CatgirlIntelligenceAgency/code/processes/index-constructor-process
2024-02-22 17:31:25 +01:00
..
src/main/java/nu/marginalia/index (zk-registry) epic jak shaving WIP 2024-02-22 14:01:23 +01:00
build.gradle * Extract the index functionality 2024-02-22 17:31:25 +01:00
readme.md (doc) Update docs 2024-02-06 16:29:55 +01:00

The index construction process is responsible for creating the indexes used by the search engine.

There are three types of indexes:

  • The forward index, which maps documents to words.
  • The full reverse index, which maps words to documents; and includes all words.
  • The priority reverse index, which maps words to documents; but includes only the most "important" words (such as those appearing in the title, or with especially high TF-IDF scores).

This is a very light-weight module that delegates the actual work to the modules:

Their respective readme files contain more information about the indexes themselves and how they are constructed.

The process is glued together within IndexConstructorMain, which is the only class of interest in this module.