danrosher commented on code in PR #935:
URL: https://github.com/apache/solr/pull/935#discussion_r940144508
##########
solr/solrj/src/test/org/apache/solr/client/solrj/io/stream/StreamDecoratorTest.java:
##########
@@ -4370,6 +4370,17 @@ public void testClassifyStream() throws Exception {
updateRequest.add(id, String.valueOf(3), "text_s", "a b e e f");
updateRequest.commit(cluster.getSolrClient(), "uknownCollection");
+ expr =
+ "classify("
+ +
+ // use cacheMillis=0 to prevent cached results. it doesn't matter
on the first run,
+ // but we want to ensure that when we re-use this expression later
after
+ // training another model, we'll still get accurate results.
+ "model(modelCollection, id=\"model\", cacheMillis=0),"
+ + "topic(checkpointCollection, uknownCollection, q=\"*:*\",
fl=\"text_s, id\", id=\"1000000\"),"
+ + "field=\"text_s\","
+ + "analyzerField=\"tv_text\")";
Review Comment:
checkpoints are persisted when the stream is closed, or if checkpointEvery >
-1 (and then every count % checkpointEvery), otherwise the checkpoints are
stored in the checkpoints hashmap, so for 'just' added docs, I think as long as
is it matches the underlying query, and those docs have been soft committed
(see caveat for topicstream SOLR-8709), I think they should be picked up,
unless I'm completely misunderstanding ?
--
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]
