Ever since upgrading to JackRabbit 2.x and relying on Tika parsers for full text indexing, people have struggled with configuring those indexes. This concept is about making full text parsing configuration explicit and thus allowing Magnolia users to easily customize it to tune performance of their instances.
ready for implementation
add <param name="tikaConfigPath" value="/info/magnolia/jackrabbit/tika-config.xml"/>
tika-config.xmlby default contain following:
The alternative to the above would be to have such config file located directly in the workspace and configure full text parsing per workspace.
The above configuration is proposed as default since it removes most of resource intensive indexing while having minimum impact on available data for full text search. For DAM resources we still rely on extracted metadata to locate such resource, we are not using full text search index to locate them.
Default configuration is placed in
org/apache/jackrabbit/core/query/lucene/tika-config.xml and contains following (JR 2.6.2):