Chess Conducto Pipeline

A pipeline that'll cobble together some data about my lichess games.

Requirements

Here are the things you need before running:

Some modern linux installation (Ubuntu 18.04 is nice).
python3.7 in your $PATH.
Docker accessible by your current user (docker run --rm hello-world should run successfully).
A conducto account. Go here to create one if you haven't yet: https://conducto.com/app/register.

How To Run

We use conducto to build our pipelines. The first stage of the pipeline pulls all the game data. Subsequent stages just analyize the data, so they run in parallel. All stages print their results to stdout.

To actually run the pipeline:

make pipeline

If you've never created a pipeline before, you'll be prompted to enter your conducto creds. If everything is successful, a browser window should pop up with the conducto UI and your pipeline ready to execute. To execute your pipeline, press the "Play" button on the top left. Expand the parallel analyze-games pipeline to see the actual game analysis.

Run Against Lichess API

Alternatively you can run pulling the newest data directly from lichess:

make pipeline-fresh

NB: If you want to pull fresh from lichess, you'll need to generate a lichess API token and save it as a secret in your conducto account called LICHESS_TOKEN.

Python venv

I use the venv python module to make venvs. On ubuntu there are sometimes weird issues where python complains that the ensurepip library is missing. You can run apt install python3.7-venv to install that missing lib.

Lichess API

Here's the lichess API doc: https://berserk.readthedocs.io/en/master/usage.html

It's pretty easy to use, and only requires an API token. You can generate an API token at https://lichess.org/account/oauth/token.

dgs3 / lichess-pipeline