#mobileClip-Search: Image Search Application

An image search application based on Apple's mobileClip-Search model, supporting both text-to-image and image-to-image search capabilities.

Features

• Text-to-image search
• Image-to-image similarity search
• Support for both local and AWS S3 storage
• Docker containerization
• AWS EC2 cloud deployment
• Real-time image processing and search
• Efficient vector indexing with FAISS

Tech Stack

• Backend: Flask
• Deep Learning: PyTorch
• Model: Apple mobileClip-Search
• Vector Search: FAISS
• Deployment: Docker, AWS EC2
• Storage: Local filesystem / AWS S3
• Web Server: Gunicorn

Quick Start

Local Development

1. Clone the repository and install dependencies:

git clone [your-repository-url]
cd mobileClip-Search
pip install -r requirements.txt

2. Download the model:

mkdir -p checkpoints
# Download mobileclip_s0.pt to the checkpoints directory

3. Configure environment variables:

cp .env.example .env
# Edit .env file with your configuration

4. Run the application:

python app.py

Docker Deployment

1. Build and run with Docker:

docker-compose up --build

AWS Deployment

1. Prepare AWS environment:

   • Create an EC2 instance
   • Configure security group (open port 80)
   • Create S3 bucket (if using S3 mode)

2. Configure deployment:

   • Copy SSH key to ~/.ssh/
   • Configure .env file

3. Deploy:

bash deploy/deploy.sh

Project Structure

mobileClip-Search/
├── app.py                 # Main application file
├── requirements.txt       # Python dependencies
├── templates/            # HTML templates
│   └── index.html        # Main page
├── checkpoints/          # Model directory
│   └── mobileclip_s0.pt  # mobileClip-Search model
├── images/              # Image library directory
├── cache/               # Cache directory
├── deploy/              # Deployment files
│   ├── deploy.sh        # Deployment script
│   ├── Dockerfile       # Docker configuration
│   └── docker-compose.yml # Docker Compose configuration
└── scripts/             # Utility scripts
    └── logs.sh          # Log viewing script

Environment Configuration

Configure the following variables in your .env file:

# AWS Credentials
AWS_ACCESS_KEY_ID=your_access_key_id
AWS_SECRET_ACCESS_KEY=your_secret_access_key
AWS_DEFAULT_REGION=ap-southeast-2
S3_BUCKET=your-bucket-name

# EC2 Configuration
EC2_IP=your.ec2.ip.address
EC2_KEY=path/to/your/key.pem

# Application Configuration
STORAGE_TYPE=s3  # or 'local'
FLASK_ENV=production

Usage Guide

1. Text Search:

   • Enter descriptive text in the search box
   • Click the "Search" button
   • View the most relevant images

2. Image Search:

   • Click the "Upload Image" button
   • Select an image file
   • View similar images

Maintenance and Monitoring

View Logs

# View real-time logs
bash scripts/logs.sh -f

# View last 100 lines
bash scripts/logs.sh -n 100

# View logs with timestamps
bash scripts/logs.sh -t

Update Deployment

# Update application
bash deploy/deploy.sh