Generate questions, 0 chunks found

[macduffhughes]

I have got the demo running, connected to DeepSeek for LLM and OpenAI for embeddings. I've created some knowledge bases with large educational PDFs. When I try to Generate Questions I don't get any. The logs show:
[Backend] [LightRAG] ○ Naive query: 0 chunks (chunk_top_k:20 cosine:0.2)
[Backend] [LightRAG] ○ Naive query: 0 chunks (chunk_top_k:20 cosine:0.2)
[Backend] [LightRAG] ○ Naive query: 0 chunks (chunk_top_k:20 cosine:0.2)
[Backend] [LightRAG] ○ [naive_query] No relevant document chunks found; returning no-result.
[Backend] [LightRAG] ○ [naive_query] No relevant document chunks found; returning no-result.
[Backend] [LightRAG] ○ [naive_query] No relevant document chunks found; returning no-result.

How to start troubleshooting? Thanks.

UPDATE:

1. I see nothing in the rag_storage directory. And in the logs of LightRAG init I notice
   [Backend] _storages_status = not_created
2. It looks like the LightRAG init is hardcoded to use gpt-4o-mini? I am using deepseek as my LLM, but in the LightRAG init I see
   tiktoken_model_name = gpt-4o-mini,
   and
   llm_model_name=gpt-4o-mini

[pancacake]

Hi, thanks for your feedback, is the problem still exist?

[macduffhughes]

I did a lot of debugging. First I found that mineru had not been properly installed: when I first did the setup pip said the dependency graph was too complex, so I did "uv pip install" instead, but apparently even that didn't completely work either. I had to make multiple runs of mineru from the command line, see it fail with a missing dependency, manually install the dependency, then try again. Finally it ran... it ran for several hours then failed because my machine did not have enough memory. All sounds too boring for the discussion board but I hope this info is helpful to you. Maybe I'll get a virtual machine with more memory and try again at some point. The project sounds really cool.

…

On Wed, Dec 31, 2025 at 5:30 AM Bingxi Zhao (Frank) < ***@***.***> wrote: Hi, thanks for your feedback, is the problem still exist? — Reply to this email directly, view it on GitHub <#5 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A7VVFTMCOJ53ETQXMBS4VRD4EPFWJAVCNFSM6AAAAACQKSL34GVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTKMZYGEZDMNI> . You are receiving this because you modified the open/close state.Message ID: ***@***.***>

[pancacake]

Yea totally get that. I’m sorry our install scripts had some compatibility issues that we missed initially. The uv installation is definitely the way to go, and I’ve already improved the scripts in the latest commit using uv methods. Please feel free to try it again and open an issue if you run into any new problems. Thanks for your patience with our project, and we wish deeptutor could be something in your daily life. btw, wish you happy new year :) 🎁