Vector Database Project

This project provides a modern admin dashboard (FastAPI + HTML/JS) for uploading, viewing, deleting, embedding, and querying documents using a Chroma vector database. It also demonstrates how to create a traditional SQL database with SQLAlchemy and build a vector database from uploaded documents.

Features

• Modern admin dashboard at /admin (upload, view, delete, embed, query)
• Upload documents via web interface
• Store document metadata and content in a SQL database (SQLAlchemy)
• Generate vector embeddings from documents (Sentence Transformers)
• Store and query embeddings using Chroma vector database
• Progress bars for embedding and querying
• View all files and delete from dashboard
• Inspect Chroma DB contents with a script

---

Getting Started

1. Clone the Repository

git clone <your-repo-url>
cd Vector-database

2. Install Dependencies

Make sure you have Python 3.8+ installed. Then install the required packages:

pip install -r requirements.txt

Main dependencies:

• fastapi, uvicorn (web server)
• sqlalchemy (SQL database)
• sentence-transformers (embeddings)
• chromadb (vector database)

---

Usage

1. Create a SQL Database with SQLAlchemy

Below is a minimal example of how to define a document table and create a database using SQLAlchemy:

from sqlalchemy import create_engine, Column, Integer, String, Text
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import sessionmaker

Base = declarative_base()

class Document(Base):
    __tablename__ = 'documents'
    id = Column(Integer, primary_key=True)
    filename = Column(String, unique=True)
    content = Column(Text)

# Create SQLite database
engine = create_engine('sqlite:///documents.db')
Base.metadata.create_all(engine)
Session = sessionmaker(bind=engine)
session = Session()

2. Use the Admin Dashboard

• Go to http://127.0.0.1:8000/admin in your browser.
• Add File: Upload documents (content and metadata stored in SQL database).
• View Files: See all uploaded files and delete as needed.
• Embed Files: Click to generate and store embeddings in Chroma (progress bar shown).
• Query: Enter a text query to search the vector database (progress bar shown, references returned).

3. Generate Embeddings and Store in Chroma

• Use the "Embed Files" button in the dashboard, or run:

  python store_in_chroma.py

4. Query the Vector Database

• Use the "Query" tab in the dashboard, or run a script to query Chroma directly.

5. Inspect Chroma DB

• Run the script:

  python view_chroma_db.py

  This prints the content of the first document in Chroma. You can modify the script to see more.

---

References

• Chroma Documentation
• SQLAlchemy Documentation
• Sentence Transformers
• FastAPI

---

License

MIT