Using git username as mlflow run ID

mlflow_git_username.py

import subprocess
from mlflow.utils.mlflow_tags import MLFLOW_USER

def get_username_from_git() -> str:
    """Retrieve a (sanitized) username from git config. Raise a runtime error if the git command fails."""
    try:
        cmd = subprocess.run(['git', 'config', 'user.name'], stdout=subprocess.PIPE, stderr=subprocess.PIPE)
        if len(cmd.stderr.strip()) > 0:
            raise ValueError(cmd.stderr.decode('utf-8').strip())
        raw_name = cmd.stdout.decode('utf-8')
        user_name = "".join(x for x in raw_name.strip().lower().replace(" ", '_') if x.isalnum() or x in '-_')
        if not user_name:
            raise ValueError("invalid user name")
    except (FileNotFoundError, ValueError) as e:
        raise RuntimeError(f"Could not retrieve name from git config.\n error={e}")

    return user_name

job_id = ...
user = get_username_from_git()

mlflow.start_run(run_name=job_id)
mlflow.set_tag(MLFLOW_USER, user)