Add ][ characters to page regex of NY Slip Op and other NY reporters

[grossir]

We may be failing to resolve (or resolve incorrectly) NY citations due to a variation in the page format in the following scenarios:

• NY Slip Op citations: We accept an optional (U), but the opinions text may show an optional [U]
• Misc 3d citations: We accept an optional (A), but the opinions text may show an optional [A]
• Probably the other editions of these reporters...

Currently this happens, which may lead to incorrect matches

In [4]: get_citations("2024 NY Slip Op 51192(U),")[0].groups['page']
Out[4]: '51192(U)'

In [3]: get_citations("2024 NY Slip Op 51192[U],")[0].groups['page']
Out[3]: '51192'


In [6]: get_citations("11 Misc 3d 134[A]")[0].groups['page']
Out[6]: '134'

In [7]: get_citations("11 Misc 3d 134(A)")[0].groups['page']
Out[7]: '134(A)'

However, how could we standardize the page after extraction?

reporters-db/reporters_db/data/reporters.json

Line 18833 in d009aae

"2023 NYSlipOp 51350(U)"

Example
Highlighted the NY Slip Op; to the left is the Misc 3d

Another

---

courtlistener=> select count(*) from search_citation where reporter = 'NY Slip Op' and page like '%U%';
 count 
-------
  6259
(1 row)

courtlistener=> select count(*) from search_citation where reporter = 'Misc 3d' and page like '%A%';
 count 
-------
    11
(1 row)

[grossir]

When we ingest a scraped opinion we do:

    return Citation(
        cluster=cluster,
        volume=citation_objs[0].groups["volume"],
        reporter=citation_objs[0].corrected_reporter(),
        page=citation_objs[0].groups["page"],
...)

When we try to match citations, we do

    filters.append(
        Q(
            "match_phrase",
            **{"citation.exact": full_citation.corrected_citation()},
        )
    )

The correction in eyecite models

    def corrected_citation(self):
        """Return citation with corrected reporter."""
        if self.edition_guess:
            return self.matched_text().replace(
                self.groups["reporter"], self.edition_guess.short_name
            )
        return self.matched_text()

So, we could probably add a corrected_page method,

• that is also called in corrected_citation
• that is also called when ingesting a scraped opinion

Which for now would only standardize NY pages so we can actually match them

What would we need to set a canonical page format in reporters-db? Or we could just hardcode in eyecite the values we know in NY, and prefer () to []

[flooie]

See #207

[flooie]

I think the answer here is to improve the regexes to enable both options- which is what we did. and we will just add more variations as we see them instead of hard coding anything.