HN Explorer

Show HN: Logis – a searchable scientific log in your Git commit history (github.com)

6 points by rorytbyrne ↗ HN
I built this tool that auto-commits your code when you run an experiment, storing structured metadata in the commit message. A bit like `wandb` but local.

There is a small CLI that lets you query for (e.g.) `metrics.accuracy > 0.9`, to find the code state that produced a good results.

This is the first release, and there are quite a few planned features if anyone wants to submit a PR.

3 comments

[ 3.1 ms ] story [ 13.7 ms ] thread
madalinab_
Looks really useful! What's the advantage of having this done locally, aside from being free and not requiring a subscription?
rorytbyrne
Thank you!

The immediate benefit is that "what commit produced my last good result" is a local problem, so a local solution is faster and more convenient than a round-trip to an API.

But more broadly, storing data locally makes it easy for others to build tools on top of the data. E.g. a visualisation tool that plots hyperparameters vs metrics, by parsing the metadata from `logis` commits.

Also scientists often don't commit regularly, so there is an opportunity to hijack the commit log for their benefit. Every cycle of edit->run-experiment is an iteration of the scientific method, and capturing that automatically is arguably more valuable than sporadic commits with "stuff" as the message.