ML Flow Integration in Moonshot Container

Feedback
#1

Hi,
I wanted to know if there would be a way to get support for ML Flow integration into the Moonshot container specifically around the autolog() feature. This would allow tracking the training itself automatically to an ML Flow Server which would be very useful. Right now I can only track the outputs and hyper parameters easily from the Jupyter container to my ML Flow server.

I was wondering would their be an option to open a PR to GitHub - quantrocket-llc/moonshot: Vectorized backtester and trading engine for QuantRocket if this is what backs the moonshot container in the docker compose yml.

#2

If you’re able to set it up the way you want in the Jupyter container, you should be able to do the same in the Moonshot container. Have you installed the mlflow package in the Moonshot container as described in the docs? Once installed, you can call mlflow.autolog() from your strategy code that runs inside the container.