SDK: support file download in launch_local_browser (#4199)

This commit is contained in:
Stanislav Novosad
2025-12-05 09:50:22 -07:00
committed by GitHub
parent fadfe0848c
commit 32fb2315f0
2 changed files with 44 additions and 5 deletions

View File

@@ -1,5 +1,7 @@
import asyncio
import os
import pathlib
import tempfile
from typing import Any, overload
import httpx
@@ -389,6 +391,7 @@ class Skyvern(AsyncSkyvern):
headless: bool = False,
port: int = DEFAULT_CDP_PORT,
args: list[str] | None = None,
user_data_dir: str | None = None,
) -> SkyvernBrowser:
"""Launch a new local Chromium browser with Chrome DevTools Protocol (CDP) enabled.
@@ -404,13 +407,26 @@ class Skyvern(AsyncSkyvern):
Returns:
SkyvernBrowser: A browser instance with Skyvern capabilities.
"""
playwright = await self._get_playwright()
launch_args = [f"--remote-debugging-port={port}"]
if user_data_dir:
user_data_path = pathlib.Path(user_data_dir)
else:
user_data_path = pathlib.Path(tempfile.gettempdir()) / "skyvern-browser"
launch_args = [
f"--remote-debugging-port={port}",
]
if args:
launch_args.extend(args)
browser = await playwright.chromium.launch(headless=headless, args=launch_args)
browser_context = await playwright.chromium.launch_persistent_context(
user_data_dir=str(user_data_path),
headless=headless,
args=launch_args,
)
browser_address = f"http://localhost:{port}"
browser_context = browser.contexts[0] if browser.contexts else await browser.new_context()
return SkyvernBrowser(self, browser_context, browser_address=browser_address)
async def connect_to_browser_over_cdp(self, cdp_url: str) -> SkyvernBrowser:

View File

@@ -18,7 +18,7 @@ from urllib.parse import parse_qsl, urlparse
import psutil
import structlog
from playwright.async_api import BrowserContext, ConsoleMessage, Download, Page, Playwright
from playwright.async_api import Browser, BrowserContext, ConsoleMessage, Download, Page, Playwright
from skyvern.config import settings
from skyvern.constants import (
@@ -33,7 +33,6 @@ from skyvern.webeye.browser_artifacts import BrowserArtifacts, VideoArtifact
LOG = structlog.get_logger()
BrowserCleanupFunc = Callable[[], None] | None
@@ -148,6 +147,23 @@ def initialize_download_dir() -> str:
)
async def _apply_download_behaviour(browser: Browser) -> None:
context = ensure_context()
download_dir = get_download_dir(
context.run_id if context and context.run_id else context.workflow_run_id or context.task_id
)
cdp_session = await browser.new_browser_cdp_session()
await cdp_session.send(
"Browser.setDownloadBehavior",
{
"behavior": "allow",
"downloadPath": download_dir,
},
)
LOG.info("setDownloadBehavior applied", download_dir=download_dir)
class BrowserContextCreator(Protocol):
def __call__(
self, playwright: Playwright, proxy_location: ProxyLocation | None = None, **kwargs: dict[str, Any]
@@ -401,6 +417,7 @@ async def _create_headless_chromium(
playwright,
remote_browser_url=str(browser_address),
extra_http_headers=extra_http_headers,
apply_download_behaviour=True,
)
user_data_dir = make_temp_directory(prefix="skyvern_browser_")
@@ -436,6 +453,7 @@ async def _create_headful_chromium(
playwright,
remote_browser_url=str(browser_address),
extra_http_headers=extra_http_headers,
apply_download_behaviour=True,
)
user_data_dir = make_temp_directory(prefix="skyvern_browser_")
@@ -499,6 +517,7 @@ async def _create_cdp_connection_browser(
playwright,
remote_browser_url=str(browser_address),
extra_http_headers=extra_http_headers,
apply_download_behaviour=True,
)
browser_type = settings.BROWSER_TYPE
@@ -555,6 +574,7 @@ async def _connect_to_cdp_browser(
playwright: Playwright,
remote_browser_url: str,
extra_http_headers: dict[str, str] | None = None,
apply_download_behaviour: bool = False,
) -> tuple[BrowserContext, BrowserArtifacts, BrowserCleanupFunc]:
browser_args = BrowserContextFactory.build_browser_args(extra_http_headers=extra_http_headers)
@@ -565,6 +585,9 @@ async def _connect_to_cdp_browser(
LOG.info("Connecting browser CDP connection", remote_browser_url=remote_browser_url)
browser = await playwright.chromium.connect_over_cdp(remote_browser_url)
if apply_download_behaviour:
await _apply_download_behaviour(browser)
contexts = browser.contexts
browser_context = None