Did you mean

Lucene has a great feature to "auto-correct" the query terms.

Anyone has tried to implement it with ES ?

Thanks

There has been an open issue about this for almost a year:
"Did you mean" spellchecking · Issue #911 · elastic/elasticsearch · GitHub. Waiting for
Lucene 4.0. Quite when that will be, who knows...

On Feb 27, 7:18 pm, Alexandre Heimburger alexheimbur...@gmail.com
wrote:

Lucene has a great feature to "auto-correct" the query terms.

Anyone has tried to implement it with ES ?

Thanks

I was able to roll a poor man's did you mean using the following
method:

  • If a query returns no results, perform fuzzy query at 0.9, 0.8, 0.7,
    and 0.6 stopping once I get results
  • Enable highlighting, which allows me to extract the matched terms

This probably won't scale well, but works for fine on the data set I'm
working against. Also, analysis doesn't happen to fuzzy queries, so
for a search like "Elasticsrch is COO", I run multiple fuzzy queries,
eg:
elasticsrch~0.9 is~0.9 coo~0.9

Notice, I've lowercased before passing in due to analysis not
occurring. YMMV with this technique, but has been good enough until
the fully baked support is available.

Best Regards,
Paul

On Feb 28, 1:28 am, Nick Dunn n...@nick-dunn.co.uk wrote:

There has been an open issue about this for almost a year:"Did you mean" spellchecking · Issue #911 · elastic/elasticsearch · GitHub. Waiting for
Lucene 4.0. Quite when that will be, who knows...

On Feb 27, 7:18 pm, Alexandre Heimburger alexheimbur...@gmail.com
wrote:

Lucene has a great feature to "auto-correct" the query terms.

Anyone has tried to implement it with ES ?

Thanks