# Advise on Chatbot development

**URL:** <https://discuss.huggingface.co/t/advise-on-chatbot-development/112313>\
**Category:** Beginners\
**Created:** [October 16, 2024, 4:12am UTC](https://discuss.huggingface.co/t/advise-on-chatbot-development/112313 "2024-10-16T04:12:12Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Arulprasath](https://avatars.discourse-cdn.com/v4/letter/a/ecae2f/32.png) [@Arulprasath](https://discuss.huggingface.co/u/Arulprasath)\
**Post date:** [October 16, 2024, 4:12am UTC](https://discuss.huggingface.co/t/advise-on-chatbot-development/112313/1 "2024-10-16T04:12:12Z")

</div>

Hi Team,

As I choose I’m a beginner , I am developing the Chatbot  
Usecase : chat with a document source(Size less than 1GB).  
Current State : Stored the document chunk and retrieving it based on the user question.  
Now I wanted to beautify and structure the junk and give as a response.

**To do that beatifying what model I have to use ? any advise ? Please help to address me with next step.**

---

<div class="post-metadata">

**Author:** ![mahmutc](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/mahmutc/32/64626_2.png) [@mahmutc](https://discuss.huggingface.co/u/mahmutc)\
**Post date:** [October 16, 2024, 6:56am UTC](https://discuss.huggingface.co/t/advise-on-chatbot-development/112313/2 "2024-10-16T06:56:36Z")

</div>

hi @Arulprasath  
We’re talking about HTML Beautifying, right? Why don’t you use simple solution like [Beautiful Soup Documentation — Beautiful Soup 4.4.0 documentation](https://beautiful-soup-4.readthedocs.io/en/latest/) to get text?

I guess you’re building kind of RAG. If your system can retrieve relevant part based on the user question, you might give a shoot `question answering` pipeline:

> **[Transformers, what can they do? - Hugging Face NLP Course](https://huggingface.co/learn/nlp-course/chapter1/3#question-answering)**
>
> We’re on a journey to advance and democratize artificial intelligence through open source and open science.

```auto
from transformers import pipeline

question_answerer = pipeline("question-answering")
question_answerer(
    question="user question",
    context="context what you retrieve and beautify",
)

```

---

<div class="post-metadata">

**Author:** ![Arulprasath](https://avatars.discourse-cdn.com/v4/letter/a/ecae2f/32.png) [@Arulprasath](https://discuss.huggingface.co/u/Arulprasath)\
**Post date:** [October 17, 2024, 1:46am UTC](https://discuss.huggingface.co/t/advise-on-chatbot-development/112313/3 "2024-10-17T01:46:31Z")

</div>

Hi @mahmutc , Thanks for the info given.

I’m not looking for the UI cosmetic changes. That streamlit can help.

I finished until the chunk matching for the user question. In the junk I am getting the exact text i loaded with a broken statement. as a next step what I can use for correcting and structuring the mach chunk ?
