[BUG] -File uploaded from Chat homescreen tab works, same files uploaded form Files - File Collection tab breaks

[Niko-La]

Description

uploaded from chat are at the bottom with token counts, when reindex or on a fresh upload fromt he Files tab it files to create token

Reproduction steps

1. Go to '...'
2. Click on '....'
3. Scroll down to '....'
4. See error

Screenshots

![DESCRIPTION](LINK.png)

Logs

No response

Browsers

No response

OS

No response

Additional information

No response

[taprosoft]

Make sure you setup default embedding model and attach correct embedding model to the File Collection following this: https://github.com/Cinnamon/kotaemon/blob/main/docs/local_model.md#use-local-models-for-rag

[taprosoft]

Home tab upload silently ignore embedding error so you can still upload it successfully, but embedding will not be created and can not be used for vector search.

[Niko-La]

Make sure you setup default embedding model and attach correct embedding model to the File Collection following this: https://github.com/Cinnamon/kotaemon/blob/main/docs/local_model.md#use-local-models-for-rag

tried this but same issue when uploading fro File tab.

@taprosoft How do we know if the uploaded file has the embedding are created correctly?

Im unable to get my local chat going at all #138 added my screenshot there. :slightly_frowning_face:

[munkito]

Hi,

I have the same problem with the local model, it's not working. It throws the following exception: raise retry_exc from fut.exception() tenacity.RetryError: RetryError[<Future at 0x237dfc10d50 state=finished raised APIConnectionError>]

I think there is a mistake with the url of the Ollama API http://localhost:11434/v1/, I'm not sure if the "v1" is correct, if I try to open this url and it returns a 404.

Thanks!!

[phv2312]

Hi @Niko-La , you can check the success of indexing steps by Files -> File Collection (or your Collection). IF you can see the chunks & its content. You're almost success Besides, we have added feature for user to test your connection recently. For example, go to Resouces -> LLM | Embedding -> Click on your option -> There are a Test tab above the Save button

[Niko-La]

Hi,

I have the same problem with the local model, it's not working. It throws the following exception: raise retry_exc from fut.exception() tenacity.RetryError: RetryError[<Future at 0x237dfc10d50 state=finished raised APIConnectionError>]

I think there is a mistake with the url of the Ollama API http://localhost:11434/v1/, I'm not sure if the "v1" is correct, if I try to open this url and it returns a 404.

Thanks!!

getting same error when visitng

[munkito]

Same error:

[munkito]

Hi @Niko-La , you can check the success of indexing steps by Files -> File Collection (or your Collection). IF you can see the chunks & its content. You're almost success Besides, we have added feature for user to test your connection recently. For example, go to Resouces -> LLM | Embedding -> Click on your option -> There are a Test tab above the Save button

Test success

[phv2312]

@munkito can you help to check whether you have change the embedding settings of FileIndex to ollama. You can refer to the document https://github.com/Cinnamon/kotaemon/blob/main/docs/local_model.md#use-local-models-for-rag

[phv2312]

Besides, in case of error still happend, can you provide the logs (the powershell, terminall,...) that you launch the app. It will help us in investigating the problem?

[munkito]

Hi,

@phv2312 thanks for the link, My configuration was wrong on embedding model for the File Collection to a local model. Then, I have news:

Good news: Now It doesn't raise an error. Bad news: It doesn't respond, It holds on in this step:

Got 10 retrieved documents
len (original) 25640
len (trimmed) 25640
Got 3 images
Trying LLM streaming
CitationPipeline: invoking LLM

What logs do you need?

Thanks!

[munkito]

And finally it throws the following errors:

LLM rerank scores [0.9, 0.8, 0.7, 0.6, 0.6, 0.5, 0.3, 0.3, 0.3, 0.3]
CitationPipeline: finish invoking LLM
2 validation errors for QuestionAnswer
answer.0.fact
  Field required [type=missing, input_value={'body': '15 días labora...go de Trabajo (1948).']}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.8/v/missing
answer.0.substring_quote
  Field required [type=missing, input_value={'body': '15 días labora...go de Trabajo (1948).']}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.8/v/missing
CitationPipeline: finish invoking LLM
1 validation error for QuestionAnswer
answer
  Input should be a valid list [type=list_type, input_value='[{"body": "{4 años inme...oplazos-(FTCRR).aspx"]}', input_type=str]
    For further information visit https://errors.pydantic.dev/2.8/v/list_type
Got 0 cited docs