Get list of PRs for release, save as cvs file

get_prs.py

"""
Must run with github token (see get_token())
Also can simply  load it into GITHUB_AUTH env var.
"""
import csv 
import datetime
from collections import namedtuple
from pathlib import Path

from github import Github
from github.PullRequest import PullRequest
from packaging.version import Version

from typing import (
    List,
    Optional,
    Set,
)

PROJECT_OWNER = "galaxyproject"
PROJECT_NAME = "galaxy"
PRItem = namedtuple("PR", ["number", "title", "url", "is_highlight", "is_uiux", "is_manual", "admin_required", "labels_kind", "labels_area", "labels_other", "author"])

MILESTONE = "24.2"
STATE = "closed"


def run():
    version = Version("24.2")
    release_date = datetime.date(2025, 1, 15)   
    items = []
    title_row = [field for field in PRItem._fields]
    items.append(title_row)
    for pr in _get_prs(version, release_date):
        labels = pr.labels
        if "merge" in (l.name for l in labels):  # ignore merge PRs
            continue
        is_manual = _to_y("[x] Instructions for manual testing" in pr.body)
        is_uiux = _to_y("area/UI-UX" in (l.name for l in labels))
        is_highlight = _to_y(any(x for x in labels if x.name.startswith("highlight")))
        labels_kind, labels_area, labels_other = _get_pr_labels(labels)

        item = PRItem(
            number = pr.number,
            title = pr.title,
            url = pr.html_url,
            is_highlight = is_highlight,
            is_uiux = is_uiux,
            is_manual = is_manual,
            admin_required = None,
            labels_kind = labels_kind,
            labels_area = labels_area,
            labels_other = labels_other,
            author = pr.user.login,
        )
        items.append(item)
    write_csv(items)

def _to_y(value):
    return "y" if value else None

def _get_prs(release_version: Version, release_date: datetime.date, state: str = "closed") -> List[PullRequest]:
    github = Github(_get_token())
    repo = github.get_repo(f"{PROJECT_OWNER}/{PROJECT_NAME}")

    _cutoff_date = release_date.replace(year=release_date.year - 1)
    cutoff_time = datetime.datetime.combine(_cutoff_date, datetime.time.min)

    prs: List[PullRequest] = []
    counter = 0 
    print("Collecting relevant pull requests...")
    for pr in repo.get_pulls(state=state, sort="updated", direction="desc"):
        assert pr.updated_at
           if pr.updated_at.replace(tzinfo=None) < cutoff_time:
            break
        counter += 1
        if counter % 100 == 0:
            print(
                f"Examined {counter} PRs; collected {len(prs)} (currently on #{pr.number} updated on {pr.updated_at.date()})"
            )
        # Select PRs that are merged + have correct milestone + have not been previously collected and added to the prs file
        proper_state = state != "closed" or pr.merged_at  # open PRs or PRs that have been merged
        if proper_state and pr.milestone and pr.milestone.title == str(release_version):
            prs.append(pr)

    print(f"Collected {len(prs)} pull requests")
    return prs


def write_csv(items):
    with open('prs.csv', 'w', newline='') as f:
        writer = csv.writer(f, dialect='excel-tab')
        for item in items:
            writer.writerow(item)


def _get_pr_labels(labels):
    kind, area, other = [], [], []
    for label in labels:
        if label.name.startswith("kind"):
            kind.append(label.name[5:])
        elif label.name.startswith("area"):
            area.append(label.name[5:])
        else:
            other.append(label.name)
    return _stringify(kind), _stringify(area), _stringify(other)


def _stringify(items):
    return ", ".join(items)


def _get_token():
    path = Path.home() / '.github'
    with open(path, 'r') as f:
        return f.read().strip()


if __name__ == '__main__':
    run()