Add issues to OBO Foundry repositories to add `obofoundry` topic

create_obofoundry_issues.py

"""

This script blasts all of the OBO Foundry Repositories with a given issue.
Use sparingly.

 ______________________________ 
/ with great power comes great \
\ responsibility               /
 ------------------------------ 
        \   ^__^
         \  (oo)\_______
            (__)\       )\/\
                ||----w |
                ||     ||

"""

from textwrap import dedent

import pandas as pd
import pystow
import requests
import yaml
from tqdm import tqdm

URL = 'https://raw.githubusercontent.com/OBOFoundry/OBOFoundry.github.io/master/_config.yml'
PREFIX = 'https://github.com/'
PATH = pystow.join('obofoundry', name='obofoundry_topics.tsv')
TITLE = 'Add obofoundry topic to repo metadata'
BODY = dedent(f'''\
The OBO Foundry has recently created a topics page at https://github.com/topics/obofoundry.
It would be great if you could add the `obofoundry` topic to your repository, since
it's currently listed as active in the OBO Foundry. If you're not sure how to do that,
follow the instructions [here](https://docs.github.com/en/github/administering-a-repository/managing-repository-settings/classifying-your-repository-with-topics). 

The main issue for this task is at https://github.com/OBOFoundry/OBOFoundry.github.io/issues/1538
in case you want some more context or to join the larger discussion.
''').replace('\n', ' ')

# Load the GitHub access token via PyStow. We'll
# need it so we don't hit the rate limit
TOKEN = pystow.get_config('github', 'access_token')


def main():
    df = make_topic_sheet()
    # print(tabulate(df.values, headers=list(df.columns), tablefmt='github'))
    post_issues(df)


def post_issues(df):
    for owner, repo in df.loc[~df.has_obofoundry_topic, ['owner', 'repo']].values:
        res = requests.get(
            f'https://api.github.com/repos/{owner}/{repo}/issues',
            headers={
                'Accept': "application/vnd.github.mercy-preview+json",
                'Authorization': f"token {TOKEN}",
            },
        ).json()

        # An issue has already been sent. Don't make duplicates
        if any(issue['title'] == TITLE for issue in res):
            continue

        tqdm.write(f'Creating issue for {owner}/{repo}')

        from github import Github
        g = Github(TOKEN)
        repo = g.get_repo(f'{owner}/{repo}')
        issue = repo.create_issue(title=TITLE, body=BODY)
        tqdm.write(f'success: {issue}')
                
        # probably enough time to avoid the abuse detection trigger
        time.sleep(2)


def make_topic_sheet(force: bool = False) -> pd.DataFrame:
    if PATH.is_file() and not force:
        df = pd.read_csv(PATH, sep='\t')
        df['topics'] = df['topics'].map(lambda x: x if pd.notna(x) else '')
        return df

    # Get the
    res = requests.get(URL)
    res = yaml.safe_load(res.content)

    rows = []
    for record in tqdm(res['ontologies']):
        tracker = record.get('tracker')

        # All active ontologies have trackers. Most of them
        # have GitHub, but not all. Don't consider the non-GitHub
        # ones
        if not tracker or not tracker.startswith(PREFIX):
            continue

        # Since we assume it's a github link, slice out the prefix then
        # parse the owner and repository out of the path
        owner, repo, *_ = tracker[len(PREFIX):].split('/')

        # Get all topics from the repository. See more information on the GH docs:
        # https://docs.github.com/en/rest/reference/repos#get-all-repository-topics
        topics = requests.get(
            f'https://api.github.com/repos/{owner}/{repo}/topics',
            headers={
                'Accept': "application/vnd.github.mercy-preview+json",
                'Authorization': f"token {TOKEN}",
            },
        ).json()['names']

        rows.append((
            record['id'],
            record['title'],
            owner,
            repo,
            ','.join(topics),
            'obofoundry' in topics
        ))

    columns = ['ontology', 'title', 'owner', 'repo', 'topics', 'has_obofoundry_topic']
    df = pd.DataFrame(rows, columns=columns)
    df.to_csv(PATH, sep='\t', index=False)
    return df


if __name__ == '__main__':
    main()