mr.cabot is a tool for allowing software projects to map where their contributions are coming from. It reads data form common sources like GitHub, gmane and stackoverflow and attempts to locate where those contributors are physically located.
You need to customise your mr.cabot.cfg config file to match your software project. The basic format is that the sources line lists the sections that supply lists of contributors and users lists the sections that supply lists of contributors.
Only the GitHub sources support user enumeration at the moment, but stackoverflow and gmane both extract geolocation information where possible.
The current example from the Plone project generates the following:
- type
github
- key
the name of the organisation to be scanned
- token
an oauth token, preferably one with the repo scope. See https://help.github.com/articles/creating-an-oauth-token-for-command-line-use
- checkout_directory
an existing directory to cache checkouts in to save on bandwidth
- type
git
- key
a clone url of a git repository
- type
stackoverflow
- key
a tag used on stack overflow
- days
the number of days of history to download
- type
gmane
- key
the full gmane newsgroup name
- messages
downloads the last x messages
Output type defaults to a google static map, but can be selected using --output type on the command line.
The available options are:
- googlestaticmap
- html
- kml
No caching of geolocation data is included as yet, but data runs are cached under var/data, with the filename yyyy-mm-dd.pickle. This allows you to re-run old data sets if you are changing display methods.
To load a pickle instead of re-scanning use the --pickle command line option:
./bin/cabot --pickle ./var/data/2013-01-05.pickle --output kmlAdditionally, there is a command line option to skip pulling git repositories. This is useful for re-running when you are tweaking the config file initially, as updating git is the slowest section:
./bin/cabot -N --output kml