Merge pull request #428 from ThePhaseless/fix/per-request-geo-lookup

fix: resolve the browser timezone once per process instead of per request

README.md

@@ -18,6 +18,7 @@
 | `PROXY_PASSWORD`     | None      | Password for proxy authentication.                                                                                                                                |
 | `OWUI_API_KEY`       | None      | Bearer token for `/load` endpoint authentication. Must match `EXTERNAL_WEB_LOADER_API_KEY` in Open WebUI.                                                         |
 | `BROWSER_LOCALE`     | None      | Override the browser's language with a [BCP-47](https://www.rfc-editor.org/rfc/bcp/bcp47.txt) tag, e.g. `en-US`, `de-DE`, `fr-FR`. When unset, the locale is derived from the egress country. |
+| `BROWSER_TIMEZONE`   | None      | Override the browser's timezone with an [IANA](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones) name, e.g. `Europe/Berlin`, `Asia/Shanghai`. When unset, the timezone is derived from the egress IP. |
 
 #### Browser language
 
@@ -25,6 +26,14 @@ Set `BROWSER_LOCALE` to a [BCP-47](https://www.rfc-editor.org/rfc/bcp/bcp47.txt)
 
 Valid tags are maintained in the [IANA Language Subtag Registry](https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry). For a friendlier list, see [List of ISO 639-1 codes](https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes) (language) combined with an [ISO 3166-1 alpha-2](https://en.wikipedia.org/wiki/ISO_3166-1_alpha-2) region code for the full tag, e.g. `pt-BR`.
 
+#### Browser timezone
+
+Set `BROWSER_TIMEZONE` to an [IANA timezone name](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones) like `Europe/Berlin` or `Asia/Shanghai` to fix the browser's timezone. An explicit value is used as is, also behind a proxy.
+
+When either `BROWSER_TIMEZONE` or `BROWSER_LOCALE` is unset, Byparr looks up its public IP through `api.ipify.org`, `icanhazip.com` or `checkip.amazonaws.com` and maps it to a timezone and language with a GeoIP database downloaded from GitHub. Without a proxy this happens once, on the first request, and the result is reused until Byparr restarts (restart it if its public IP moves to another country). If that lookup fails, the host timezone (`TZ`, otherwise the system's, otherwise UTC) and `en-US` are used, and the lookup is retried after 10 minutes. With a proxy (`PROXY_SERVER` or the `X-Proxy-Server` header) it runs for every request, through the proxy.
+
+If those hosts are slow or blocked on your network, set both `BROWSER_TIMEZONE` and `BROWSER_LOCALE`: the lookup is then skipped entirely and no request is made before the browser starts.
+
 ## Proxy Recommendation
 
 Recently I've partnered with a _new in town_ proxy service - ProxyBase - to offer affordable proxy services that seems to work seamlessly with Byparr! Using my affiliate code `byparr` (case sensitive!) when signing up will not only get you access to their cost-effective (**$0.69/GB with occasional promotions** _at the time of writing_) proxy network but will also help support the continued development of this project. ProxyBase's proxies can significantly improve your success rate when bypassing anti-bot challenges. [Check out ProxyBase](https://client.proxybase.org/signup?ref=byparr) and enhance your Byparr experience!

pyproject.toml

@@ -15,6 +15,7 @@ dependencies = [
     "pydantic==2.*",
     "pydantic-settings==2.*",
     "trafilatura==2.3.*",
+    "tzlocal==5.*",
 ]
 urls = { repository = "https://github.com/ThePhaseless/Byparr" }
 [dependency-groups]

src/consts.py

@@ -20,6 +20,7 @@ class Settings(BaseSettings):
     return_only_cookies: bool = False
     owui_api_key: str | None = None
     browser_locale: str | None = None
+    browser_timezone: str | None = None
 
 
 settings = Settings()
@@ -39,3 +40,4 @@ RETURN_ONLY_COOKIES = settings.return_only_cookies
 
 OWUI_API_KEY = settings.owui_api_key
 BROWSER_LOCALE = settings.browser_locale
+BROWSER_TIMEZONE = settings.browser_timezone

src/utils.py

@@ -1,15 +1,20 @@
+import asyncio
 import logging
+import math
 import time
 from collections.abc import AsyncGenerator
 from typing import Annotated, NamedTuple, cast
 
 from fastapi import Header
+from invisible_core import prepare_session_geo
 from invisible_playwright.async_api import InvisiblePlaywright
 from playwright.async_api import Browser, BrowserContext, Page
 from pydantic import BaseModel, Field
+from tzlocal import get_localzone_name
 
 from src.consts import (
     BROWSER_LOCALE,
+    BROWSER_TIMEZONE,
     LOG_LEVEL,
     PROXY_PASSWORD,
     PROXY_SERVER,
@@ -52,6 +57,46 @@ class BrowserDepClass(NamedTuple):
     context: BrowserContext
 
 
+GEO_RETRY_SECONDS = 600
+
+
+class BrowserGeo(NamedTuple):
+    timezone: str
+    locale: str
+    expires_at: float = math.inf
+
+
+_geo_lock = asyncio.Lock()
+_geo: BrowserGeo | None = None
+
+
+def resolve_browser_geo() -> BrowserGeo:
+    """Resolve the timezone and locale of a session without a proxy."""
+    geo = prepare_session_geo(BROWSER_TIMEZONE or "", None, BROWSER_LOCALE or "auto")
+    locale = BROWSER_LOCALE or (geo.locale.primary if geo.locale else "en-US")
+    if geo.timezone:
+        logger.info("Browser timezone %s, locale %s", geo.timezone, locale)
+        return BrowserGeo(geo.timezone, locale)
+    timezone = get_localzone_name() or "UTC"
+    logger.warning(
+        "Could not resolve the timezone from the egress IP, using %s and locale %s "
+        "for %d seconds. Set BROWSER_TIMEZONE and BROWSER_LOCALE to skip the lookup.",
+        timezone,
+        locale,
+        GEO_RETRY_SECONDS,
+    )
+    return BrowserGeo(timezone, locale, time.monotonic() + GEO_RETRY_SECONDS)
+
+
+async def get_browser_geo() -> BrowserGeo:
+    """Per-process cache of resolve_browser_geo()."""
+    global _geo  # noqa: PLW0603
+    async with _geo_lock:
+        if _geo is None or time.monotonic() >= _geo.expires_at:
+            _geo = await asyncio.to_thread(resolve_browser_geo)
+        return _geo
+
+
 async def get_browser(
     x_proxy_server: Annotated[
         str | None,
@@ -79,6 +124,8 @@ async def get_browser(
     header_password = x_proxy_password
 
     proxy_config = None
+    timezone = BROWSER_TIMEZONE or ""
+    locale = BROWSER_LOCALE or "auto"
 
     if header_server:
         proxy_config = {
@@ -92,12 +139,15 @@ async def get_browser(
             "username": PROXY_USERNAME,
             "password": PROXY_PASSWORD,
         }
+    else:
+        timezone, locale, _ = await get_browser_geo()
 
     async with InvisiblePlaywright(
         headless=True,
         proxy=proxy_config,
         humanize=True,
-        locale=BROWSER_LOCALE or "auto",
+        locale=locale,
+        timezone=timezone,
         extra_prefs={
             "devtools.jsonview.enabled": False,
             "browser.tabs.remote.useCrossOriginOpenerPolicy": False,

tests/utils_test.py

@@ -0,0 +1,118 @@
+import asyncio
+import math
+from unittest.mock import AsyncMock, MagicMock, patch
+
+import pytest
+from invisible_core import SessionGeo, SessionLocale
+
+from src import utils
+
+BERLIN = SessionGeo(
+    "Europe/Berlin",
+    "203.0.113.7",
+    locale=SessionLocale(languages=("de", "en-US", "en"), region="DE"),
+)
+FAILED = SessionGeo(
+    "", None, locale=SessionLocale(languages=("en-US", "en"), region="US")
+)
+
+
+@pytest.fixture(autouse=True)
+def fresh_geo_cache(monkeypatch: pytest.MonkeyPatch):
+    """Every test starts with nothing resolved."""
+    monkeypatch.setattr(utils, "_geo", None)
+    monkeypatch.setattr(utils, "_geo_lock", asyncio.Lock())
+
+
+@pytest.mark.asyncio
+async def test_egress_lookup_runs_once_per_process():
+    """Concurrent first requests and every later one share a single lookup."""
+    with patch("src.utils.prepare_session_geo", return_value=BERLIN) as lookup:
+        results = await asyncio.gather(*(utils.get_browser_geo() for _ in range(3)))
+        results.append(await utils.get_browser_geo())
+
+    lookup.assert_called_once_with("", None, "auto")
+    assert set(results) == {utils.BrowserGeo("Europe/Berlin", "de", math.inf)}
+
+
+@pytest.mark.asyncio
+async def test_failed_lookup_uses_host_timezone_then_retries(
+    monkeypatch: pytest.MonkeyPatch,
+):
+    """A failed lookup is not repaid per request, but is retried later."""
+    with (
+        patch("src.utils.get_localzone_name", return_value="Asia/Shanghai"),
+        patch("src.utils.prepare_session_geo", return_value=FAILED) as lookup,
+    ):
+        geo = await utils.get_browser_geo()
+        assert (geo.timezone, geo.locale) == ("Asia/Shanghai", "en-US")
+        await utils.get_browser_geo()
+        assert lookup.call_count == 1
+
+        monkeypatch.setattr(utils, "_geo", geo._replace(expires_at=0))
+        await utils.get_browser_geo()
+        assert lookup.call_count == 2  # noqa: PLR2004
+
+
+@pytest.mark.asyncio
+async def test_failed_lookup_without_a_host_zone_uses_utc():
+    """No zone from the host either is what the Docker image has: UTC."""
+    with (
+        patch("src.utils.get_localzone_name", return_value=None),
+        patch("src.utils.prepare_session_geo", return_value=FAILED),
+    ):
+        geo = await utils.get_browser_geo()
+
+    assert geo.timezone == "UTC"
+
+
+@pytest.mark.asyncio
+async def test_explicit_timezone_and_locale_touch_no_network():
+    """BROWSER_TIMEZONE with BROWSER_LOCALE needs neither egress IP nor geoip."""
+    with (
+        patch("src.utils.BROWSER_TIMEZONE", "Asia/Shanghai"),
+        patch("src.utils.BROWSER_LOCALE", "zh-CN"),
+        patch("invisible_core._geo.discover_egress_ip") as discover,
+        patch("invisible_core.download.ensure_geoip_mmdb") as geoip,
+    ):
+        geo = await utils.get_browser_geo()
+
+    assert (geo.timezone, geo.locale) == ("Asia/Shanghai", "zh-CN")
+    discover.assert_not_called()
+    geoip.assert_not_called()
+
+
+def fake_invisible_playwright() -> MagicMock:
+    """Stand-in for InvisiblePlaywright that records how it was built."""
+    browser = MagicMock()
+    browser.new_context = AsyncMock(return_value=AsyncMock())
+    session = MagicMock()
+    session.__aenter__ = AsyncMock(return_value=browser)
+    session.__aexit__ = AsyncMock(return_value=False)
+    return MagicMock(return_value=session)
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize(
+    ("proxy_header", "expected"),
+    [
+        (None, {"timezone": "Europe/Berlin", "locale": "de"}),
+        ("socks5://proxy.test:1080", {"timezone": "", "locale": "auto"}),
+    ],
+)
+async def test_proxy_sessions_resolve_their_own_egress(
+    proxy_header: str | None, expected: dict[str, str]
+):
+    """The cache only covers direct sessions; a proxy's egress is its own."""
+    launcher = fake_invisible_playwright()
+    with (
+        patch("src.utils.PROXY_SERVER", None),
+        patch("src.utils.InvisiblePlaywright", launcher),
+        patch("src.utils.prepare_session_geo", return_value=BERLIN) as lookup,
+    ):
+        async for _ in utils.get_browser(x_proxy_server=proxy_header):
+            pass
+
+    kwargs = launcher.call_args.kwargs
+    assert {"timezone": kwargs["timezone"], "locale": kwargs["locale"]} == expected
+    assert lookup.called is (proxy_header is None)

uv.lock

@@ -86,6 +86,7 @@ dependencies = [
     { name = "pydantic" },
     { name = "pydantic-settings" },
     { name = "trafilatura" },
+    { name = "tzlocal" },
 ]
 
 [package.dev-dependencies]
@@ -111,6 +112,7 @@ requires-dist = [
     { name = "pydantic", specifier = "==2.*" },
     { name = "pydantic-settings", specifier = "==2.*" },
     { name = "trafilatura", specifier = "==2.3.*" },
+    { name = "tzlocal", specifier = "==5.*" },
 ]
 
 [package.metadata.requires-dev]