update-maker.py

import hashlib
import os
import traceback
import json
import time


def checksum_file(path):
    with open(path, mode='rb') as f:
        return hashlib.md5(f.read()).hexdigest()


def make_index(dir_path):
    result = {}

    for root, dirs, files in os.walk(dir_path):
        for name in files:
            path = os.path.join(root, name)
            result[path] = checksum_file(path)

    return result


def load_index(path):
    try:
        with open(path, mode='r') as f:
            return json.load(f)
    except:
        return {}


def save_index(index, path):
    with open(path, mode='w') as f:
        json.dump(index, f, sort_keys=True, ensure_ascii=False, indent=4)


def diff_index(index1, index2):
    result = {path: None for path in set(index1.keys()) | set(index2.keys())}

    for path, checksum in index2.items():
        if path not in index1 or checksum != index1[path]:
            result[path] = 1

    for path, checksum in index1.items():
        if path not in index2:
            result[path] = 0

    result = {k: v for k, v in result.items() if v is not None}

    return result


def add_diff_to_changelog(changelog, diff):
    changelog[str(int(time.time()))] = diff


def print_diff(diff):
    print('NEW DIFF:')
    print(json.dumps(diff, sort_keys=True, ensure_ascii=False, indent=4))


def load_changelog(path):
    try:
        with open(path, mode='r') as f:
            return json.load(f)
    except:
        return {}


def save_changelog(changelog, path):
    with open(path, mode='w') as f:
        json.dump(changelog, f, sort_keys=True, ensure_ascii=False, indent=4)


def main():
    try:
        source_dir_path = os.sys.argv[1]
        output_index_path = os.sys.argv[2]
        output_changelog_path = os.sys.argv[3]
    except ValueError:
        traceback.print_exc()
        print('USAGE: python3 update-maker.py SOURCE_DIR OUTPUT_INDEX OUTPUT_CHANGELOG')
        exit(1)

    old_index = load_index(output_index_path)
    new_index = make_index(source_dir_path)
    diff = diff_index(old_index, new_index)
    changelog = load_changelog(output_changelog_path)
    add_diff_to_changelog(changelog, diff)

    print_diff(diff)

    if len(diff) != 0:
        save_index(new_index, output_index_path)
        save_changelog(changelog, output_changelog_path)


if __name__ == '__main__':
    main()