arbtt-graph

arbtt - Automatic Rule Based Time Tracker

Note: Requires arbtt-stats version 0.9, since it has for-each=minute, which is required for the barcode chart to work.

I love arbtt. Just remember to add arbtt-capture to your startup applications and let it record active windows in the background. Then you can use arbtt-stats to categorize recorded data. For example, running arbtt-stats with the included categorize.cfg file:

$ arbtt-stats --categorizefile=./categorize.cfg -c Graph
Statistics for category "Graph"
===============================
________________Tag_|_________Time_|_Percentage_
Graph:social        | 10d01h18m00s |      11.23
Graph:mail          |  9d04h08m00s |      10.25
Graph:read-pdf      |  6d23h17m00s |       7.79
Graph:term          |  6d17h47m00s |       7.53
Graph:irc           |  2d16h47m00s |       3.02
Graph:tv-local      |  2d08h28m00s |       2.63
Graph:communities   |  2d05h19m00s |       2.48
Graph:write         |  1d17h06m00s |       1.91
(9 entries omitted) |  3d15h44m00s |       4.08
(unmatched time)    | 43d22h02m00s |      49.07

It is pretty cool. But you would agree that colored graphs are much more striking than a plain ascii table? So taking a few ideas from karpathy's ulogme, I wrote this tool to plot my daily stats as logged by arbtt-capture.

Getting Started

Start recording

1. Install arbtt. Run arbtt-capture at startup.
2. Probably let it run for some time to have some data.

Using arbtt-graph

1. Clone/download this repository.
2. Have a look at the categorize.cfg file. Add similar categories to your categorize.cfg file. (Usually at ~/.arbtt/categorize.cfg).
3. Important: Modify path of the file in update.
4. Important: Accordingly modify render/settings.js for barcode categories.
5. Once everything is set up, run update once.
6. Now you can run ./arbtt-serve.py (or python3 arbtt-serve.py) and go to the displayed address. (Default: <http://localhost:9999>)
7. You can refresh anytime using the reload button to the right.

Structure

All html/js UI is inside the render folder. Also no information other than the percentage usage is inside the json files, so you can even make the plots public for extra extrinsic motivation. 😄

Suggestions and pull requests are always welcome.

Thanks

• Joachim Breitner for arbtt and numerous suggestions.
• Andrej Karpathy for ulogme.

Happy tracking!