Add unique case paths to static export

This PR tries to handle the tricky issue of how to find one case in the static file export from another.

Problem: If you have a case like /us/123/cases/0012-01.json that cites to case id 4567, or a list of case IDs in /us/123/CaseMetadatas.json, it's hard to find the files that contain the data for those cases.

Solution: Everywhere that we have cases referred to by ID, also refer to them by path. A path for a case is like "/us/123/0012-01", or just "0012-01" if the reporter and volume are already implicit.

So:

In CaseMetadatas.json and single case json files, add a "file_name": "0012-01" field. This makes it possible to jump from CaseMetadatas.json to the associated json or html file, allowing the frontend to render the list of cases in a volume for example.
In case json, when we link to a case this case cites with "cites_to": [{"case_ids": [1,2,3...]}], add a field for each citation like "case_paths":["/us/1/0001-01", "/us/2/0002-01", ...] that parallels "case_ids".
In case html, when we link citations to their targets, edit them to point to case paths instead. This has a few options:
- We know of no target case for the citation: link to href="/citations/?q=123 Wtf. 456". (This is common for statutes and regulations).
- We know of one or more target case for the citation: link to href="/us/1/0012-01" data-case-paths="/us/1/0012-01,/us/1/0012-02".

Deployment notes:

Case paths are hard to calculate, because you have to add an increment for two cases starting on the same page of the same book, like 0123-01 and 0123-02. So this PR adds a column in a migration to store that calculated string on CaseMetadata, and after deployment we should run fab set_case_static_file_names (which runs in a tqdm loop rather than in celery) before redoing fab export_cap_static_volumes for the non-redacted jurisdictions.

harvard-lil / capstone

Add unique case paths to static export #2185

Codecov Report