LUCENE-10312: Make stemming configurable on PersianAnalyzer #906
Conversation
…mer-configurability
| * @param stemExclusionSet a set of terms not to be stemmed | ||
| */ | ||
| public PersianAnalyzer(CharArraySet stopwords, CharArraySet stemExclusionSet) { | ||
| public PersianAnalyzer( | ||
| CharArraySet stopwords, boolean useStemming, CharArraySet stemExclusionSet) { |
There was a problem hiding this comment.
How about this three-args constructor private and keep the two-args constructor public PersianAnalyzer(CharArraySet stopwords, CharArraySet stemExclusionSet) so that we make the API changes minimum on the next major release?
There was a problem hiding this comment.
I'd suppose when stemExclusionSet is set useStemming flag is always set to true, so I think the three-args constructor can be internal-use only.
|
I don't think we should make this configurable. Please let's not do this to the analyzers. Just add the stemmer in 10.x, done. |
That also works for me, the TokenFilter will be available in 9.2 but you have to create an analyzer yourself to use it, and we delegate BWC handling to factories. I'll update this PR to just add a MIGRATE entry explaining how the analyzer has changed, and open a separate one against 9x to remove the stem filter from PersianAnalyzer. |
I wonder if my first PR #904 would work? |
|
Closing in favour of #904 |
No description provided.