(from github.com/andreierdoss)
Hello,
How can I customize Fess to index correctly Romanian language sites? I would like to replace the Japanese specific files with the Romanian ones.
Thank you
(from github.com/andreierdoss)
Hello,
How can I customize Fess to index correctly Romanian language sites? I would like to replace the Japanese specific files with the Romanian ones.
Thank you
(from github.com/andreierdoss)
Thank you for the response.
What would I have to do for Kuromoji equivalent?
Where is the synonyms dictionary being used?
(from github.com/marevol)
What would I have to do for Kuromoji equivalent?
Analyzers for each languages affect search quality.
For reference to create an analyzer, Solr uses https://github.com/apache/lucene-solr/blob/master/solr/example/files/conf/managed-schema
You can create the analyzer definition from Solr’s one.
Where is the synonyms dictionary being used?
synonym data is in .configsync index.
(from github.com/andreierdoss)
I want to use the Romanian Elasticsearch’s analyzer. How do I configure this? Do I just change it in the doc.json to from “empty_analyzer” to “romanian_analyzer”?
Also what is the format for the synonyms doc? Currently it has 2 comma separated words. Can it be multiple words?
(from github.com/marevol)
Analyzer: https://www.elastic.co/guide/en/elasticsearch/reference/current/analysis-analyzers.html
synonym file: https://www.elastic.co/guide/en/elasticsearch/reference/2.3/analysis-synonym-tokenfilter.html#_solr_synonyms
(from github.com/andreierdoss)
Thank you. I updated fess.json and also doc.json for Romanian. Does this look ok? Am I missing something?
(from github.com/andreierdoss)
I edited doc.json and fess.json files as you see above (https://gist.github.com/andreierdoss/a601d20bd14207f9f6f314f47297d6a5), but these changes are not reflected in the fess index. How would I go about making these changes take effect?
(from github.com/marevol)
The index is created at startup phase if it does not exist.
So, please delete old fess index.
© 2020. All Rights Reserved - CodeLibs, Inc.