[
https://issues.apache.org/jira/browse/LUCENE-1794?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12742994#action_12742994
]
Shai Erera commented on LUCENE-1794:
------------------------------------
bq. I only call reset() on streams.result when there is a state-keeping
TokenFilter on that chain. it is not necessary to invoke it if reset() is a
no-op...
What if one of those streams will become state-keeper some day? I don't think
that calling reset() and have it done nothing will be expensive, no?
bq. but now my problem child is memory/PatternAnalyzer, its source of tokens is
not a Tokenizer.
It could just override reusableTokenStream and do what it wants, no?
I must admit that I still have in mind the current TokenStream and Analyzer
API. Therefore my suggestion may not be 100% compatible w/ AttributeSource and
the new stuff. But my gut feeling tells me there has to be a way to remove all
those unnecessary impls in all Analyzers. The following default impl seems too
obvious than to say we cannot do it:
{code}
protected TokenStream internalTokenStream() {
// do something
}
protected Tokenizer getTokenizer() {
// do something
}
public TokenStream tokenStream() {
TokenStream result = getTokenizer();
result = internalTokenStream(result);
return result;
}
public TokenStream reusableTokenStream() {
Streams streams = getPrevTS();
if (streams == null) {
streams = new Streams();
streams.tokenizer = getTokenizer();
streams.tokenStream = internalTokenStream(streams.tokenizer);
setPrevTS(streams);
} else {
streams.reset();
}
return streams.tokenStream;
}
{code}
I'll try to impl it tomorrow, using the new API and see how it goes.
> implement reusableTokenStream for all contrib analyzers
> -------------------------------------------------------
>
> Key: LUCENE-1794
> URL: https://issues.apache.org/jira/browse/LUCENE-1794
> Project: Lucene - Java
> Issue Type: Improvement
> Components: contrib/analyzers
> Reporter: Robert Muir
> Assignee: Robert Muir
> Priority: Minor
> Fix For: 2.9
>
> Attachments: LUCENE-1794.patch, LUCENE-1794.patch, LUCENE-1794.patch,
> LUCENE-1794.patch, LUCENE-1794.patch
>
>
> most contrib analyzers do not have an impl for reusableTokenStream
> regardless of how expensive the back compat reflection is for indexing speed,
> I think we should do this to mitigate any performance costs. hey, overall it
> might even be an improvement!
> the back compat code for non-final analyzers is already in place so this is
> easy money in my opinion.
--
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: [email protected]
For additional commands, e-mail: [email protected]
