About CLI

[dibetgonzalez]

Search before asking

• I have searched the AnomaVision issues, discussions, and documentation and found no similar questions

Question Category

Installation & Setup

Question

I can't see the CLI commands.
I make git clone and i use into a docker-compose service
Any help here?

Operating System

None

Package Manager Used

None

Target Hardware / Extra Bracket

None

Dataset Size/Type

No response

Code Example

No response

Expected vs Actual Behavior

No response

Additional Context

No response

[github-actions]

👋 Hello @dibetgonzalez, thank you for opening this issue! Welcome to the AnomaVision community.

Our triage bot has categorized this as a Question. A community member or maintainer will jump in to help as soon as possible.

Please hold tight while we route this to the right team member. 🚀

[DeepKnowledge1]

Hi @dibetgonzalez ,
The cli commands are mentioned in the readme file

https://github.com/DeepKnowledge1/AnomaVision#cli

Also have a look in tye cli file

https://github.com/DeepKnowledge1/AnomaVision/blob/main/anomavision/cli.py

If you have any questions, please let me know

[DeepKnowledge1]

@dibetgonzalez

Thanks for the report! This is a packaging/Docker setup issue, not a bug in the CLI itself.

Root Cause

The anomavision command is registered as an entry point in pyproject.toml, but entry points only become available after the package is installed. Simply cloning the repo and mounting it as a volume is not enough.

Fix

Make sure your Dockerfile installs the package:

COPY . /app
WORKDIR /app
RUN pip install -e ".[cpu]"  # or cu118 / cu121 / cu124 for GPU

If you are using uv, the binary lands in .venv/bin which is not on PATH by default:

RUN uv sync --extra cpu
ENV PATH="/app/.venv/bin:$PATH"

Watch out for volume mounts

If your docker-compose.yml mounts the project directory over /app, it can shadow the installation. In that case, either use the Python module fallback:

python -m anomavision.cli train --config config.yml

or ensure pip install -e runs after the volume is mounted (e.g. as the container entrypoint).

Quick diagnostic

Run this inside the container to see what is happening:

which anomavision        # should print a path
pip show anomavision     # confirms installation
python -m anomavision.cli --help  # always works as a fallback

Please share your Dockerfile and docker-compose.yml if the above does not resolve it and we can take a closer look.

[DeepKnowledge1]

@dibetgonzalez

I also noted one additional thing worth mentioning in the reply — the GitStatusChecker running at module level in train.py will likely cause problems in Docker even after the install is fixed, since it tries to run git fetch inside the container on every import. Worth flagging to the user so they don't hit a second issue right after resolving the first.

I have fixed it right away

[DeepKnowledge1]

@dibetgonzalez

I’ll close this issue for now, but if it’s still not working, feel free to reopen it.