feat: allow disabling tldextract HTTP requests

namsnath · namsnath · commit 365defa871b0 · 2025-02-17T19:47:55.000+05:30
- Workaround till tldextract#233 is implemented
- Adds flag to control disabling HTTP requests
- Adds `pyfakefs` to use in tests
diff --git a/.gitignore b/.gitignore
@@ -8,7 +8,7 @@ __pycache__
 releasePassword
 apiPassword
 venv/
-env/
+./env/
 .env
 .DS_Store
 bin/
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -8,6 +8,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [unreleased]
 
+## [0.28.1] - 2025-02-17
+- Adds option to disable `tldextract` HTTP calls by setting `SUPERTOKENS_TLDEXTRACT_DISABLE_HTTP=1`
+
 ## [0.28.0]
 - **[Breaking] Updates pre-commit hooks to use `pre-commit`**
   - Migration:
diff --git a/dev-requirements.txt b/dev-requirements.txt
@@ -9,6 +9,7 @@ flask-cors==5.0.0
 nest-asyncio==1.6.0
 pdoc3==0.11.0
 pre-commit==3.5.0
+pyfakefs==5.7.4
 pylint==3.2.7
 pyright==1.1.393
 python-dotenv==1.0.1
diff --git a/setup.py b/setup.py
@@ -83,7 +83,7 @@
 
 setup(
     name="supertokens_python",
-    version="0.28.0",
+    version="0.28.1",
     author="SuperTokens",
     license="Apache 2.0",
     author_email="team@supertokens.com",
diff --git a/supertokens_python/constants.py b/supertokens_python/constants.py
@@ -15,7 +15,7 @@
 from __future__ import annotations
 
 SUPPORTED_CDI_VERSIONS = ["5.2"]
-VERSION = "0.28.0"
+VERSION = "0.28.1"
 TELEMETRY = "/telemetry"
 USER_COUNT = "/users/count"
 USER_DELETE = "/user/remove"
diff --git a/supertokens_python/env/__init__.py b/supertokens_python/env/__init__.py
diff --git a/supertokens_python/env/base.py b/supertokens_python/env/base.py
@@ -0,0 +1,12 @@
+from os import environ
+
+from supertokens_python.env.utils import str_to_bool
+
+
+def FLAG_tldextract_disable_http():
+    """
+    Disable HTTP calls from `tldextract`.
+    """
+    val = environ.get("SUPERTOKENS_TLDEXTRACT_DISABLE_HTTP", "0")
+
+    return str_to_bool(val)
diff --git a/supertokens_python/env/utils.py b/supertokens_python/env/utils.py
@@ -0,0 +1,5 @@
+def str_to_bool(val: str) -> bool:
+    """
+    Convert ENV values to boolean
+    """
+    return val.lower() in ("true", "t", "1")
diff --git a/supertokens_python/utils.py b/supertokens_python/utils.py
@@ -35,8 +35,9 @@
 from urllib.parse import urlparse
 
 from httpx import HTTPStatusError, Response
-from tldextract import extract  # type: ignore
+from tldextract import TLDExtract
 
+from supertokens_python.env.base import FLAG_tldextract_disable_http
 from supertokens_python.framework.django.framework import DjangoFramework
 from supertokens_python.framework.fastapi.framework import FastapiFramework
 from supertokens_python.framework.flask.framework import FlaskFramework
@@ -288,7 +289,16 @@ def get_top_level_domain_for_same_site_resolution(url: str) -> str:
     if hostname.startswith("localhost") or is_an_ip_address(hostname):
         return "localhost"
 
-    parsed_url: Any = extract(hostname, include_psl_private_domains=True)
+    extract = TLDExtract(fallback_to_snapshot=True, include_psl_private_domains=True)
+    # Explicitly disable HTTP calls, use snapshot bundled into library
+    if FLAG_tldextract_disable_http():
+        extract = TLDExtract(
+            suffix_list_urls=(),  # Ensures no HTTP calls
+            fallback_to_snapshot=True,
+            include_psl_private_domains=True,
+        )
+
+    parsed_url: Any = extract(hostname)
     if parsed_url.domain == "":  # type: ignore
         # We need to do this because of https://github.com/supertokens/supertokens-python/issues/394
         if hostname.endswith(".amazonaws.com") and parsed_url.suffix == hostname:
diff --git a/tests/auth-react/fastapi-server/app.py b/tests/auth-react/fastapi-server/app.py
@@ -1117,7 +1117,7 @@ async def exception_handler(a, b):  # type: ignore
     return JSONResponse(status_code=500, content={})
 
 
-app.add_middleware(ExceptionMiddleware, handlers=app.exception_handlers)
+app.add_middleware(ExceptionMiddleware, handlers=app.exception_handlers)  # type: ignore
 
 
 @app.post("/beforeeach")
diff --git a/tests/test_utils.py b/tests/test_utils.py
@@ -1,18 +1,21 @@
+import os
 import threading
+from contextlib import ExitStack
 from typing import Any, Dict, List, Union
+from unittest.mock import patch
 
-import pytest
+from pytest import mark, param, raises
 from supertokens_python.utils import (
     RWMutex,
     get_top_level_domain_for_same_site_resolution,
     humanize_time,
     is_version_gte,
 )
 
-from tests.utils import is_subset
+from tests.utils import is_subset, outputs
 
 
-@pytest.mark.parametrize(
+@mark.parametrize(
     "version,min_minor_version,is_gte",
     [
         (
@@ -72,7 +75,7 @@ def test_util_is_version_gte(version: str, min_minor_version: str, is_gte: bool)
 HOUR = 60 * MINUTE
 
 
-@pytest.mark.parametrize(
+@mark.parametrize(
     "ms,out",
     [
         (1 * SECOND, "1 second"),
@@ -91,7 +94,7 @@ def test_humanize_time(ms: int, out: str):
     assert humanize_time(ms) == out
 
 
-@pytest.mark.parametrize(
+@mark.parametrize(
     "d1,d2,result",
     [
         ({"a": {"b": [1, 2]}, "c": 1}, {"c": 1}, True),
@@ -176,7 +179,7 @@ def balance_is_valid():
     assert actual_balance == expected_balance, "Incorrect account balance"
 
 
-@pytest.mark.parametrize(
+@mark.parametrize(
     "url,res",
     [
         ("http://localhost:3001", "localhost"),
@@ -196,3 +199,41 @@ def balance_is_valid():
 )
 def test_tld_for_same_site(url: str, res: str):
     assert get_top_level_domain_for_same_site_resolution(url) == res
+
+
+@mark.parametrize(
+    ["internet_disabled", "env_val", "expectation"],
+    [
+        param(True, "False", raises(RuntimeError), id="Internet disabled, flag unset"),
+        param(True, "True", outputs("google.com"), id="Internet disabled, flag set"),
+        param(False, "False", outputs("google.com"), id="Internet enabled, flag unset"),
+        param(False, "True", outputs("google.com"), id="Internet enabled, flag set"),
+    ],
+)
+def test_tldextract_http_toggle(
+    internet_disabled: bool,
+    env_val: str,
+    expectation: Any,
+    # pyfakefs fixture, mocks the filesystem
+    # Mocking `tldextract`'s cache path does not work in repeated tests
+    fs: Any,
+):
+    import socket
+
+    # Disable sockets, will raise errors on HTTP calls
+    socket_patch = patch.object(socket, "socket", side_effect=RuntimeError)
+    environ_patch = patch.dict(
+        os.environ,
+        {"SUPERTOKENS_TLDEXTRACT_DISABLE_HTTP": env_val},
+    )
+
+    stack = ExitStack()
+    stack.enter_context(environ_patch)
+    if internet_disabled:
+        stack.enter_context(socket_patch)
+
+    # if `expectation` is raises, checks for raise
+    # if `outputs`, value used in `assert` statement
+    with stack, expectation as expected_output:
+        output = get_top_level_domain_for_same_site_resolution("https://google.com")
+        assert output == expected_output
diff --git a/tests/utils.py b/tests/utils.py
@@ -16,6 +16,7 @@
 
 # Import AsyncMock
 import sys
+from contextlib import contextmanager
 from datetime import datetime
 from http.cookies import SimpleCookie
 from os import environ, kill, remove, scandir
@@ -613,3 +614,17 @@ async def create_users(
             await manually_create_or_update_user(
                 "public", user["provider"], user["userId"], user["email"], True, None
             )
+
+
+@contextmanager
+def outputs(val: Any):
+    """
+    Outputs a value to assert.
+
+    Usage:
+        @mark.parametrize(["input", "expectation"], [(1, outputs(1)), (0, raises(Exception))])
+        def test_fn(input, expectation):
+            with expectation as expected_output:
+                assert 1 / input == expected_output
+    """
+    yield val
