Handle downloads in playwright browser, solved the issue #127

src/magentic_ui/tools/playwright/browser/headless_docker_playwright_browser.py

@@ -2,6 +2,7 @@
 
 import asyncio
 import logging
+from typing import Any, Dict
 
 from autogen_core import Component
 import docker
@@ -87,21 +88,21 @@ async def create_container(self) -> Container:
         )
 
         client = docker.from_env()
-        return await asyncio.to_thread(
-            client.containers.create,
-            name=f"magentic-ui-headless-browser_{self._playwright_port}",
-            image="mcr.microsoft.com/playwright:v1.51.1-noble",
-            detach=True,
-            auto_remove=True,
-            ports={
+        container_config: Dict[str, Any] = {
+            "name": f"magentic-ui-headless-browser_{self._playwright_port}",
+            "image": "mcr.microsoft.com/playwright:v1.51.1-noble",
+            "detach": True,
+            "auto_remove": True,
+            "ports": {
                 f"{self._playwright_port}/tcp": self._playwright_port,
             },
-            command=[
+            "command": [
                 "/bin/sh",
                 "-c",
                 f"npx -y [email protected] run-server --port {self._playwright_port} --host 0.0.0.0",
             ],
-        )
+        }
+        return await asyncio.to_thread(client.containers.create, **container_config)
 
     def _to_config(self) -> HeadlessBrowserConfig:
         return HeadlessBrowserConfig(

src/magentic_ui/tools/playwright/browser/local_playwright_browser.py

@@ -2,9 +2,10 @@
 
 from typing import Optional, Any, Dict
 from pathlib import Path
+import os
 
 from autogen_core import Component
-from playwright.async_api import BrowserContext, Browser
+from playwright.async_api import BrowserContext, Browser, Route
 from pydantic import BaseModel
 
 from playwright.async_api import async_playwright, Playwright
@@ -99,15 +100,19 @@ async def _start(self) -> None:
             # Ensure the browser data directory exists
             Path(self._browser_data_dir).mkdir(parents=True, exist_ok=True)
 
-            # Launch persistent context
+            # Launch persistent context with automatic downloads
             self._context = await self._playwright.chromium.launch_persistent_context(
                 self._browser_data_dir,
-                accept_downloads=self._enable_downloads,
+                accept_downloads=True,  # Always accept downloads
                 **launch_options,
                 args=["--disable-extensions", "--disable-file-system"],
                 env={},
                 chromium_sandbox=True,
             )
+
+            # Set up download behavior for persistent context
+            if self._enable_downloads:
+                await self._context.route("**/*", self._handle_download)
         else:
             # Launch regular browser and create new context
             self._browser = await self._playwright.chromium.launch(
@@ -117,11 +122,47 @@ async def _start(self) -> None:
                 env={} if self._headless else {"DISPLAY": ":0"},
             )
 
+            # Create context with automatic downloads
             self._context = await self._browser.new_context(
                 user_agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36 Edg/122.0.0.0",
-                accept_downloads=self._enable_downloads,
+                accept_downloads=True,  # Always accept downloads
             )
 
+            # Set up download behavior for regular context
+            if self._enable_downloads:
+                await self._context.route("**/*", self._handle_download)
+
+    async def _handle_download(self, route: Route) -> None:
+        """
+        Handle download requests by intercepting them and saving to the .webby directory.
+        """
+        response = await route.fetch()
+        headers = response.headers
+
+        # Check if this is a download (Content-Disposition header)
+        if "content-disposition" in headers:
+            # Extract filename from Content-Disposition header
+            content_disposition = headers["content-disposition"]
+            filename = None
+            if "filename=" in content_disposition:
+                filename = content_disposition.split("filename=")[1].strip('"')
+
+            if filename and self._browser_data_dir:
+                # Save to .webby directory
+                webby_dir = os.path.join(self._browser_data_dir, ".webby")
+                os.makedirs(webby_dir, exist_ok=True)
+
+                filepath = os.path.join(webby_dir, filename)
+                with open(filepath, "wb") as f:
+                    f.write(await response.body())
+
+                # Continue with the download in browser
+                await route.continue_()
+            else:
+                await route.continue_()
+        else:
+            await route.continue_()
+
     async def _close(self) -> None:
         """
         Close the browser resource.

src/magentic_ui/tools/playwright/browser/vnc_docker_playwright_browser.py

@@ -2,6 +2,7 @@
 
 import asyncio
 import logging
+from typing import Any, Dict, Tuple
 
 from pathlib import Path
 import secrets
@@ -101,7 +102,7 @@ def __init__(
         )
         self._docker_name = f"magentic-ui-vnc-browser_{self._playwright_websocket_path}_{self._novnc_port}"
 
-    def _get_available_port(self) -> tuple[int, socket.socket]:
+    def _get_available_port(self) -> Tuple[int, socket.socket]:
         s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
         s.bind(("127.0.0.1", 0))
         port = s.getsockname()[1]
@@ -160,27 +161,26 @@ async def create_container(self) -> Container:
         )
 
         client = docker.from_env()
-
-        return await asyncio.to_thread(
-            client.containers.create,
-            name=self._docker_name,
-            image=self._image,
-            detach=True,
-            auto_remove=True,
-            network=self._network_name if self._inside_docker else None,
-            ports={
+        container_config: Dict[str, Any] = {
+            "name": self._docker_name,
+            "image": self._image,
+            "detach": True,
+            "auto_remove": True,
+            "network": self._network_name if self._inside_docker else None,
+            "ports": {
                 f"{self._playwright_port}/tcp": self._playwright_port,
                 f"{self._novnc_port}/tcp": self._novnc_port,
             },
-            volumes={
+            "volumes": {
                 str(self._bind_dir.resolve()): {"bind": "/workspace", "mode": "rw"}
             },
-            environment={
+            "environment": {
                 "PLAYWRIGHT_WS_PATH": self._playwright_websocket_path,
                 "PLAYWRIGHT_PORT": str(self._playwright_port),
                 "NO_VNC_PORT": str(self._novnc_port),
             },
-        )
+        }
+        return await asyncio.to_thread(client.containers.create, **container_config)
 
     def _to_config(self) -> VncDockerPlaywrightBrowserConfig:
         return VncDockerPlaywrightBrowserConfig(