Uploaded image for project: 'Nuxeo Platform'
  1. Nuxeo Platform
  2. NXP-2581

Fix lucene french analyzer weaknesses

    XMLWordPrintable

    Details

    • Type: Bug
    • Status: Resolved
    • Priority: Minor
    • Resolution: Fixed
    • Affects Version/s: 5.1.5
    • Fix Version/s: 5.1.6, 5.2 M3
    • Component/s: None
    • Impact type:
      Configuration format change
    • Upgrade notes:
      Hide

      The change is fully compatible for existing instances.
      However, there's a configuration change to perform to benefit from the new analyzer, in the infamous compass.cfg.xml, the analyzer class has to be changed to org.nuxeo.ecm.core.search.backend.compass.lucene.analysis.fr.FrenchAnalyzer. Please note that the typical instance with SQL backend has an overridden version of the compass.cfg.xml file.

      Upgrade of all default configurations (vanilla Nuxeo, postgresql-setup, oracle-setup, derby-setup) is included with the fix.

      Show
      The change is fully compatible for existing instances. However, there's a configuration change to perform to benefit from the new analyzer, in the infamous compass.cfg.xml, the analyzer class has to be changed to org.nuxeo.ecm.core.search.backend.compass.lucene.analysis.fr.FrenchAnalyzer. Please note that the typical instance with SQL backend has an overridden version of the compass.cfg.xml file. Upgrade of all default configurations (vanilla Nuxeo, postgresql-setup, oracle-setup, derby-setup) is included with the fix.

      Description

      Problems :

      • stemming is case sensitive
      • french apostroph not properly interpreted
      • dot is considered as a token separator

      For that, we need to make a Nuxeo specific version of the french analyzer.

        Attachments

          Activity

            People

            • Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: