DGX Spark Playbooks: Agentic RAG: Chat UI errors out

[ms6174]

Hi,

I’m trying out the “RAG Application in AI Workbench“ RAG chat UI playbook. I could start the container and view the GUI:

However, when I send a user input, I get the following 403 error:
File “/home/workbench/.local/lib/python3.10/site-packages/langchain_core/language_models/chat_models.py”, line 842, in generate

self.\_generate_with_cache(

File “/home/workbench/.local/lib/python3.10/site-packages/langchain_core/language_models/chat_models.py”, line 1091, in _generate_with_cache

result = self.\_generate(

File “/home/workbench/.local/lib/python3.10/site-packages/langchain_nvidia_ai_endpoints/chat_models.py”, line 382, in _generate

response = self.\_client.get_req(payload=payload, extra_headers=extra_headers)

File “/home/workbench/.local/lib/python3.10/site-packages/langchain_nvidia_ai_endpoints/_common.py”, line 473, in get_req

response, session = self.\_post(

File “/home/workbench/.local/lib/python3.10/site-packages/langchain_nvidia_ai_endpoints/_common.py”, line 369, in _post

self.\_try_raise(response)

File “/home/workbench/.local/lib/python3.10/site-packages/langchain_nvidia_ai_endpoints/_common.py”, line 462, in _try_raise

raise Exception(f"{header}\\n{body}") from None

Exception: [403] Forbidden

Authorization failed

-–ROUTE QUESTION—

Hi

I got the NVIDIA personal key with the following authorizations:

Appreciate your help in debugging this issue. Thanks in advance!

[aniculescu]

According to the playbook, your API key needs Public API Endpoints permissions

[ms6174]

Thank you for getting back. I got the key from the NGC platform and I didn’t see an option to check the key’s visibility. Could you please share how to confirm whether the key is public or private.

Thanks!

[aniculescu]

In the key permissions you have “Secrets Manager” and “NGC Catalog”. You can edit these permissions to add “Public API Endpoints” as a permission which will allow you to access the right resources to run the playbook.