Research: Legal search queries should ignore periods

fecgov / openFEC

The first RESTful API for the Federal Election Commission. We're aiming to make campaign finance more accessible for journalists, academics, developers, and other transparency seekers.

https://api.open.fec.gov/developers

Other

479 stars 106 forks source link

Research: Legal search queries should ignore periods #5946

Closed cnlucas closed 6 days ago

cnlucas commented 3 weeks ago

What we’re after

Our users want their queries to ignore periods, but standard analyzer ignores all punctuation except for periods. We should look into normalizing periods ideally without having to switch to a custom analyzer.

(This is not as important as other issues, but is nice to have)

Action items:

Can this be handled on the front-end? Back-end?
Idea: More keyword box - strip periods
Idea: Keywords search box - remove the period when sending the call to the API

Completion criteria

[x] Research and next steps are documented, any followup tickets have been opened

cnlucas commented 2 weeks ago

Currently a search for "52 USC" (no internal periods) will return 76 results and exclude results for "52 U.S.C.," and "52 U.S.C." (internal periods) will return 1255 results and appears to exclude "52 USC"

Because citations and names have periods (102.2, commissioner names..) I think this behavior should not be changed. Stripping the periods on the FE or ignoring them on the backend would ultimately limit our users ability to search by specific parts and sections. We would have to create a custom analyzer to fix this on the back end.

cnlucas commented 6 days ago

We are holding off on this for now, but may consider a custom analyzer later.