Ensures that fiction metadata element is being stored in the search i…

[dbernstein]

…ndex in lowercase so that searches for "fiction" and "nonfiction" produce the expected results.

Resolves: https://ebce-lyrasis.atlassian.net/browse/PP-1332

Description

This is the story as I understand it:

The "Fiction" and "Nonfiction" values associated with "fiction" search document property have been stored in a mix case way since at least 3/23/2022. At that time we did not have a classifier list filter in the custom list builder in circulation-admin. When that feature was introduced on 3/28/2024, the list values of "fiction" and "nonfiction" were defined. The client sends those values to the backend for processing. When @jonathangreen made that addition we had fuzzy filtering turned on so "fiction" and "nonfiction" in the query matched for "Fiction" and "Nonfiction" in the search index. However, when we disabled fuzzy they are no longer exact matches. I believe we are both normalizing our queries to use lowercase matching and lowercasing any text going into the search index in the author, title, summary, genre, audience etc fields. It appears that fiction was not being lowercased and thus the non-matching once fuzziness was disabled.

This update simply lowercases the fiction field value on the way into the search index. This way, search queries coming into palace using "fiction" and "Fiction" will produce the expected matches.

Motivation and Context

https://ebce-lyrasis.atlassian.net/browse/PP-1332

How Has This Been Tested?

I verified that the matches were not occurring. Then I verified that changing the incoming value from the client to mix case caused search results to appear. I wrote a failing end to end test that demonstrated the problem. I then put in the fix and confirmed the test was passing. Finally I verified that the UI was giving the expected results post fix.

I'll be adding an end to end test to bring this one to completion.

Checklist

• I have updated the documentation accordingly.
• All new and existing tests passed.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 90.13%. Comparing base (d14b879) to head (b44fabc).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1886   +/-   ##
=======================================
  Coverage   90.12%   90.13%           
=======================================
  Files         322      322           
  Lines       39626    39626           
  Branches     8629     8629           
=======================================
+ Hits        35713    35716    +3     
+ Misses       2604     2602    -2     
+ Partials     1309     1308    -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jonathangreen]

Made a couple comments on this one that would be nice to clean up before merging, but it looks good.

Since we've had this field in mixed case for some time, I'm a bit concerned that this change is going to break some other use of this field in the search index, where we are currently depending on it to be mixed case, instead of lower case. I'm assuming you've looked for uses like this already.

[jonathangreen]

I don't think the test name here has to refer to a ticket. It would be nice to have the test name stand on its own, so we don't have to go digging in jira in the future when we are refactoring this test.

[jonathangreen]

This isn't really the default filter, the default filter is what you get when you don't pass a filter into the query. This should probably be called filter or json_filter.

[dbernstein]

@jonathangreen : yes, I reviewed all the instances of "Fiction" that may be affected by this change and I didn't see anything that would present a problem.