Details
-
Improvement
-
Resolution: Fixed
-
Medium
-
4.0.0
-
None
Description
While the existing version of Solr in eZ Find supports solr.MappingCharFilterFactory , it is possibly an old version (see https://issues.apache.org/jira/browse/SOLR-822 ), and it doesn't support solr.CharStreamAwareWhitespaceTokenizerFactory which appears to be needed (see http://wiki.apache.org/solr/SchemaDesign?highlight=(MappingCharFilterFactory)#head-cbd09984c67526fbfde825739d72e9c37139f52c ).
I've tried using MappingCharFilterFactory to map characters, with no success.
Steps to reproduce
1. Edit ezfind/java/solr/conf/schema.xml and add the charFilter line below to both the "query" and "index" analyzers:
<fieldType name="text" class="solr.TextField" positionIncrementGap="100"> <analyzer type="index"> <charFilter class="solr.MappingCharFilterFactory" mapping="mapping-macrons.txt"/>
2. Create a file in the same directory called mapping-macrons and paste the following:
"Ä" => a "Ä€" => A "Ä“" => e "Ä’" => E "Ä«" => i "Ī" => I "Å" => o "ÅŒ" => O "Å«" => u "Ū" => U
3. Put some of the Unicode chars in mapping-macrons.txt into an XML-block in a content object on your eZ Publish site e.g "MÄori".
4. Re-start Solr & re-index your site using the --clean option
5. Search for the mapped equivalent, e.g "maori"