[BUG]Upload files failed

[xjh6]

Describe the bug

I deployed the project natively using Python and configured the .env file. I encountered the same problem as #36. When logging in to the backend, it showed "detail": "Not Found".

To Reproduce

steps on readme.md

I deployed the project natively on a server. And used ssh to connect. The remote connection command is like this:
ssh -L 8889:localhost:5173 -L 8890:localhost:8000 -p xxxx xxx@xxxx
I visited the front-end page: http://localhost:8889. Click add data to upload the file. After loading for a while, the uploaded file was not displayed.
Then I visited the back-end page: http://localhost:8890, and it was displayed as follows:
{
"detail": "Not Found"
}
The following content popped up on the back-end server interface:
INFO: 127.0.0.1:36656 - "GET / HTTP/1.1" 404 Not Found

I am not using OpenAI's model, I am using qwen and qwen's embeddings model. I changed some codes in the backend openai_llm_service.py and openai_embedding_service.py.
self.client = OpenAI(base_url="https://dashscope.aliyuncs.com/compatible-mode/v1", api_key=settings.openai_api_key)
Just added base_url in this line

Logs and Error Reports

On the backend show like this:
{
"detail": "Not Found"
}
In the server show like this:
INFO: 127.0.0.1:36656 - "GET / HTTP/1.1" 404 Not Found

Screenshots

[tomsmoker]

Hey @xjh6, can you share the code where you changed the model? To add a new model, you need to extend the factory to allow for an alternative one. Happy to help show you how.

[tomthebuzz]

After quite some while of waiting and now re-installing from scratch I seem to be at the same place:

• all tests complete without any errors
• backend and front-end are working
• i can't upload any files => simply nothing happens after I select a .pdf to upload

Also no errors show in the backend or in the frontend. However when I access the backend via browser I get the same 404 error response as @xjh6:

INFO:     Loading config settings from the environment...
INFO:     Started server process [72808]
INFO:     Waiting for application startup.
INFO:     Application startup complete.
INFO:     Uvicorn running on http://127.0.0.1:8000 (Press CTRL+C to quit)
INFO:     127.0.0.1:59349 - "GET / HTTP/1.1" 404 Not Found

[TreyWard-VT]

I am also having the same issues; my goal is to link the knowledge table to OWUI setup w/Ollama & docker. I setup Knowledge tables by giving the .env file my Ollama API url. I think i may need to feed it an API key frop Open AI & give that same key to my OWUI.

any advice @tomsmoker ?

[tomsmoker]

@TreyWard-VT the factories we have in place only support OpenAI at the moment. Supporting Ollama is harder given the variety of models, and the factory will need to be extended. Which model are you trying to use specifically?

[TreyWard-VT]

@TreyWard-VT the factories we have in place only support OpenAI at the moment. Supporting Ollama is harder given the variety of models, and the factory will need to be extended. Which model are you trying to use specifically?

I have been running Gemma3's 4 billion parameter model. Sounds like for the short term, my best bet is to make use of OpenAI platform.