Convert Google Fonts Protobuf to JSON

google-fonts-protobuf-to-json.py

"""
This will read through the Google Fonts repo and make unified JSON of it
"""

import os
import sys
import io
from collections import defaultdict
import json

def is_font_dir(path, font_folders):
    for f in os.listdir(path):
        f_path = os.path.join(path, f)

        if os.path.isfile(f_path) and f == "METADATA.pb":
            font_folders.append(path)
            return
        elif os.path.isdir(f_path):
            is_font_dir(f_path, font_folders)
        
def metadata_to_json(metadata_file_path):

    lines = io.open(metadata_file_path, encoding="utf-8").readlines()
    font_json = defaultdict(list)
    nested = False
    nested_json = defaultdict(list)
    for line in lines:
        line = line.strip()

        if ":" in line:
            key, value = line.split(":")[0:2]
            if nested is True:
                nested_json[key].append(value)
            else:
                font_json[key].append(value)

        elif " {" in line:
            nested = True
            nested_key, _ = line.split(" ")
            
        elif "}" in line:
            nested = False
            font_json[nested_key].append(nested_json)
            nested_json = defaultdict(list)
    
    return font_json

def clean(val):
    if type(val) is not str:
        return val

    val = val.strip()
    if val.startswith("\"") and val.endswith("\""):
        val = val[1:-1]
    
    return val.strip()

def simplify_json(font_json):

    for k,v in font_json.items():

        for i, val in enumerate(v):
            if type(val) is dict or type(val) is defaultdict:
                font_json[k][i] = simplify_json(val)
            else:
                font_json[k][i] = clean(val)

    for k,v in font_json.items():
        if len(v) == 1 and k != "fonts":
            font_json[k] = clean(v[0])
            
    
    return font_json

if __name__ == "__main__":

    if len(sys.argv) < 2:
        print("Please provide path to Google Fonts repo")    
        exit(0)

    # Get Google Fonts directory
    gfonts_path = sys.argv[1]

    # Recurse and find all fonts directory
    font_folders = []
    is_font_dir(gfonts_path, font_folders)

    print("Found {} font folders.".format(len(font_folders)))
    print("Parsing...")


    fonts_info = []
    with io.open("fonts.json", "w", encoding="utf-8") as f:
        for folder in font_folders:
            metadata_path = os.path.join(folder, "METADATA.pb")
            font_json = metadata_to_json(metadata_path)
            font_json = simplify_json(font_json)
            font_json["files_path"] = folder
            fonts_info.append(font_json)
        json.dump({"info": fonts_info}, f)

    print("Fonts info written to fonts.json")

Google Fonts repo unfortunately stopped keeping metadata as JSON.
And I do not know how to parse Protobuf natively in Python.

This scripts converts all the metadata into one big nice JSON.
Clone https://github.com/google/fonts and then run

python google-fonts-protobuf-to-json.py path-to-google-fonts-directory

It will create fonts.json in the current directory if if all went well which will look like:

{
  "info": [
    {
      "fonts": [
        {
          "post_script_name": "CraftyGirls-Regular",
          "style": "normal",
          "name": "Crafty Girls",
          "full_name": "Crafty Girls Regular",
          "copyright": "Copyright (c) 2010 by Font Diner, Inc DBA Tart Workshop. All rights reserved.",
          "filename": "CraftyGirls-Regular.ttf",
          "weight": "400"
        }
      ],
      "designer": "Tart Workshop",
      "name": "Crafty Girls",
      "date_added": "2011-01-06",
      "category": "HANDWRITING",
      "subsets": [
        "latin",
        "menu"
      ],
      "license": "APACHE2",
      "files_path": "google-fonts/apache/craftygirls"
    },
    {
      "fonts": [
        {
          "post_script_name": "Arimo-Regular",
          "style": "normal",
          "name": "Arimo",
          "full_name": "Arimo Regular",
          "copyright": "Digitized data copyright (c) 2010-2012 Google Corporation.",
          "filename": "Arimo-Regular.ttf",
          "weight": "400"
        },
        {
          "post_script_name": "Arimo-Italic",
          "style": "italic",
          "name": "Arimo",
          "full_name": "Arimo Italic",
          "copyright": "Digitized data copyright (c) 2010-2012 Google Corporation.",
          "filename": "Arimo-Italic.ttf",
          "weight": "400"
        },
        {
          "post_script_name": "Arimo-Bold",
          "style": "normal",
          "name": "Arimo",
          "full_name": "Arimo Bold"