Car Model Vector Search System

A semantic search system for car models using vector embeddings and LLM-powered response generation.

Overview

This project implements a vector similarity search system for car models using sentence transformers and ChromaDB. It includes capabilities to:

• Create and store vector embeddings for car model data
• Perform semantic similarity searches
• Generate contextual responses using LLMs (via Ollama)

Technologies Used

• Python 3.x
• ChromaDB (Vector Database)
• Sentence Transformers (BAAI/bge-small-en-v1.5)
• Ollama (Local LLM integration)
• Pandas (Data handling)
• LangChain (LLM orchestration)

Project Structure

.
├── embeddings.py
├── query.py
├── vector_store.py
├── README.md
├── LICENSE.txt
├── requirements.txt
├── .gitignore

Setup

1. Create a virtual environment:

python -m venv env
source env/bin/activate  # On Windows: env\Scripts\activate

2. Install dependencies:

pip install -r requirements.txt

3. Install Ollama and download required model:

ollama pull deepseek-r1:1.5b

4. Prepare your data:

• Place your car model dataset in train.csv
• Required columns: Id, Key, Value

Usage

1. Generate Embeddings

Run the embedding generation script:

python embeddings.py

This will:

• Load data from train.csv
• Create vector embeddings
• Store them in ChromaDB

2. Perform Queries

Use the query system:

from vector_store import VectorStore
from query import Query

# Initialize
vector_store = VectorStore("BAAI/bge-small-en-v1.5", "car_model", "Car Model Collection", "./vector_embedding")
query = Query(vector_store, "deepseek-r1:1.5b")

# Search
input_query = "aviator"
retrieved_data = query.query(input_query, 10)

# Get LLM-enhanced response
query_obj = {
    "query": input_query,
    "keys": retrieved_data['keys'],
    "values": retrieved_data['values'],
    "distances": retrieved_data['distances']
}
response = query.get_response(query_obj)

Configuration

The vector store can be configured with the following parameters:

• hnsw_space: Vector similarity metric (default: 'cosine')
• hnsw_construction: HNSW construction parameter (default: 100)
• hnsw_search_ef: HNSW search parameter (default: 100)
• hnsw_M: HNSW M parameter (default: 100)

Contributing

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Push to the branch
5. Create a Pull Request

License

This project is open-sourced under the MIT License - see the LICENSE file for details.