Hakan Karaoguz

Any updates on this?

I recently saw a bug in the browser tool. This could be related. If the file names contain a "." before the extension then the annotations are not saved, e.g.,...

I will try it out but according to this [Article](https://akash97715.medium.com/data-preparation-for-rag-part-1-b8bee1130115) `detection_class_prob` is about the class confidence of the extracted section (Table, Header etc) in the PDF. I am more interested...

I have worked with UnstructuredURLLoader a bit too. Under the [hood](https://github.com/Unstructured-IO/unstructured/blob/main/unstructured/partition/html.py#L70) it uses requests library to fetch data. I have implemented my own solution to fetch the URL and parse...

I think you should check [here](https://github.com/hwchase17/langchain/blob/d85f57ef9cbbbd5e512e064fb81c531b28c6591c/langchain/chains/question_answering/__init__.py#L146) that `refine chain` has no parameter called `prompt` instead you should use `question_prompt` as input.

As a short term solution, you can decouple the Serp tool and the LLM. In the first prompt you can add the question to the prompt and order LLM to...