[Feature Request]: Somehow weaken typo matches (not exact, but caught in Levenshtein distance)

[dpasque]

"What": What is your idea?

Right now, we allow a little bit of search fuzziness in our searches -- essentially allowing for typos. We scale based on length, but for most words, you can have 1 character wrong. This makes the search more forgiving, allowing for common typos both for searchers and issue reporters.

But, this can also be a nuisance on occasion! See issue #150 -- photon is one edit away from photos -- which is unfortunately a pretty common word.

We should find a way to soften those cases.

"Why?": Why would this be impactful?

In certain cases, helps you more accurately find the issue you're looking for.

"How?" How should this feature work?

Couple of ideas:

• See if ElasticSearch has the ability to score issues that typo match lower than ones that don't.
• See if there's a way to turn off typo matching if the term is put in quotes.