DiseaseDetectionAPI

A python flask rest api that use a trained tensorflow keras model from AWS S3 for predicting disease , deployed on AWS EC2 Ubuntu instance with NGINX . Application Performance Monitoring done by Elastic APM through elastic apm agent configured on the flask app.

Take a look at the right side of the block diagram

AI Disease Detector API Documentation

Endpoints

Predict Disease

• Endpoint: /predict
• Method: POST
• Description: Predicts the disease in a plant image.
• Request Body:
  • Form data with key file containing the plant image file.
• Response:
  • If successful, returns a JSON object with the predicted disease class.
  • If unsuccessful, returns an error message.

Health Check

• Endpoint: /
• Method: GET
• Description: A simple health check endpoint to verify if the API is running.
• Response: Returns a welcome message indicating that the API is operational.

Error Handling

• If any errors occur during the prediction process, appropriate error messages are returned along with HTTP status codes.

Usage Example

import requests

API endpoint

url = 'https://ec2_instance_ip_or_domain/predict'

Load image file

files = {'file': open('plant_image.jpg', 'rb')}

Send prediction request

response = requests.post(url, files=files)

Print prediction result

print(response.json())

Installation and Setup

1. Clone the repository or download the source code.
2. Install required Python packages using pip install -r requirements.txt.
3. Set up AWS S3 bucket with the trained model file.
4. Configure NGINX to forward requests to the Flask app.
5. Set up Elastic APM and configure the Flask application with the Elastic APM agent.
6. Obtain an SSL/TLS certificate using Python Certbot for securing communications.
7. Assign a domain name to your AWS EC2 instance.
8. Implement CORS to allow requests only from the Floravision frontend.

Configuration

AWS S3 Configuration

• Ensure that the AWS credentials with access to the S3 bucket are properly configured.
• Update the download_model_from_s3 function with the appropriate S3 bucket name and model file name.

NGINX Configuration

• Configure NGINX to forward requests to the Flask app running on a port.
• Set up SSL/TLS certificate obtained from Python Certbot for secure communicatio

Elastic APM Configuration

• Set up an Elastic APM instance and obtain the service name, secret token, and server URL.
• Update the app.config['ELASTIC_APM'] dictionary with the obtained credentials.

CORS Configuration

• Implement CORS to only allow requests from the Floravision frontend. Update the CORS configuration in the Flask app accordingly.

Dependencies

• Flask
• Flask-CORS
• ElasticAPM
• TensorFlow 2.11
• Boto3