Luke is the GUI tool for introspecting your Lucene / Solr index. It allows:
- Viewing your documents and analyzing their field contents (for stored fields)
- Searching in the index
- Performing index maintenance: index sanity checking, index optimization (take a backup before running this!)
- Reading index from hdfs
- Exporting the index or portion of it into an xml format
- Testing your custom Lucene analyzers
See CHANGES.txt
This lightning talk was delivered by Dmitry Kan on ApacheCon 2015 in Austin, Texas.
- The original author is Andrzej Bialecki
- The project has been mavenized by Neil Ireson (see google group discussion here)
- The project has been ported to Lucene trunk (marked as 5.0 at the time) by Dmitry Kan
- The project has been back-ported to Lucene 4.3 by sonarname, who later decided not to continue supporting the project
- There are updates to the (non-mavenized) project done by tarzanek
- Keep the project mavenized (compatible with Apache Lucene and Solr style)
- To port the thinlet UI to an ASL compliant license framework so that it can be contributed back to Apache Lucene. Current work is done with Apache Pivot (https://github.com/dmitrykey/luke/tree/pivot-luke).
- To revive the project and establish a single point of trust for the development and updates of the tool. This said, everyone is welcome to join.
Luke has become somewhat greedy on the perm gen size and thus requires an extra parameter to start:
java -XX:MaxPermSize=512m -jar luke-with-deps.jar
See luke.sh for an example of launching luke with a custom analyzer.
The documentation is being written along the way of discovering the luke's usage scenarios.
https://github.com/DmitryKey/luke/wiki
You are free and welcome to contribute you findings!
Found bug? Missing a useful feature? Raise an issue here:
https://github.com/DmitryKey/luke/issues
Or, and btw, the pull requests are welcome as well!
All of the releases you find under the "releases" link above are versioned after the Lucene's version they use.
Usually we don't do releases for minor version upgrades, because the major release usually can read the index of the next minor release.
An example: lucene 4.8.0 can read the index generated by lucene 4.8.1. Hence the luke 4.8.0 can read too.
Conclusion: in order to find a luke release that can read an index of your version of Lucene, pick the closest major version and download the luke for that from the releases page above.
There is no separate luke 4.4.0 release, but luke 4.5.0 should open the Lucene 4.4.0 index just fine.
The license for the code is ALv2, excluding the license of the thinlet library that is used for the UI.