mkhludnev opened a new pull request, #1233:
URL: https://github.com/apache/solr/pull/1233
The way how distributed MLT works now lacks of perfection, you know.
Here's not a real proposal, it just demonstrates current state.
1. MLT respond with Query instance. However, it might be list of terms, or
so. I keep it as-is
2. These queries should be responded to query coordinator. In HEAD it's just
`BoolQ.toString()` that's a pity. There are two options instead: Lucene query
can be structured as named list, or as a blob. I started from the former
approach, but then decided to dump it to blob with Jackson+Smile. So, in the
patch MLT query successfully arrives to coordinator.
3. Then, we need to spread it across shards. And now it's prototyped in a
little bit weird way. Since, it's deadly hard to send shard request with
payload in body, I have to encode it to base64 and hook it up by dedicated
QParser. Yeah...
So, where to tweak it further?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
