CAR CHATBOT

Overview

This project is a chatbot designed for evaluating car-related queries. It leverages a combination of retrieval-augmented generation (RAG) and NLP-based evaluation metrics.

Project Structure

CAR CHATBOT
│-- car_chroma_db/               # ChromaDB storage
│   │-- chroma.sqlite3           # SQLite database for embeddings
│-- src/
│   │-- testing_codes/           # Contains evaluation scripts and datasets
│   │-- chunking.py              # Script for data preprocessing and chunking
│   │-- data_eval.py             # Evaluation script for the dataset
│   │-- embedding_store.py       # Script to store and retrieve embeddings
│   │-- eval_testing_03.py       # Main evaluation script
│   │-- evaluation_results.json  # JSON file with evaluation results
│   │-- evaluation.py            # Evaluation logic implementation
│   │-- llm_rag.py               # Core chatbot logic
│   │-- main.py                  # Main execution script
│   │-- more_details_prompt.txt  # Additional prompts for chatbot
│   │-- requirement.txt          # Required dependencies
│   │-- scrapper.py              # Web scraping script
│   │-- streamlit.py             # UI implementation using Streamlit
│-- car_evaluation_dataset.json  # Dataset for car evaluation
│-- cartrade_cars_chunked.json   # Processed data fed to the chatbot
│-- chatbot_evaluation_detailed.csv  # Detailed chatbot evaluation
│-- chatbot_evaluation_report.json  # JSON report of chatbot evaluation

Setup Instructions

1. Install Dependencies:

   pip install -r requirement.txt

2. Run the Chatbot:

   python llm_rag.py

3. Run the UI Interface:

   streamlit run streamlit.py

4. Run Evaluation:

   python eval_testing_03.py

   This script calculates BERT, ROUGE, METEOR, and F1 scores for evaluation.

Data Sources

• cartrade_cars_chunked.json contains the actual data fed to the model.
• The dataset has been chunked and stored for efficient retrieval.

Description of Key Files

• llm_rag.py - Contains the core logic for the chatbot.
• scrapper.py - Scrapes data from car-related sources.
• streamlit.py - Streamlit-based UI for interacting with the chatbot.
• eval_testing_03.py - Computes NLP-based evaluation metrics.