1. Marcin Kuzminski
  2. RhodeCode
  3. Issues

Issues

Issue #322 resolved

Tracking with partial/without indexing

Anonymous created an issue

We use RhodeCode for a few projects that contain a very large number of files(in the 50K+ zone) and growing. I was asked to enable statistics, so I added the needed extensions to ADDITIONAL_MAPPINGS, but the index got to about 100MB after an hour and was still going at it, and most of that data wasn't relevant in searches anyway. Would it be possible to allow for custom indexing handlers for files (instead of the current read() and feed to whoosh) or at least a clean way of excluding files from the index but still add them to the statistics? I've done a small hack to skip adding the my extensions to the index, but this would be much nicer in production.ini or in the settings panel(except for custom handlers)

Comments (3)

  1. Marcin Kuzminski repo owner

    I think it would be good to split the stats extension map from whoosh index extensions, i like this idea, and i think it's worth implementation.

    I'm not sure how usefull would be handlers for particular files but i can say that implementing a mapping dict with regex_key:method_handler would be quite easy, unfortunetly due to limitations of ini files i would have put it in a somkind of config.py file located somewhere in source. I'll see if i can make a smarte solution, with a config next to. ini files.

  2. Log in to comment