stats_contribution.py

# TODO : cpprefjp以外にも対応する
import glob
import re
import os
import sys
from subprocess import Popen

point_dict = {
    "cpprefjp/typo": 1,
    "cpprefjp/link": 2,
    "cpprefjp/addref": 20,
    "cpprefjp/addlang": 20,
    "cpprefjp/addpage": 20,
    "cpprefjp/fixs": 2,
    "cpprefjp/fixm": 5,
    "cpprefjp/fixl": 10,
    "cpprefjp/compiler": 2,
    "boostjp/typo": 1,
    "boostjp/releases": 2,
    "boostjp/releasem": 5,
    "boostjp/releasem": 10,
    "boostjp/releasel": 20,
    "boostjp/addrefs": 10,
    "boostjp/addrefm": 20,
    "boostjp/boosts": 5,
    "tool/fixbug": 30,
    "tool/improves": 10,
    "tool/improvem": 30,
    "tool/improvem": 50,
    "tool/updatelib": 20,
    "tool/updatelang": 10,
    "tool/updatelang": 30,
    "tool/updatelang": 50,
    "tool/adds": 30,
    "tool/addm": 50,
    "tool/addl": 100,
}

def stats_contribution(text: str, filename: str) -> set[str]:
    user_name: str | None = None
    user_point = 0
    commit_set = set()

    users = dict()
    for line in text.split("\n"):
        m = re.fullmatch(r'## (.*?)', line)
        if m:
            user_name = m.group(1)
            user_point = 0
            continue

        if not user_name:
            continue

        if line.startswith("| ["):
            cols = line.split("|")

            commits = cols[1].split(",")
            for commit in commits:
                m = re.fullmatch(r'\[(.*?)\]\((.*?)/commit/(.*?)\)', commit.strip())
                full_commit_id = m.group(3)
                commit_set.add(full_commit_id)

            points = cols[2].split(",")
            for point in points:
                point_values = point.split(":")
                point_name = point_values[0].strip()
                point_quantity = int(point_values[1].strip())
                point_value = point_dict.get(point_name)
                if not point_value:
                    raise KeyError("{}: invalid point tag `{}`".format(filename, point_name))
                user_point += point_value * point_quantity
            users[user_name] = user_point

    for name, point in sorted(users.items(), key=lambda item: item[1], reverse=True):
        print("{}: {}".format(name, point))
    return commit_set

def check_commit_set(commit_set: set[str]) -> None:
    command = "git log --after \'2023-01-01\' --pretty=oneline --no-merges".split(" ")
    proc = Popen(command, stdin=-1,stdout=-1,stderr=-1)
    out, err = proc.communicate()
    if len(err) > 0:
        print(err)
        return

    commit_log_set = set()
    for line in out.decode().split("\n"):
        if len(line) <= 0:
            continue
        cols = line.split(" ")
        full_commit_id = cols[0]
        commit_log_set.add(full_commit_id)

    diff = commit_log_set - commit_set
    if len(diff) > 0:
        print("unstats commits: {}".format(diff))

if __name__ == '__main__':
    commit_set = set()
    for p in sorted(list(glob.glob("start_editing/*.md", recursive=True))):
        filename = os.path.basename(p)
        if not filename.startswith("contribution_stats_"):
            continue

        with open(p) as f:
            text = f.read()

        commit_set = commit_set.union(stats_contribution(text, p))

    check_commit_set(commit_set)