lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Stefan Seidel (JIRA)" <>
Subject [jira] Commented: (SOLR-1316) Create autosuggest component
Date Thu, 17 Jun 2010 15:07:38 GMT


Stefan Seidel commented on SOLR-1316:

Hi. First of all, thanks for making this effort for a usable auto-suggest feature.

I have a question which is (hopefully) relevant to all this: we tried using the facet method
for autosuggest as presented in David's book about solr. It works well and usually fast, however,
there are two issues: (a) initializing the faceting takes time (>300 sec. for 8.3 million
docs here), which would be ok if not (b) after any commit (even without new documents added)
the faceting needs exactly this time again. On a heavy load server with commits every 60 seconds,
this doesn't work at all. Also, it makes solr use >1G of RAM, and I can get it into HeapSpace
errors when multiple queries are running and the faceting hasn't been initialized.

Question: has this scenario (frequent commits) been considered a use case when implementing
these new components? Will the tree get rebuilt at every commit as well? (Judging from the
previous numbers this should take about 10 sec. in our scenario?) I think this is a crucial
question for the whole feature.

> Create autosuggest component
> ----------------------------
>                 Key: SOLR-1316
>                 URL:
>             Project: Solr
>          Issue Type: New Feature
>          Components: search
>    Affects Versions: 1.4
>            Reporter: Jason Rutherglen
>            Assignee: Shalin Shekhar Mangar
>            Priority: Minor
>             Fix For: Next
>         Attachments: SOLR-1316.patch, SOLR-1316.patch, SOLR-1316.patch, SOLR-1316.patch,
SOLR-1316.patch, SOLR-1316.patch, suggest.patch, suggest.patch, suggest.patch,
>   Original Estimate: 96h
>  Remaining Estimate: 96h
> Autosuggest is a common search function that can be integrated
> into Solr as a SearchComponent. Our first implementation will
> use the TernaryTree found in Lucene contrib. 
> * Enable creation of the dictionary from the index or via Solr's
> RPC mechanism
> * What types of parameters and settings are desirable?
> * Hopefully in the future we can include user click through
> rates to boost those terms/phrases higher

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message