When writing documents of sufficient length, I find it extremely helpful to have a good indication of how far along I am. In this case I’m going one last time through my documentation and tying up any loose ends. The whole thing is 86 pages long front to back.
It started as this:
Which is nice and helpful and easy to do:
\subsection{Codec\statusdone{}}
But I wanted more. So I wrote a parser to extract that information et voilà:
Here’s a longer excerpt:
…
installation.tex: 0%
concepts.tex: 57%
pipeline.tex: 63%
performance.tex: 83%
cython.tex: 100%
…
65 of 95 sections done, 30 left
▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓░░░░░░░░ 68%
{u'canonical': u'02 30 "via bmndr at 22:10:40"',
u'comment': u'via bmndr at 22:10:40',
u'daystamp': u'20200602',
u'fulltext': u'2020-Jun-02 entered at 22:10 via api',
u'id': u'5ed6b2403b53bf26d3008392',
u'origin': u'api',
u'requestid': None,
u'status': u'created',
u'timestamp': 1591128640,
u'updated_at': 1591128640,
u'value': 30.0}
Because this is just something I need for the next few days I didn’t pour lots of time into this so I botched together a simple Beeminder integration using bmndr (hence the Python output).
And it worked great, so I threw it in a watch -n 30 'sbt sectionscanner/run' so it would run every 30s and update Beeminder accordingly.
But… it was a bit too simple because… well let’s just say Beeminder only allows a not-so-large number of datapoints per day 
For more LaTeX goodness, see this other thing I made, which is actually closely related:
