Hi All, Have been trying to get Brandon's CrewAI Rag code running that scrapes youtube for the five top videos (crewai-rag-deep-dive/2_youtube_and_web). I have installed and re-installed the poetry shell. The code is exactly the same - just a git clone of Brandon's code. That being said I have followed the vid twice to ensure all is well - I cannot see where my env starts to go pear shaped. The first task runs as per Brandon's output, It is when the code starts executing the second agent/task that it fails to add any of the five videos to the vector store. I have let the code run to the end and it finishes all the tasks and does find a few items from the web searches. But it is the failed vectorDB that has me stumped! Youtube cc are enabled. Help please! Thanks Alan. ****** > Entering new CrewAgentExecutor chain... I need to process each of the extracted video URLs and add them to the vector database using the given tool. I will proceed to add each video URL one by one. Action: Add Video to Vector DB Action Input: {"video_url": "https://www.youtube.com/watch?v=8PtGcNE01yo"} success=False Thought: The first attempt to add the video to the vector database was unsuccessful. I will try adding the next video URL to see if it works. Action: Add Video to Vector DB Action Input: {"video_url": "https://www.youtube.com/watch?v=cyPZsbO5i5U"} success=False Thought: The second attempt also failed. I will proceed to try adding the next video URL to the vector database to check for any successful additions. Action: Add Video to Vector DB Action Input: {"video_url": "https://www.youtube.com/watch?v=2qU3SPPojDA"} success=False Thought: I need to continue attempting to add the remaining video URLs to the vector database to see if any are successful. Action: Add Video to Vector DB Action Input: {"video_url": "https://www.youtube.com/watch?v=np1cwfDjCF4"}

Hi All, Have been trying to get Brandon's CrewAI Rag code running that scrapes youtube for the five top videos (crewai-rag-deep-dive/2_youtube_and_web). I have installed and re-installed the poetry shell. The code is exactly the same - just a git clone of Brandon's code. That being said I have followed the vid twice to ensure all is well - I cannot see where my env starts to go pear shaped. My output seems different - less verbose? It is when the code starts executing the second agent/task that it fails to add any of the five videos to the vector store. I have let the code run to the end and it finishes all the tasks and does find a few items from the web searches. But it is the failed vectorDB that has me stumped! Help please! Thanks Alan. *** terminal output *** crewai-rag-basics-py3.10(base) a_shipley@ShipleyAI:~/Python_poetry/crewai-rag-deep-dive/2_youtube_and_web$ python crew.py Please enter the YouTube handle to analyze: bhancock_ai > Entering new CrewAgentExecutor chain... To proceed with scraping the latest five videos from the specified YouTube channel, I will use the available tool to fetch the latest videos from the channel handle `bhancock_ai`. Action: Fetch Latest Videos for Channel Action Input: {"youtube_channel_handle": "@bhancock_ai", "max_results": 5} videos=[VideoInfo(video_id='8PtGcNE01yo', title='CrewAI Flows Crash Course', publish_date=datetime.datetime(2024, 10, 25, 13, 45, 29, tzinfo=datetime.timezone.utc), video_url='https://www.youtube.com/watch?v=8PtGcNE01yo'), VideoInfo(video_id='cyPZsbO5i5U', title='Supabase Postgres Vector DB Crash Course', publish_date=datetime.datetime(2024, 10, 18, 20, 48, 14, tzinfo=datetime.timezone.utc), video_url='https://www.youtube.com/watch?v=cyPZsbO5i5U'), VideoInfo(video_id='2qU3SPPojDA', title='How to Build a NextJS MVP using v0, Claude, and Cursor', publish_date=datetime.datetime(2024, 10, 10, 17, 13, 35, tzinfo=datetime.timezone.utc), video_url='https://www.youtube.com/watch?v=2qU3SPPojDA'), VideoInfo(video_id='np1cwfDjCF4', title='Build a Full Stack AI Marketing Platform SaaS [Premium Course Announcement]', publish_date=datetime.datetime(2024, 9, 19, 17, 13, 33, tzinfo=datetime.timezone.utc), video_url='https://www.youtube.com/watch?v=np1cwfDjCF4'), VideoInfo(video_id='fMa2zQIkQwM', title='Claude 3.5 Crash Course for Developers: Code 10x Faster in 2024 [Claude 3.5 artifacts]', publish_date=datetime.datetime(2024, 8, 21, 13, 39, 24, tzinfo=datetime.timezone.utc), video_url='https://www.youtube.com/watch?v=fMa2zQIkQwM')]

Not sure what might be causing this... this is the agent output whilst running the crewai.py from Brandon's crewai rag video. All 5 failed. It did pick up most of the info from the scrape but obviously the youtube rag is the key element. > Entering new CrewAgentExecutor chain... I need to add each of the provided YouTube video URLs to the vector database using the available tool. I will begin with the first video in the list and proceed sequentially. Action: Add Video to Vector DB Action Input: {"video_url": "https://www.youtube.com/watch?v=8PtGcNE01yo"} success=False Thought: The first attempt to add the video to the vector database failed. I will attempt to add the video again and check for a successful operation. If it fails again, I will proceed with the next video. Action: Add Video to Vector DB Action Input: {"video_url": "https://www.youtube.com/watch?v=8PtGcNE01yo"} I tried reusing the same input, I must stop using this action input. I'll try something else instead.

Hi All, Hoping you may be able to lend some pointers to a noob. I have been trying to get crewai flows working but I fail when trying to install crewai 0.75.1 into a condo env. I have added a screen shots of the fail reason. I have been trying to install all day and troubleshoot for the last few hours but to no avail. I have tried changing python versions 3.10 and 3.11 same error. Would be great of someone could help!! Thanks Al.

Hi All, Hoping for some help. I have followed instructions and I get the following error - /home/a_shipley/.cache/pypoetry/virtualenvs/langchain-crash-course-rMBVWZ4b-py3.10/lib/python3.10/site-packages/google/auth/_default.py:76: UserWarning: Your application has authenticated using end user credentials from Google Cloud SDK without a quota project. You might receive a "quota exceeded" or "API not enabled" error. See the following page for troubleshooting: https://cloud.google.com/docs/authentication/adc-troubleshooting/user-creds. I have followed all troubleshooting - setting quota, ensuring firestorm db is up, api is enabled in gcloud, i have the right user logged into cli and the right project assigned. Thanks all!