[
https://issues.apache.org/jira/browse/SOLR-4587?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17896132#comment-17896132
]
Luke Kot-Zaniewski commented on SOLR-4587:
------------------------------------------
Hello [~janhoy]! I am indeed still around. The main outstanding things I can
see:
1. Documentation (probably best I start this soon)
2. I have to revisit any outstanding PRs by [~cpoerschke] on my branch, i.e.
[https://github.com/kotman12/solr/pull/3] .. these remove the highlighting
feature as I've implemented it. I'm ok with this but it also prompted me to
rethink the API .. which brings me to 3 ..
3. Currently there are really two APIs but I'm leaning heavily towards one of
them. I want to remove the "bad API" and all the cruft code in it. The
highlight feature, for instance, was implemented in the bad API and I would
like to port it to the "good API". Unfortunately there are situations where the
bad API is much faster but that will hopefully be ameliorated by the latest
segment search parallelization changes in solr ("reverse" search can be CPU
intensive depending on how efficient the pre-matching-phase pruning is). On top
of that, I implemented the expensive matching phase as a solr post-filter
which, I believe, makes it difficult to take advantage of the latest
{_}intra-{_}segment parallelization work in [lucene
|https://github.com/apache/lucene/pull/13542]. I kind of regret doing it this
way now but changing this would require more effort than I can commit to at
this point.
4. A bit of an implementation detail but I manually "shade" lucene's
CandidateMatcher to expose a certain private method as protected. Ideally this
change would just be upstreamed in Lucene. Hopefully I can convince the lucene
community to do accept this small change, otherwise I'll have to implement a
solr equivalent of it.
With all that said I'd be happy to collaborate with this potential user to see
what features they are looking for. Perhaps it can guide the decisions in some
of these outstanding tasks.
Would you be able to add me l.kotzaniewski AT gmail.com to the apache solr
slack channel? I think that might be a better forum for discussing this in more
detail (unless you think otherwise).
> Implement Saved Searches a la ElasticSearch Percolator
> ------------------------------------------------------
>
> Key: SOLR-4587
> URL: https://issues.apache.org/jira/browse/SOLR-4587
> Project: Solr
> Issue Type: New Feature
> Components: SearchComponents - other, SolrCloud
> Reporter: Otis Gospodnetic
> Priority: Major
> Labels: pull-request-available
> Time Spent: 15h
> Remaining Estimate: 0h
>
> Use Lucene MemoryIndex for this.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
