diff asset manifests

diff_asset_manifests.py

import codecs
import json
import os
import sys
from os import path
from urllib import urlopen

asset_diff_dir = '/tmp/_asset_diff'

def load_file(file):
    if file.startswith('gs://'):
        file_name = path.basename(file)
        dest_path = '{}/{}'.format(asset_diff_dir, file_name)
        cmd = 'gsutil cp {} {}'.format(file, dest_path)
        print('# {}'.format(cmd))
        os.system(cmd)
        file = dest_path

    with open(file, 'r') as f:
        return json.load(f)

def unminify_str(minified_str):
    # Basically re-indent by parsing {}; characters
    chars = list(minified_str)
    indent = 0
    for i, char in enumerate(chars):
        if char == ';' or char == '{' or char == '}':
            if char == '{':
                chars[i] = '\n' + ('    ' * indent) + char
                indent += 1

            elif char == '}':
                indent -= 1
                chars[i] = '\n' + ('    ' * indent) + char

            chars[i] += '\n' + ('    ' * indent)

    return ''.join(chars)


def write_unminified_assets(file_name, changed_assets, asset_paths):
    dir_path = path.join(asset_diff_dir, path.splitext(path.basename(file_name))[0])
    if not os.path.exists(dir_path):
        os.makedirs(dir_path)

    for asset_url in changed_assets:
        asset_path = asset_paths[asset_url]
        asset_path = asset_path.replace('/', '_')
        asset_path = asset_path.replace('compiled_', '') # remove compiled prefix
        asset_url = 'https:' + asset_url

        asset_dir_path = path.join(dir_path, asset_path)
        if not os.path.exists(asset_dir_path):
            print('Downloading {}'.format(asset_url))
            asset_content = urlopen(asset_url).read().decode('utf-8', errors='ignore')

            if asset_path.endswith('.min.js') or asset_path.endswith('.min.css'):
                asset_content = unminify_str(asset_content)

            with codecs.open(asset_dir_path, 'w', 'utf-8') as f:
                f.write(asset_content)

    return dir_path

def diff_manifests(manifest_old_file, manifest_new_file):
    manifest_files = [manifest_old_file, manifest_new_file]
    manifests = list(map(load_file, manifest_files))
    inverted = list(map(lambda m: dict([[v,k] for k,v in m.items()]), manifests))
    manifest_assets = list(map(lambda m: set(m.keys()), inverted))

    same_assets = manifest_assets[0].intersection(manifest_assets[1])
    changed_assets = list(map(lambda assets: assets - same_assets, manifest_assets))

    dir_paths = [
        write_unminified_assets(manifest_files[0], changed_assets[0], inverted[0]),
        write_unminified_assets(manifest_files[1], changed_assets[1], inverted[1]),
    ]

    print('\nChanged Assets:')
    print('\n'.join(changed_assets[0]))
    return dir_paths

if __name__ == '__main__':
    if len(sys.argv) < 3:
        print('Usage: {} manifest_old.json manifest_new.json'.format(sys.argv[0]))
        sys.exit(1)

    dir_paths = diff_manifests(sys.argv[1], sys.argv[2])
    diff_path = path.join(asset_diff_dir, 'asset_manifests.diff')
    cmd = 'diff -urb {dir1} {dir2} > {diff_path}'.format(
        **{'dir1': dir_paths[0], 'dir2': dir_paths[1], 'diff_path': diff_path}
    )
    print('# {}'.format(cmd))
    os.system(cmd)
    print('\nDiff saved at {}'.format(diff_path))