|
--- |
|
title: RAG Excel Chatbot |
|
emoji: 📊 |
|
colorFrom: green |
|
colorTo: blue |
|
sdk: streamlit |
|
app_file: app.py |
|
pinned: false |
|
license: apache-2.0 |
|
sdk_version: 1.44.1 |
|
--- |
|
|
|
# Excel-Aware RAG Chatbot |
|
|
|
This Streamlit app lets you upload an Excel file (with a 'Data Base' sheet), processes the data into a retrievable vector database, and allows question answering using a RAG pipeline powered by `flan-t5-base`. |
|
|
|
## Features |
|
- Upload `.xlsm` or `.xlsx` files |
|
- Automatically cleans and processes the 'Data Base' sheet |
|
- Embeds entries using `MiniLM` embeddings |
|
- Uses `flan-t5-base` for fast CPU-friendly QA |
|
- Works on Hugging Face Spaces (L4 hardware) |
|
|
|
## Run Locally |
|
|
|
```bash |
|
pip install -r requirements.txt |
|
streamlit run app.py |
