This tool is designed for two purposes:
- an overview of the state of the review queues of Servo's reviewers
- a historical record of the trends in review queues
Based on the list of reviewers in the Homu configuration file:
- The number of PRs assigned to a particular github username at a given moment
- The number of days since the oldest assigned PR was created
- The average number of days since each assigned PR was created
These measurements are intended to reflect trends in reviewing. If a reviewer consistently has a large queue size but the average age is very small, that means they are being efficient in their reviews and no PR is lingering. On the other hand, if the queue size fluctuates but the oldest PR age remains constant, that means that a PR is being ignored.
It is unclear how meaningful average PR age is, but its intent is to indicate something about how long PRs linger in a particular reviewer's queue.
$ sqlite3 queues.db <schema.sql
$ python scrape.py
$ python generate.py >index.html
scrape.py is intended to be run once per day to populate queues.db with that day's data scraped from github's APIs. Any missed day will result in missing data that cannot be recovered or synthesized.
generate.py uses the data in queues.db to fill out the various template HTML files and generate the static page that contains the massaged and charted data.