Lucene for JRipples analysis module
Lucene for JRipples analysis module computes ranks for components based on similarity between pre-processed (indexed) source code of these nodes and the text of a query, entered by a user.
Lucene for JRipples plugin is based on an open-source text search engine Lucene. Libraries of the Lucene engine are bundled with Eclipse; thus, any improvements to Lucene engine will be automatically propagated to Eclipse and Lucene for JRipples plugin.
Using Lucene for JRipples
Lucene for JRipples module provides two commands, available through JRipples context menus, as well as through the main JRipples menu:
- Set query - using this command, a user can enter a text query that will be matched to the indexed source code of components. Query can be updated as many times as necessary; every time the user updates the query, node ranks are re-computed and, if necessary, updated in JRipples views.
- Index - this command tells Lucene engine to index (or re-index if the index already exists) the source code of components. Usually, index is computed only once, and then re-used throughout the maintanence process. Index is stored in workspace_adress/.metadata/.plugins/org.severe.jripples.lucene/project_name folder, where workspace_adress is the path to a current workspace and project_name is the name of a project under analysis. Index can be analyzed with third-party tools, e.g. Luke.