merge: bring all updates into main

DouYinSparkFlow/core/browser.py

@@ -52,13 +52,26 @@ def _browser_args():
     ]
 
 
-def _douyin_browser_proxy():
-    """Return an explicit proxy for Douyin browser traffic, or None for direct."""
+def _douyin_network_mode():
     settings = get_app_settings(force_reload=True)
-    mode = str(
+    return str(
         os.getenv("SPARKFLOW_DOUYIN_NETWORK_MODE")
         or settings.get("douyin_network_mode", "direct")
     ).strip().lower()
+
+
+def douyin_network_modes():
+    # Direct is the default; Mihomo is used only when explicitly selected.
+    mode = _douyin_network_mode()
+    if mode == "mihomo":
+        return ("mihomo",)
+    return ("direct",)
+
+
+def _douyin_browser_proxy(network_mode=None):
+    # Return an explicit proxy URL for Douyin traffic, or None for direct.
+    settings = get_app_settings(force_reload=True)
+    mode = str(network_mode or _douyin_network_mode()).strip().lower()
     if mode != "mihomo":
         return None
     return str(
@@ -67,13 +80,45 @@ def _douyin_browser_proxy():
     ).strip() or None
 
 
-def _browser_launch_options(GUI=False):
+def _browser_launch_options(GUI=False, network_mode=None):
     args = _browser_args()
-    proxy = _douyin_browser_proxy()
+    proxy = _douyin_browser_proxy(network_mode=network_mode)
     if proxy:
-        return {"headless": _headless_for(GUI), "args": args, "proxy": {"server": proxy}}
+        return {
+            "headless": _headless_for(GUI),
+            "args": args,
+            "proxy": {"server": proxy},
+        }
     args.append("--no-proxy-server")
-    return {"headless": _headless_for(GUI), "args": args}
+    return {
+        "headless": _headless_for(GUI),
+        "args": args,
+    }
+
+
+async def select_douyin_network_mode(target_url):
+    # Select the first route that can load the target before a task starts.
+    failures = []
+    for network_mode in douyin_network_modes():
+        playwright = browser = page = None
+        try:
+            playwright, browser = await get_browser(network_mode=network_mode)
+            page = await browser.new_page()
+            response = await page.goto(target_url, wait_until="commit", timeout=30000)
+            status = response.status if response is not None else None
+            if status is not None and status < 500:
+                return network_mode
+            failures.append(f"{network_mode}: HTTP {status}")
+        except Exception as exc:
+            failures.append(f"{network_mode}: {exc}")
+        finally:
+            if page:
+                await page.close()
+            if browser:
+                await browser.close()
+            if playwright:
+                await playwright.stop()
+    raise RuntimeError(f"Douyin network preflight failed: {'; '.join(failures)}")
 
 
 def sanitize_profile_name(value):
@@ -102,12 +147,12 @@ async def install_browser():
         console.print(f"[bold red]Browser install failed: {exc}[/bold red]")
 
 
-async def get_browser(GUI=False):
+async def get_browser(GUI=False, network_mode=None):
     configure_playwright_environment()
 
     try:
         playwright = await async_playwright().start()
-        browser = await playwright.chromium.launch(**_browser_launch_options(GUI))
+        browser = await playwright.chromium.launch(**_browser_launch_options(GUI, network_mode=network_mode))
         return playwright, browser
     except Exception as exc:
         if "Executable doesn't exist" in str(exc) and get_environment() != Environment.GITHUBACTION:
@@ -118,7 +163,7 @@ async def get_browser(GUI=False):
         raise
 
 
-async def get_persistent_browser_context(profile_name, GUI=False, root=None):
+async def get_persistent_browser_context(profile_name, GUI=False, root=None, network_mode=None):
     configure_playwright_environment()
 
     profile_dir = browser_profile_root(root) / sanitize_profile_name(profile_name)
@@ -126,7 +171,7 @@ async def get_persistent_browser_context(profile_name, GUI=False, root=None):
 
     try:
         playwright = await async_playwright().start()
-        launch_options = _browser_launch_options(GUI)
+        launch_options = _browser_launch_options(GUI, network_mode=network_mode)
         launch_options["viewport"] = {"width": 1600, "height": 1000}
         context = await playwright.chromium.launch_persistent_context(
             str(profile_dir),