Build a Visual Search Engine with CLIP and Qdrant

1. 1
   Set Up Environment and Qdrant

   Clone the example repository and install dependencies. This project uses Qdrant as the vector database, so you'll need to start a Qdrant instance using Docker.
2. 2
   Prepare Your Image Dataset

   Place your image files (.jpg, .png) into the `data/` directory. If you don't have images, you can run the provided script to download the Unsplash Lite dataset.
3. 3
   Index Your Image Collection

   Run the `index.py` script. It will process each image in the `data/` folder, generate embeddings using the CLIP model, and upload them to your running Qdrant instance for searching.
4. 4
   Launch the Search Service and UI

   The project includes a Gradio interface for easy interaction. Run `app.py` to launch a web server that provides both the backend API and the frontend user interface.
5. 5
   Perform Visual Searches

   Open your browser to the URL provided by Gradio (usually http://127.0.0.1:7860). You can now search your image collection by typing a text description or uploading another image.

#!/bin/bash

# This script sets up and runs the Visual Search Engine.
# Make sure you have Docker, Python 3.8+, and pip installed.

# 1. Clone the repository and navigate into the project directory
echo "Cloning repository..."
git clone https://github.com/qdrant/examples.git
cd examples/visual-search-engine

# 2. Start Qdrant in a separate terminal
#    You must run this command in another terminal window before proceeding.
echo "*********************************************************************"
echo "IMPORTANT: In a SEPARATE terminal, run the following command:"
echo "docker run -p 6333:6333 -p 6334:6334 qdrant/qdrant"
echo "*********************************************************************"
read -p "Press [Enter] after you have started Qdrant..."

# 3. Install Python dependencies
echo "Installing dependencies..."
pip install -r requirements.txt

# 4. Download a sample dataset (optional, can be skipped if you have your own)
echo "Downloading sample dataset..."
python download_data.py

# 5. Index the images
echo "Indexing images... This may take a few minutes."
python index.py

# 6. Launch the Gradio application
echo "Launching the search UI..."
echo "Open your browser to the URL provided below (e.g., http://127.0.0.1:7860)"
python app.py