css-resource-embedder.py

import asyncio
from base64 import urlsafe_b64encode
from hashlib import sha384
import os

import aiofiles
from aiohttp.client import ClientSession

# Fake user agent to trick the google font api to give us the .woff2 fonts
# Retrive one by typing `navigator.userAgent` in the browser's console
UA = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/132.0.0.0 Safari/537.36'


async def download_css(url: str, session: ClientSession) -> str:
  async with session.get(url, headers={'User-Agent': UA}) as response:
    return await response.text()


async def download_resource(
  url: str, session: ClientSession, dir_prefix: str, url_prefix: str
) -> str:
  url = url.strip()
  while len(url) and url[0] == url[-1] and url[0] in "'\"":
    url = url[1:-1]

  async with session.get(url, headers={'User-Agent': UA}) as response:
    font = await response.read()

    # Create a 192-bit hash by xor-ing the two halves of a 384-bit hash
    hash = sha384(font).digest()
    h1 = int.from_bytes(hash[:24], byteorder='big')
    h2 = int.from_bytes(hash[24:], byteorder='big')
    h = (h1 ^ h2).to_bytes(24, byteorder='big')

    subtype, ext = response.content_type.split(';')[0].split('/')
    name = urlsafe_b64encode(h).decode('ascii')

    prefix = subtype + 's'
    filename = f'{prefix}/{name}.{ext}'

    os.makedirs(dir_prefix + prefix, exist_ok=True)

    async with aiofiles.open(dir_prefix + filename, 'wb') as f:
      await f.write(font)

    return url_prefix + filename


async def main(
  fonts_url: list[str], dir_prefix: str, url_prefix: str, output: str
) -> None:
  async with ClientSession() as session:
    css_tasks = [download_css(url, session) for url in fonts_url]

    css = ''.join(await asyncio.gather(*css_tasks))
    segments: list[str] = []
    urls: list[str] = []

    # Extract anything between these two patterns
    pat_start = 'url('
    pat_end   = ')'

    while (start := css.find(pat_start)) != -1:
      start += len(pat_start)

      segments.append(css[:start])
      end = css.find(pat_end, start)

      urls.append(css[start:end])
      css = css[end:]

    segments.append(css)

    resource_tasks = [
      download_resource(url, session, dir_prefix, url_prefix)
      for url in urls
    ]

    filenames = await asyncio.gather(*resource_tasks)
    result = ''.join(
      f'{segment}{filename}'
      for segment, filename in zip(segments, filenames)
    ) + segments[-1]

    async with aiofiles.open(output, 'w') as f:
      await f.write(result)


DIR_PREFIX = './'
URL_PREFIX = '/'

asyncio.run(main([
  'https://fonts.googleapis.com/css2?family=Open+Sans:ital@0;1&display=swap',
  'https://fonts.googleapis.com/css2?family=Open+Sans:ital,wght@0,700;1,700&display=swap',
  'https://fonts.googleapis.com/css2?family=Merriweather+Sans:ital,wght@0,700;1,700&display=swap',
  'https://fonts.googleapis.com/css2?family=Fira+Code:wght@500&display=swap',
], DIR_PREFIX, URL_PREFIX, 'fonts.css'))