phantomfetch 0.5.8__tar.gz → 0.5.10__tar.gz

{phantomfetch-0.5.8 → phantomfetch-0.5.10}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: phantomfetch
-Version: 0.5.8
+Version: 0.5.10
 Summary: High-performance agentic web scraping library combining curl-cffi speed with Playwright browser capabilities
 Keywords: web-scraping,playwright,curl-cffi,async,browser-automation,http-client,agentic,anti-detection
 Author: CosmicBull
@@ -28,11 +28,13 @@ Requires-Dist: beautifulsoup4>=4.14.3
 Requires-Dist: cloakbrowser[geoip]>=0.3.22 ; extra == 'all'
 Requires-Dist: camoufox[geoip]>=0.4.11 ; extra == 'all'
 Requires-Dist: rebrowser-playwright>=1.52.0 ; extra == 'all'
+Requires-Dist: patchright>=1.52.0 ; extra == 'all'
 Requires-Dist: maxminddb>=2.0.0 ; extra == 'all'
 Requires-Dist: camoufox[geoip]>=0.4.11 ; extra == 'camoufox'
 Requires-Dist: cloakbrowser>=0.3.22 ; extra == 'cloakbrowser'
 Requires-Dist: cloakbrowser[geoip]>=0.3.22 ; extra == 'geoip'
 Requires-Dist: maxminddb>=2.0.0 ; extra == 'geoip'
+Requires-Dist: patchright>=1.52.0 ; extra == 'patchright'
 Requires-Dist: rebrowser-playwright>=1.52.0 ; extra == 'rebrowser'
 Requires-Python: >=3.13
 Project-URL: Homepage, https://github.com/iristech-systems/PhantomFetch
@@ -44,6 +46,7 @@ Provides-Extra: all
 Provides-Extra: camoufox
 Provides-Extra: cloakbrowser
 Provides-Extra: geoip
+Provides-Extra: patchright
 Provides-Extra: rebrowser
 Description-Content-Type: text/markdown
 

{phantomfetch-0.5.8 → phantomfetch-0.5.10}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "phantomfetch"
-version = "0.5.8"
+version = "0.5.10"
 description = "High-performance agentic web scraping library combining curl-cffi speed with Playwright browser capabilities"
 readme = "README.md"
 requires-python = ">=3.13"
@@ -47,7 +47,8 @@ cloakbrowser = ["cloakbrowser>=0.3.22"]
 geoip = ["cloakbrowser[geoip]>=0.3.22", "maxminddb>=2.0.0"]
 camoufox = ["camoufox[geoip]>=0.4.11"]
 rebrowser = ["rebrowser-playwright>=1.52.0"]
-all = ["cloakbrowser[geoip]>=0.3.22", "camoufox[geoip]>=0.4.11", "rebrowser-playwright>=1.52.0", "maxminddb>=2.0.0"]
+patchright = ["patchright>=1.52.0"]
+all = ["cloakbrowser[geoip]>=0.3.22", "camoufox[geoip]>=0.4.11", "rebrowser-playwright>=1.52.0", "patchright>=1.52.0", "maxminddb>=2.0.0"]
 
 [project.urls]
 Homepage = "https://github.com/iristech-systems/PhantomFetch"

{phantomfetch-0.5.8 → phantomfetch-0.5.10}/src/phantomfetch/engines/browser/actions.py

@@ -186,6 +186,46 @@ async def execute_actions(
                             result.error = "Loop requires selector and child actions"
                         else:
                             limit = action.max_iterations or 100
+                            js_supported_actions = {
+                                "wait",
+                                "click",
+                                "extract",
+                                "evaluate",
+                            }
+                            has_deferred_actions = any(
+                                sub_action.action not in js_supported_actions
+                                for sub_action in action.actions
+                            )
+
+                            if has_deferred_actions:
+                                import msgspec
+
+                                logger.debug(
+                                    f"[browser] Loop '{action.selector}' contains deferred actions; using Python fallback (limit={limit})"
+                                )
+
+                                loop_locator = ctx.locator(action.selector)
+                                loop_count = min(await loop_locator.count(), limit)
+                                loop_results = []
+
+                                for i in range(loop_count):
+                                    iter_ctx = loop_locator.nth(i)
+                                    iter_results = await execute_actions(
+                                        iter_ctx, action.actions
+                                    )
+                                    loop_results.append(
+                                        {
+                                            "index": i,
+                                            "results": [
+                                                msgspec.to_builtins(ir)
+                                                for ir in iter_results
+                                            ],
+                                        }
+                                    )
+
+                                result.data = loop_results
+                                continue
+
                             logger.debug(
                                 f"[browser] Compiling loop for '{action.selector}' to native JS (limit={limit})"
                             )

{phantomfetch-0.5.8 → phantomfetch-0.5.10}/src/phantomfetch/engines/browser/cdp.py

@@ -217,6 +217,7 @@ class CDPEngine:
         self._browser: Any = None
         self._existing_context: Any = None
         self._existing_page: Any = None
+        self._connect_lock = asyncio.Lock()
 
     async def connect(self) -> None:
         """Initialize Playwright and connect to browser with retry."""
@@ -442,6 +443,17 @@ class CDPEngine:
             self._camoufox_context = None
             self._camoufox_proxy_key = None
 
+    async def start(self) -> None:
+        """Start the browser engine."""
+        async with self._connect_lock:
+            await self.connect()
+
+    async def restart(self) -> None:
+        """Restart the browser engine."""
+        async with self._connect_lock:
+            await self.disconnect()
+            await self.connect()
+
     async def _handle_route(self, route: "Route") -> None:
         """Handle network requests for caching."""
         if not self.cache:
@@ -601,12 +613,13 @@ class CDPEngine:
 
             # Ensure browser is running.
             # CloakBrowser persistent mode uses _cloak_context (not _browser).
-            needs_connect = not self._browser and not (
-                (self._cloak_browser_available and self.persistent_context_dir)
-                or (self._camoufox_available and self.persistent_context_dir)
-            )
-            if needs_connect:
-                await self.connect()
+            async with self._connect_lock:
+                needs_connect = not self._browser and not (
+                    (self._cloak_browser_available and self.persistent_context_dir)
+                    or (self._camoufox_available and self.persistent_context_dir)
+                )
+                if needs_connect:
+                    await self.connect()
 
             # Create context
             # We create a fresh context for each request to ensure isolation

{phantomfetch-0.5.8 → phantomfetch-0.5.10}/src/phantomfetch/fetch.py

@@ -203,6 +203,13 @@ class Fetcher:
         if self._browser:
             await self._browser.disconnect()
 
+    async def restart_browser(self) -> None:
+        """
+        Restart the browser engine.
+        """
+        if self._browser and hasattr(self._browser, "restart"):
+            await self._browser.restart()
+
     def save_session(self, path: str) -> None:
         """
         Save the current session storage (cookies, localStorage) to a file.