Chapter 7.  Changes

Since 0.4, with 0.5: Solr and Lucene 5. Highlight and MoreLikeThis.

  opennlpmodelpath=/home/user/data/opennlp/LANG-doccat.bin (before only en)

new property

  mahoutbasepath=/home/user/data/mahoutc/LANG

if present the rest just added, otherwise, use with old full path. mahoutlabelindexfilepath=/labelindex (before /home/user/data/mahoutc/labelindex) Classifier multilanguage support new property comma-separated languages

0.6: Metadata change handling, stored, multi-valued. No more limit check for using alternative converters, only do that if tika crashes. Java 8. Usage of mime type probing. DataNucleus table changes.

0.7: New distributed mode, with usage of Curator and Hazelcast. A lot of refactoring have been done.

0.8: Redoing configuration, more validity checking. Multinode configuration.

0.9: Mahout spark.