Video Frame Extractor API

A Python API that extracts a frame from a video at a specific time.

Installation

pip install -r requirements.txt

Running the API

uvicorn main:app --reload

The API will be available at http://localhost:8000

API Endpoints

POST /extract-frame

Extract a frame from a video URL at a specific time.

Request Body:

{
  "video_url": "https://example.com/video.mp4",
  "time": 5.5
}

Parameters:

• video_url (string): URL of the video file
• time (float): Time in seconds (e.g., 5.5 for 5.5 seconds)

Response:

• Returns a PNG image of the frame at the specified time

Example using curl:

curl -X POST "http://localhost:8000/extract-frame" \
  -H "Content-Type: application/json" \
  -d '{"video_url": "https://example.com/video.mp4", "time": 5.5}' \
  --output frame.png

Example using Python:

import requests

response = requests.post(
    "http://localhost:8000/extract-frame",
    json={
        "video_url": "https://example.com/video.mp4",
        "time": 5.5
    }
)

with open("frame.png", "wb") as f:
    f.write(response.content)

API Documentation

Once the server is running, you can access:

• Interactive API docs: http://localhost:8000/docs
• Alternative docs: http://localhost:8000/redoc

Deployment to Google Cloud Run

Prerequisites

1. Install Google Cloud SDK
2. Install Docker
3. Authenticate with Google Cloud:

   gcloud auth login
   gcloud auth configure-docker

Option 1: Deploy using the deployment script

1. Set your Google Cloud project ID:

   export GOOGLE_CLOUD_PROJECT="your-project-id"

2. Make the script executable and run it:

   chmod +x deploy.sh
   ./deploy.sh

Option 2: Deploy using Cloud Build

1. Set your Google Cloud project ID:

   export GOOGLE_CLOUD_PROJECT="your-project-id"

2. Make the script executable and run it:

   chmod +x deploy-cloud-build.sh
   ./deploy-cloud-build.sh

Option 3: Manual deployment

1. Build and push the Docker image:

   export PROJECT_ID="your-project-id"
   docker build -t gcr.io/${PROJECT_ID}/video-frame-extractor:latest .
   docker push gcr.io/${PROJECT_ID}/video-frame-extractor:latest

2. Deploy to Cloud Run:

   gcloud run deploy video-frame-extractor \
     --image gcr.io/${PROJECT_ID}/video-frame-extractor:latest \
     --platform managed \
     --region us-central1 \
     --allow-unauthenticated \
     --memory 2Gi \
     --cpu 2 \
     --timeout 300

Configuration

The deployment uses:

• Memory: 2GB (adjustable for larger videos)
• CPU: 2 cores
• Timeout: 300 seconds (5 minutes)
• Region: us-central1 (changeable)

After deployment, you'll receive a URL like: https://video-frame-extractor-xxxxx-uc.a.run.app

You can then use this URL instead of localhost:8000 in your API calls.