeah, no need to do a haphazard fix here, rather than resolving the upstream issue.

[v-anne]

I think I found a fix to this issue over in #5186 for CourtListener. I had resolved 5292 and unintentionally fixed the leading 0 issue as well by using RECAPDocument.get_absolute_url(). Is it possible to import that function to this codebase? My understanding is that function treats the document number as a CharField instead of an int.

[mlissner]

Hm, are you sure this is fixed? the get_absolute_url method is:

    def get_absolute_url(self) -> str:
        if not self.document_number:
            # Numberless entries don't get URLs
            return ""
        if self.document_type == self.PACER_DOCUMENT:
            return reverse(
                "view_recap_document",
                kwargs={
                    "docket_id": self.docket_entry.docket.pk,
                    "doc_num": self.document_number,
                    "slug": self.docket_entry.docket.slug,
                },
            )
        elif self.document_type == self.ATTACHMENT:
            return reverse(
                "view_recap_attachment",
                kwargs={
                    "docket_id": self.docket_entry.docket.pk,
                    "doc_num": self.document_number,
                    "att_num": self.attachment_number,
                    "slug": self.docket_entry.docket.slug,
                },
            )

Noting special there about char fields and whatnot. Since we store it as an int, I'm not sure how application code could fix it, actually...

[v-anne]

Sorry Mike, you're correct. I looked at the definition of document number in the RECAPDocument model itself, but that doesn't help us here since we're relying on a separate model. Disregard this suggestion...