Automate browser based workflows with AI
Python aiapiautomationbrowserbrowser-automationcomputergptllmplaywrightpowerautomatepuppeteerpythonrpaseleniumvisionworkflow
improve download handler (#3642)
LawyZheng04e768987ef5
diff --git a/skyvern/webeye/utils/dom.py b/skyvern/webeye/utils/dom.pyindex 7a5a914e8..ed5ad653c 100644--- a/skyvern/webeye/utils/dom.py+++ b/skyvern/webeye/utils/dom.py@@ -296,6 +296,12 @@ class SkyvernElement:)return False+ async def is_child_of_pdf_object(self, timeout: float = settings.BROWSER_ACTION_TIMEOUT_MS) -> bool:+ parent_locator = self.get_locator().locator("..")+ tag_name: str | None = await parent_locator.evaluate("el => el.tagName", timeout=timeout)+ type_attr = await parent_locator.get_attribute("type", timeout=timeout)+ return tag_name is not None and tag_name.lower() == "object" and type_attr == "application/pdf"+async def is_parent_of(self, target: ElementHandle) -> bool:skyvern_frame = await SkyvernFrame.create_instance(self.get_frame())return await skyvern_frame.is_parent(await self.get_element_handler(), target)@@ -361,7 +367,7 @@ class SkyvernElement:return handlerasync def should_use_navigation_instead_click(self, page: Page) -> str | None:- if await self.get_attr("target", mode="static") != "_blank":+ if await self.get_attr("target", mode="static") != "_blank" and not await self.is_child_of_pdf_object():return Nonehref: str | None = await self.get_attr("href", mode="static")@@ -837,7 +843,12 @@ class SkyvernElement:except Exception as e:# some cases use this method to download a file. but it will be redirected away soon# and agent will run into ABORTED error.- if "net::ERR_ABORTED" in str(e):+ error = str(e)+ if "net::ERR_ABORTED" in error:+ return href++ # some cases playwright will raise error like "Page.goto: Download is starting"+ if "Page.goto: Download is starting" in error:return hrefLOG.warning("Failed to navigate to the <a> href link", exc_info=True, href=href, current_url=page.url)