phantomfetch 0.4.7__tar.gz → 0.4.8__tar.gz

{phantomfetch-0.4.7 → phantomfetch-0.4.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: phantomfetch
-Version: 0.4.7
+Version: 0.4.8
 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

{phantomfetch-0.4.7 → phantomfetch-0.4.8}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "phantomfetch"
-version = "0.4.7"
+version = "0.4.8"
 description = "High-performance agentic web scraping library combining curl-cffi speed with Playwright browser capabilities"
 readme = "README.md"
 requires-python = ">=3.13"

{phantomfetch-0.4.7 → phantomfetch-0.4.8}/src/phantomfetch/fetch.py

@@ -1,11 +1,11 @@
 import asyncio
 import json
 import os
-from typing import Any, Literal, cast
+from typing import Any, Literal
 
 from loguru import logger
 
-from .cache import Cache, FileSystemCache
+from .cache import Cache
 from .engines import CDPEngine, CurlEngine
 from .pool import ProxyPool
 from .telemetry import get_tracer
@@ -46,6 +46,7 @@ class Fetcher:
         cache: Cache | bool | None = None,
         # Advanced CDP
         cdp_use_existing_page: bool = True,
+        cdp_connection_type: str = "cdp",
         # BrowserForge fingerprinting
         fingerprint: bool = True,
         fingerprint_options: dict[str, Any] | None = None,
@@ -97,9 +98,11 @@ class Fetcher:
             timeout=timeout,
             max_retries=max_retries,
         )
-        
+
         if browser_engine == "baas":
-            raise ValueError("BaaSEngine has been removed. Please use 'cdp' engine or contact support.")
+            raise ValueError(
+                "BaaSEngine has been removed. Please use 'cdp' engine or contact support."
+            )
 
         # Browser engine
         # Always use CDPEngine now
@@ -109,6 +112,7 @@ class Fetcher:
             timeout=browser_timeout,
             cache=self.cache,
             use_existing_page=cdp_use_existing_page,
+            cdp_connection_type=cdp_connection_type,
             fingerprint=fingerprint,
             fingerprint_options=fingerprint_options,
             browser_type=browser_type,
@@ -268,7 +272,7 @@ class Fetcher:
             proxy: Specific proxy to use (overrides pool)
             location: Geo location for proxy selection
             actions: List of `Action` objects or dicts (implies engine="browser")
-            
+
             ...
         """
         # Normalize actions - implies browser
@@ -319,21 +323,21 @@ class Fetcher:
                 cache_key_parts = [engine, url]
                 if location:
                     cache_key_parts.append(f"loc={location}")
-                
+
                 # Use proxy URL for cache key if explicit proxy is used
                 # Note: We rely on the proxy *argument* (manual override) for this separation.
-                # If using pool, we typically don't split cache by specific pool proxy, 
+                # If using pool, we typically don't split cache by specific pool proxy,
                 # UNLESS location was requested (handled above).
-                if proxy: 
+                if proxy:
                     # If manual proxy string/object provided, include it
                     p_url = proxy if isinstance(proxy, str) else proxy.url
-                    # Sanitize sensitive info? user:pass might be sensitive, 
-                    # but for cache key uniqueness it's needed. 
+                    # Sanitize sensitive info? user:pass might be sensitive,
+                    # but for cache key uniqueness it's needed.
                     # Since MD5 is used downstream, it's somewhat obscured.
                     cache_key_parts.append(f"proxy={p_url}")
 
                 cache_key = ":".join(cache_key_parts)
-                
+
                 cached_resp = await self.cache.get(cache_key)
                 if cached_resp:
                     cached_resp.from_cache = True
@@ -347,10 +351,12 @@ class Fetcher:
             selected_proxy: Proxy | None = None
             if proxy:
                 if isinstance(proxy, str):
-                    selected_proxy = Proxy(url=proxy, metadata={"source": "manual_override"})
+                    selected_proxy = Proxy(
+                        url=proxy, metadata={"source": "manual_override"}
+                    )
                 else:
                     selected_proxy = proxy
-            
+
             # 2. From Pool (if no override)
             if not selected_proxy:
                 selected_proxy = self.proxy_pool.get(url=url, location=location)
@@ -358,20 +364,27 @@ class Fetcher:
             if selected_proxy:
                 span.set_attribute("phantomfetch.proxy", selected_proxy.url)
                 if selected_proxy.vendor:
-                    span.set_attribute("phantomfetch.proxy.vendor", selected_proxy.vendor)
+                    span.set_attribute(
+                        "phantomfetch.proxy.vendor", selected_proxy.vendor
+                    )
                 if selected_proxy.proxy_type:
-                    span.set_attribute("phantomfetch.proxy.type", selected_proxy.proxy_type)
+                    span.set_attribute(
+                        "phantomfetch.proxy.type", selected_proxy.proxy_type
+                    )
                 if selected_proxy.location:
-                    span.set_attribute("phantomfetch.proxy.location", selected_proxy.location)
+                    span.set_attribute(
+                        "phantomfetch.proxy.location", selected_proxy.location
+                    )
                 if selected_proxy.provider:
-                    span.set_attribute("phantomfetch.proxy.provider", selected_proxy.provider)
+                    span.set_attribute(
+                        "phantomfetch.proxy.provider", selected_proxy.provider
+                    )
 
             # Route to engine
             if engine == "browser":
                 resp = await self._fetch_browser(
                     url=url,
                     proxy=selected_proxy,
-
                     headers=headers,
                     cookies=cookies,
                     actions=normalized_actions,
@@ -406,7 +419,7 @@ class Fetcher:
             # For simplicity, if we have a pool, and this proxy matches one in the pool, we could update it.
             # But the 'selected_proxy' might be a new instance specific to this request (manual override).
             # The pool.mark_* methods take a Proxy object.
-            
+
             # Logic: If 'proxy' argument was None, it came from pool -> Update stats.
             # If 'proxy' argument was set -> Do NOT update pool stats (it's a manual override).
             if not proxy and selected_proxy:
@@ -421,10 +434,10 @@ class Fetcher:
                 cache_key_parts = [engine, url]
                 if location:
                     cache_key_parts.append(f"loc={location}")
-                if proxy: 
+                if proxy:
                     p_url = proxy if isinstance(proxy, str) else proxy.url
                     cache_key_parts.append(f"proxy={p_url}")
-                
+
                 cache_key = ":".join(cache_key_parts)
                 await self.cache.set(cache_key, resp)
 
@@ -451,26 +464,26 @@ class Fetcher:
             async with self._browser_semaphore:
                 # Direct CDP usage (BaaSEngine removed)
                 if not self._cdp_engine:
-                     # Lazy init if not done (though currently init in __init__)
-                     # But wait, logic in __init__ was: 
-                     # self._cdp_engine = CDPEngine(...) if browser_engine == "cdp"
-                     # Since we removed engine selection, we should ensure it's initialized.
-                     # In __init__ I see: self._cdp_engine: CDPEngine | None = None
-                     # And the constructor logic for it was tied to the "if browser_engine == 'cdp'" block which I might have messed up or need to fix.
-                     # Let's check __init__ again.
-                     pass 
-
-                # Actually, looking at the previous file content, I removed the 'if browser_engine ==' logic in __init__? 
+                    # Lazy init if not done (though currently init in __init__)
+                    # But wait, logic in __init__ was:
+                    # self._cdp_engine = CDPEngine(...) if browser_engine == "cdp"
+                    # Since we removed engine selection, we should ensure it's initialized.
+                    # In __init__ I see: self._cdp_engine: CDPEngine | None = None
+                    # And the constructor logic for it was tied to the "if browser_engine == 'cdp'" block which I might have messed up or need to fix.
+                    # Let's check __init__ again.
+                    pass
+
+                # Actually, looking at the previous file content, I removed the 'if browser_engine ==' logic in __init__?
                 # No, I only updated the arguments. I need to make sure __init__ initializes _cdp_engine unconditionally.
                 # Let's assume I fix __init__ in a separate recursive step if needed, or I can fix it here if I see it.
-                # In the previous `view_file` output (Step 356), lines 87-101 show the old init logic. 
-                # I need to clean that up too! 
-                
+                # In the previous `view_file` output (Step 356), lines 87-101 show the old init logic.
+                # I need to clean that up too!
+
                 # Let's just assume _browser attribute is now _cdp_engine.
                 # Wait, the __init__ logic was:
                 # self._browser = CDPEngine(...)
                 # So here I should just use self._browser (which is typed as CDPEngine in the new world).
-                
+
                 # Correct implementation for _fetch_browser:
                 return await self._cdp_engine.fetch(
                     url=url,