Basic changelog

changelog.py

import git

r = git.Repo("/Users/fries2/projects/uvcdat")

t = None
for tag in r.tags:
    if tag.tag:
        tag = tag.tag
    
    if t is None or tag.object.committed_date > t.object.committed_date:
        t = tag

end = t.object.binsha

commit = r.head.commit

prs = []

def traverse_commits(c, end, pull_requests):
    if c.binsha == end:
        return True

    if c.message[:18] == "Merge pull request":
        pull_requests.append(c)

    for parent in c.parents:
        if traverse_commits(parent, end, pull_requests):
            return True
    return False

traverse_commits(commit, end, prs)

import os
import requests
import re

fix_commit_re = re.compile("\d{3,}")

key = os.environ["CHANGELOG_GITHUB_TOKEN"]

headers = {
    "Authorization": "token %s" % key,
    "Accept": "application/vnd.github.v3+json",
    "User-Agent": "chaosphere2112-uvcdat-changelogger",
}


api_base = "https://api.github.com/%s"
pr_base = api_base % "repos/UV-CDAT/uvcdat/pulls/%s"
user_base = api_base % "users/%s"
issue_base = api_base % "repos/UV-CDAT/uvcdat/issues/%s"

pull_data = []

for pr in prs:
    # Get the #PRID part of the message
    pr_id = pr.message[19:].split()[0]
    # Get the actual ID number
    pr_id = pr_id[1:]
    response = requests.get(pr_base % pr_id, headers=headers)
    if response.status_code == 404:
        continue
    pr_info = response.json()
    data = {}
    data["title"] = pr_info["title"]
    match = fix_commit_re.search(data["title"])
    if match is not None:
        # Grab the matching issue
        issue_response = requests.get(issue_base % match.group(0), headers=headers)
        if issue_response.status_code != 404:
            issue_data = issue_response.json()
            data["issue"] = issue_data["html_url"]
            data["issue_title"] = issue_data["title"]
    data["url"] = pr_info["html_url"]
    authors= set()
    author_commits = {}

    pr_commits = requests.get(pr_info["commits_url"], headers=headers)
    commits = pr_commits.json()

    while "next" in pr_commits.links:
        pr_commits = requests.get(pr_commits.links["next"]["url"], headers=headers)
        commits.extend(pr_commits.json())

    for commit in commits:
        author = commit["commit"]["author"]["name"]
        if author not in authors:
            authors.add(author)
            author_commits[author] = 1
        else:
            author_commits[author] += 1

    # Check if author is a github user, deduplicate names
    for author in authors:
        if len(author.split()) == 1:
            # Might be a user, let's check
            user_data = requests.get(user_base % author, headers=headers)
            if user_data.status_code == 404:
                continue
            user = user_data.json()
            user_real_name = user["name"]
            if user_real_name in author_commits:
                author_commits[user_real_name] += author_commits[author]
            else:
                author_commits[user_real_name] = author_commits[author]
            del author_commits[author]

    # Sort authors by commit count
    authors = sorted(author_commits, key=lambda x: author_commits[x])
    authors.reverse()
    data["authors"] = ", ".join(authors)

    if "issue" in data:
        print "* [{title}]({url}) by {authors} ([{issue_title}]({issue}))".format(**data)
    else:
        print "* [{title}]({url}) by {authors}".format(**data)