Fetch GitHub repo and run cloc on it.

gh_cloc.py

#!/usr/bin/env python

"""
Find the default branch of a GitHub repository online, 
download its zipped archive, unpack it in a temporary
directory, and run cloc over it and show the output table.

$ python gh_cloc.py zslayton cron
      22 text files.
classified 19 files
      19 unique files.                              
       3 files ignored.

github.com/AlDanial/cloc v 1.96  T=0.02 s (1043.8 files/s, 155853.3 lines/s)
-------------------------------------------------------------------------------
Language                     files          blank        comment           code
-------------------------------------------------------------------------------
Rust                            16            311            195           2240
Markdown                         1              7              0             38
TOML                             1              3              0             19
YAML                             1              5              0             19
-------------------------------------------------------------------------------
SUM:                            19            326            195           2316
-------------------------------------------------------------------------------
"""

import requests
import tempfile
import zipfile
import os
import subprocess
import typer
import shutil

# Constants
REPO_URL = 'https://api.github.com/repos/{owner}/{repo}'
ZIP_URL = 'https://github.com/{owner}/{repo}/archive/refs/heads/{branch}.zip'

app = typer.Typer()

def download_repo(owner, repo):
    # Fetch default branch name
    response = requests.get(REPO_URL.format(owner=owner, repo=repo))
    response.raise_for_status()
    default_branch = response.json()['default_branch']

    # Download the zip archive
    response = requests.get(ZIP_URL.format(owner=owner, repo=repo, branch=default_branch), stream=True)
    response.raise_for_status()

    # Save to a temporary file
    with tempfile.NamedTemporaryFile(delete=False, suffix='.zip') as tmp:
        for chunk in response.iter_content(chunk_size=8192):
            tmp.write(chunk)
        tmp_file_name = tmp.name

    return tmp_file_name, default_branch

def unpack_zip(zip_file):
    # Create a temporary directory
    tmp_dir = tempfile.mkdtemp()

    # Unpack the zip archive
    with zipfile.ZipFile(zip_file, 'r') as zip_ref:
        zip_ref.extractall(tmp_dir)

    return tmp_dir

def run_cloc_on_directory(directory):
    # Run cloc and print the output
    result = subprocess.run(['cloc', directory], capture_output=True, text=True)
    print(result.stdout)

@app.command()
def main(owner: str, repo: str):
    """
    Download and run cloc on a given GitHub repository.

    Args:
    - owner: GitHub username.
    - repo: GitHub repository name.
    """
    zip_file, _ = download_repo(owner, repo)
    directory = unpack_zip(zip_file)
    run_cloc_on_directory(directory)

    # Clean up
    os.remove(zip_file)
    shutil.rmtree(directory)

if __name__ == '__main__':
    app()