CenterForOpenScience / SHARE

SHARE is building a free, open, data set about research and scholarly activities across their life cycle.
http://share-research.readthedocs.io/en/latest/index.html
Apache License 2.0
101 stars 58 forks source link

[SHARE-925][Feature] Allow case insensitive keyword search #685

Closed laurenbarker closed 7 years ago

laurenbarker commented 7 years ago

Purpose

Allow case insensitive search for sources.

Changes

Add a normalizer for keywords.

Side effects

Need to re-index everything.

aaxelb commented 7 years ago

Does using a normalizer like this mean the sources facet/donut will be all lower case?

laurenbarker commented 7 years ago

Ew. Yes it does.

aaxelb commented 7 years ago

Maybe we should add a sources.exact field for aggregations/filtering? Then once preprints/dashboard/whoever have switched over, sources could be text, which would be nicer for typed queries.

laurenbarker commented 7 years ago

That would probably make the most sense.

laurenbarker commented 7 years ago

@chrisseto would you be okay with splitting the case insensitive search stuff into two releases? One for adding an exact field for sources and another for making the sources field a text field. It would probably be easier than coordinating a release with preprints and the dashboard.

chrisseto commented 7 years ago

Would it make more sense to use an analyzer/tokenizer given that they are officially supported?

laurenbarker commented 7 years ago

@chrisseto, analyzer/tokenizer can't be used on a keyword field

Keyword fields are only searchable by their exact value.

So are you saying your on board with switching to a text field or am I missing something?

chrisseto commented 7 years ago

@laurenbarker Do you remember what the final decision on this was?

laurenbarker commented 7 years ago

@chrisseto, we still haven't heard back from product on this one. I would say move it out of the sprint and close the PR until a decision is made.

chrisseto commented 7 years ago

@laurenbarker alright, thank you!