Searching for "digital" finds "digit"

[eaquigley]

Pointed out by someone on Twitter:

"try search "digital" and you find files labelled "digit". add phrase support."

In production, I searched for digital and narrowed down to files. Digital did show up in the description of some files (on page 66, I was logged in) but not in the name of the file as far as I could see in my quick look. The first page of results (and many after) had Digit in the title but not Digital. We should look at the relevancy for searches; would make sense that if the exact term showed up anywhere in the metadata of an object, it would be one of the first results displayed.

[pdurbin]

Ah, this tweet: https://twitter.com/jonintweet/status/637223429175947264

Hmm. We worked a bit on search relevance in #1928 . Maybe the comment from @tercer at #1928 (comment) should be copied to this issue since that was for Beta 15 (but still open in QA).

[pdurbin]

"try search "digital" and you find files labelled "digit"

Easy to reproduce at https://dataverse.harvard.edu/dataverse/andrewleigh/?q=digital (running 5.11.1):

[qqmyers]

FWIW - there is phrase support - searching for "digital code" only returns the first hit above. However, stemming is enabled so the various forms of words - digital/digit, code/codes are found, even in the phrase.

[pdurbin]

Yes, stemming. Exactly. What we should probably do is first write some tests to make assertion about how our search works. Then try to fix this issue and hopefully not cause other problems. 😄

If anyone wants to pick this up, here is where we keep the search tests: https://github.com/IQSS/dataverse/blob/v5.12/src/test/java/edu/harvard/iq/dataverse/api/SearchIT.java

[cmbz]

To focus on the most important features and bugs, we are closing issues created before 2020 (version 5.0) that are not new feature requests with the label 'Type: Feature'.

If you created this issue and you feel the team should revisit this decision, please reopen the issue and leave a comment.