microsoft / bonsai-cli

A python library for making API calls to Bonsai BRAIN.

Bonsai CLI

A python library for making API calls to Bonsai BRAIN.

Documentation for the Bonsai CLI can be found on the Bonsai Docs website.

Guide to CLI installation and usage: http://docs.bons.ai/guides/cli-guide.html

Full CLI Reference Documentation: http://docs.bons.ai/references/cli-reference.html

Quick Installation

Install the latest stable from PyPI:

$ pip install bonsai-cli

Install the latest in-development version:

$ pip install https://github.com/BonsaiAI/bonsai-cli

bonsai-cli's Issues

Provide an override flag so an unmanaged simulator can be reused in another training.

In the https://github.com/microsoft/bonsai-cli/blob/main/bonsaicli2/bonsai_cli/commands/simulator_unmanaged.py lines 501 to 518 is doing a check if the unmanaged simulator is already registered.
This prevents reusing an unmanaged simulator. Can we have an override flag that allows setting the unmanaged sim to a new brain or new version of the brain even if it was previously assigned to another another brain?

Bonsai commands removed

Only help, config, brain, simulator commands are available. How do we push files without the push command?

Support service principal login

I'd like to autoscale the number of unmanaged simulators (via Kubernetes). As part of that, I would like to be able to automatically use the new simulators for training. The only way I can see right now is to run bonsai simulator unmanaged connect, which requires authentication.

Feature request: service principal logins to authenticate the Bonsai CLI (because I can't possibly log in interactively every time)