Add first version of google fonts check

Pipfile

+[[source]]
+url = "https://pypi.org/simple"
+verify_ssl = true
+name = "pypi"
+
+[packages]
+requests = "*"
+beautifulsoup4 = "*"
+
+[dev-packages]
+
+[requires]
+python_version = "3.10"

Pipfile.lock

+{
+    "_meta": {
+        "hash": {
+            "sha256": "73352366ebbba8a1b3952ae48124bd44d7414d5f62dd76c10b00d5772e972461"
+        },
+        "pipfile-spec": 6,
+        "requires": {
+            "python_version": "3.10"
+        },
+        "sources": [
+            {
+                "name": "pypi",
+                "url": "https://pypi.org/simple",
+                "verify_ssl": true
+            }
+        ]
+    },
+    "default": {
+        "beautifulsoup4": {
+            "hashes": [
+                "sha256:58d5c3d29f5a36ffeb94f02f0d786cd53014cf9b3b3951d42e0080d8a9498d30",
+                "sha256:ad9aa55b65ef2808eb405f46cf74df7fcb7044d5cbc26487f96eb2ef2e436693"
+            ],
+            "index": "pypi",
+            "version": "==4.11.1"
+        },
+        "certifi": {
+            "hashes": [
+                "sha256:0d9c601124e5a6ba9712dbc60d9c53c21e34f5f641fe83002317394311bdce14",
+                "sha256:90c1a32f1d68f940488354e36370f6cca89f0f106db09518524c88d6ed83f382"
+            ],
+            "markers": "python_version >= '3.6'",
+            "version": "==2022.9.24"
+        },
+        "charset-normalizer": {
+            "hashes": [
+                "sha256:5a3d016c7c547f69d6f81fb0db9449ce888b418b5b9952cc5e6e66843e9dd845",
+                "sha256:83e9a75d1911279afd89352c68b45348559d1fc0506b054b346651b5e7fee29f"
+            ],
+            "markers": "python_full_version >= '3.6.0'",
+            "version": "==2.1.1"
+        },
+        "idna": {
+            "hashes": [
+                "sha256:814f528e8dead7d329833b91c5faa87d60bf71824cd12a7530b5526063d02cb4",
+                "sha256:90b77e79eaa3eba6de819a0c442c0b4ceefc341a7a2ab77d7562bf49f425c5c2"
+            ],
+            "markers": "python_version >= '3.5'",
+            "version": "==3.4"
+        },
+        "requests": {
+            "hashes": [
+                "sha256:7c5599b102feddaa661c826c56ab4fee28bfd17f5abca1ebbe3e7f19d7c97983",
+                "sha256:8fefa2a1a1365bf5520aac41836fbee479da67864514bdb821f31ce07ce65349"
+            ],
+            "index": "pypi",
+            "version": "==2.28.1"
+        },
+        "soupsieve": {
+            "hashes": [
+                "sha256:3b2503d3c7084a42b1ebd08116e5f81aadfaea95863628c80a3b774a11b7c759",
+                "sha256:fc53893b3da2c33de295667a0e19f078c14bf86544af307354de5fcf12a3f30d"
+            ],
+            "markers": "python_version >= '3.6'",
+            "version": "==2.3.2.post1"
+        },
+        "urllib3": {
+            "hashes": [
+                "sha256:3fa96cf423e6987997fc326ae8df396db2a8b7c667747d47ddd8ecba91f4a74e",
+                "sha256:b930dd878d5a8afb066a637fbb35144fe7901e3b209d1cd4f524bd0e9deee997"
+            ],
+            "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4, 3.5' and python_version < '4'",
+            "version": "==1.26.12"
+        }
+    },
+    "develop": {}
+}

README.md

+# Google Fonts Check
+
+This tool searches for references to Google Fonts dependencies included by
+```html
+<link rel="stylesheet" href="https://fonts.googleapis.com/css?family=[…]">
+```
+for all entries in urls.txt.
+
+This project uses Pipenv to set up the virtual environment for Python 3.
+Use
+```bash
+pipenv run python check.py
+```
+

check.py

+import concurrent.futures
+import re
+from threading import Lock
+
+import requests as r
+from bs4 import BeautifulSoup
+
+DOMAINS_FILE = "domains.txt"
+
+google_fonts_matcher = re.compile(r"https://fonts\.googleapis\.com/css2?\?family=.*")
+lock = Lock()
+tasks_total = 0
+tasks_completed = 0
+
+
+def main() -> None:
+    global tasks_total
+
+    with open(DOMAINS_FILE) as f:
+        with concurrent.futures.ThreadPoolExecutor() as executor:
+
+            futures = [executor.submit(check_google_fonts_usage, domain) for domain in f.read().split()]
+            tasks_total = len(futures)
+            for future in futures:
+                future.add_done_callback(progress_indicator)
+
+            results = [future.result() for future in concurrent.futures.as_completed(futures)]
+
+            for url, _, duration in sorted(filter(lambda x: not x[1], results), key=lambda x: x[1]):
+                if duration < 0:
+                    print(f"{url}: Connection Error")
+                else:
+                    print(f"{url} does not use Google Fonts (took {duration}s)")
+
+            for url, _, duration in sorted(filter(lambda x: x[1], results), key=lambda x: x[1]):
+                print(f"{url} uses Google Fonts (took {duration}s)")
+
+
+def progress_indicator(future: concurrent.futures.Future):
+    global lock, tasks_total, tasks_completed
+
+    with lock:
+        tasks_completed += 1
+        print(f"{tasks_completed}/{tasks_total} completed, {tasks_total - tasks_completed} remain.", end="\r")
+
+
+def check_google_fonts_usage(domain: str) -> (str, str, float):
+    url = f"https://{domain}/"
+    try:
+        response = r.get(url)
+    except r.exceptions.ConnectionError:
+        url = f"http://{domain}/"
+        try:
+            response = r.get(url)
+        except r.exceptions.ConnectionError:
+            return url, False, -1.0
+
+    if response.status_code == r.codes.ok:
+        soup = BeautifulSoup(response.text, "html.parser")
+
+        return url, uses_google_fonts(soup), response.elapsed.total_seconds()
+    else:
+        return url, False, -1.0
+
+
+def uses_google_fonts(soup: BeautifulSoup) -> (str, str):
+    for link in soup("link", rel="stylesheet", href=True):
+        if google_fonts_matcher.match(link.get("href")):
+            return True
+
+    return False
+
+
+if __name__ == "__main__":
+    main()