Automate browser based workflows with AI
Python aiapiautomationbrowserbrowser-automationcomputergptllmplaywrightpowerautomatepuppeteerpythonrpaseleniumvisionworkflow
fix uppload file imaging issue (#1234)
LawyZheng9f2fedae4e66
diff --git a/skyvern/webeye/actions/handler.py b/skyvern/webeye/actions/handler.pyindex 800c28f10..2ca4b9c96 100644--- a/skyvern/webeye/actions/handler.py+++ b/skyvern/webeye/actions/handler.py@@ -620,7 +620,12 @@ async def handle_upload_file_action(# ************************************************************************************************************** #file_url = await get_actual_value_of_parameter_if_secret(task, action.file_url)decoded_url = urllib.parse.unquote(file_url)- if file_url not in str(task.navigation_payload) and decoded_url not in str(task.navigation_payload):+ if (+ file_url not in str(task.navigation_payload)+ and file_url not in str(task.navigation_goal)+ and decoded_url not in str(task.navigation_payload)+ and decoded_url not in str(task.navigation_goal)+ ):LOG.warning("LLM might be imagining the file url, which is not in navigation payload",action=action,