Apache Lucene and Solr 3.1 released
The Lucene and Solr developers have released version 3.1 of the text search engine Lucene 3.1 and enterprise search platform Solr 3.1. This is the first major release of both applications since the projects moved to a merged single development team in March 2010. Solr's version number has jumped from 1.4.1 to 3.1 to reflect how the developers will be keeping both Lucene and Solr synchronised. The developers are, though, keeping Lucene and Solr seperately downloadable, as Lucence addresses developers building applications with search, whereas Solr is more suited to systems integrators building a platform for searching.
Solr 3.1 adds a number of new features beyond the improvements in the Lucene engine. Search improvements include numerical range facets, a new spatial search, a faster termvector-based highlighter and distributed search support for SpellCheckComponent and TermsComponent. Solr is now also capable of indexing JSON documents and responding in CSV format. Apache UIMA is now used to extract metadata from documents. New Solr installations will now have an example Velocity driven search UI available on http://localhost:8983/solr/browse and developers will be able to make use of Suggester, a new trie-based autocomplete component. Full details of the changes, and details on how to upgrade from Solr 1.4, are available in the release notes.
The highlight of the new release of the Lucene engine is performance improvements, with faster exact phrase queries, faster primary key lookups, IndexWriter index addition using file copying rather than merging and directory performance improvements. Other changes include improved Unicode support in analysis, easier TokenStream reuse and use of memory mapped directories by default on 64-bit Windows and Solaris JVMs. More details are contained in the announcement and in the Lucene change log.
Both releases are licensed under the Apache Licence 2.0. The applications can be downloaded from the Solr mirror page and Lucene mirror page where an appropriate mirror site will be suggested for downloads.