size.py

#!/usr/bin/env python3

import datetime as dt
import shlex
import subprocess


def get_files(commit):
    hashes = []

    cmd = ('git', 'ls-tree', '-r', commit)
    proc = subprocess.run(cmd, capture_output=True, text=True)
    for line in proc.stdout.strip().split('\n'):
        stats, path = line.split('\t')
        _, _, object_hash = stats.split(' ')
        hashes.append(object_hash)

    return hashes


def get_size(object_hash):
    cmd = ('git', 'cat-file', '-s', object_hash)
    proc = subprocess.run(cmd, capture_output=True, text=True)
    return int(proc.stdout.strip())


def get_charcount(object_hash):
    # TODO unused, but more precise results (ignores indent etc)
    cmd_pattern = "git cat-file blob {} | tr '\\n' ' ' | sed -r 's/[ \\-]+/ /g' | wc -m"
    cmd = cmd_pattern.format(shlex.quote(object_hash))
    proc = subprocess.run(cmd, capture_output=True, text=True, shell=True)
    return int(proc.stdout.strip())


cmd = ('git', 'log', '--pretty=%H\t%at', '--reverse')
proc = subprocess.run(cmd, capture_output=True, text=True)
lines = proc.stdout.strip().split('\n')

for commit_data in lines:
    commit, timestamp_str = commit_data.strip().split('\t')
    author_datetime = dt.datetime.fromtimestamp(int(timestamp_str))

    files = get_files(commit)
    total_size = sum([get_size(file) for file in files])

    print(author_datetime, total_size, sep='\t')