app.py

from flask import Flask, request, jsonify
import cv2
import numpy as np
import tensorflow as tf
from flask_cors import CORS
import base64

app = Flask(__name__)
API_URL = '127.0.0.1'
API_PORT = 5000
LARAVEL_APP_URL = 'http://127.0.0.1:8000'
CORS(app, origins=LARAVEL_APP_URL, allow_methods=["GET", "POST"])

IMAGE_SHAPE = (224, 224)
demo_model = tf.keras.models.load_model("./models/mobilenet_model.h5")
# model_v2 = tf.keras.models.load_model("./models/NIH_Seresnet152_model.h5")

demo_model_classes = {
    0: 'Atelectasis',
    1: 'Cardiomegaly',
    2: 'Effusion',
    3: 'Infiltration',
    4: 'Mass',
    5: 'Nodule',
    6: 'Pneumonia',
    7: 'Pneumothorax',
    8: 'Consolidation',
    9: 'Edema',
    10: 'Emphysema',
    11: 'Fibrosis',
    12: 'Pleural_Thickening',
    13: 'Hernia'
}
model_v2_classes = [
    'Cardiomegaly', 'Hernia', 'Infiltration', 'Nodule', 'Emphysema', 'Effusion',
    'Atelectasis', 'Pleural_Thickening', 'Pneumothorax', 'Mass', 'Fibrosis',
    'Consolidation', 'Edema', 'Pneumonia'
]


def decode_image64(image64):
    try:
        if image64 is None:
            raise (Exception('Image64 is Null'))
        prefix, image64 = image64.split(',', 1)
        image_type = prefix.split(';')[0].split(':')[1]
        image64 += '=' * ((4 - len(image64) % 4) % 4)
        image_data = base64.b64decode(image64)
        return image_data
    except Exception as e:
        return jsonify({'error': 'decoding base64 image failed'}), 400


@app.route('/', methods=['GET'])
def welcome():
    response = "<h1 style='color:#04aa6d'>welcome to doctor ai collab web api</h1>"
    return response


@app.route('/info', methods=['GET'])
def info():
    return {
        "demo_model": {
            "accuracy": 20,
            "classes": 'demo_model',
            "predict": 'demo_model/predict',
        },
        # "model_v2": {
        #     "accuracy": 65,
        #     "classes": 'demo_model',
        #     "predict": 'demo_model/predict',
        # },
    }


@app.route('/demo_model', methods=['GET'])
def demo_model_info():
    response = "<h1 style='color:#04aa6d'>Supported classes by model-v1:</h1>"
    response += "<ul>"
    for val in demo_model_classes.values():
        response += f"<li>{val}</li>"
    response += "</ul>"
    return response


@app.route('/demo_model/predict', methods=['POST'])
def demo_model_predict():
    try:
        data = request.get_json()
        image64 = data.get('image64')
        if image64 is None:
            return jsonify({'error': 'Base64 image not provided'}), 400
    except Exception as e:
        return jsonify({'error': 'Bad request format'}), 400

    try:
        image_data = decode_image64(image64)
        image = cv2.imdecode(np.frombuffer(image_data, np.uint8), cv2.IMREAD_COLOR)
    except Exception as e:
        return jsonify({'error': 'Image Processing Failed!'}), 400

    resized_image = cv2.resize(image, IMAGE_SHAPE)
    processed_image = np.expand_dims(resized_image, axis=0)
    predictions = demo_model.predict(processed_image)
    results = {
        demo_model_classes[class_index]: "{:.2f}".format(float(predictions[0][class_index]) * 100)
        for class_index in range(len(demo_model_classes))
    }
    return jsonify(results)


# @app.route('/model_v2', methods=['GET'])
# def model_v2_info():
#     response = "<h1 style='color:#04aa6d'>Supported classes by model-v2:</h1>"
#     response += "<ul>"
#     for val in model_v2_classes:
#         response += f"<li>{val}</li>"
#     response += "</ul>"
#     return response


# @app.route('/model_v2/predict', methods=['POST'])
# def model_v2_predict():
#     img_size = 600
#     try:
#         data = request.get_json()
#         image64 = data.get('image64')
#         if image64 is None:
#             return jsonify({'error': 'Base64 image not provided'}), 400
#     except Exception as e:
#         return jsonify({'error': 'Bad request format'}), 400
#
#     try:
#         image_data = decode_image64(image64)
#         image = cv2.imdecode(np.frombuffer(image_data, np.uint8), cv2.IMREAD_COLOR)
#     except Exception as e:
#         return jsonify({'error': 'Image Processing Failed!'}), 400
#
#     resized_image = cv2.resize(image, (img_size, img_size))
#     processed_image = np.expand_dims(resized_image, axis=0)
#
#     # Use the loaded model for prediction
#     predictions = model_v2.predict(processed_image)
#
#     results = {
#         model_v2_classes[class_index]: "{:.2f}".format(float(predictions[0][class_index]) * 100)
#         for class_index in range(len(model_v2_classes))
#     }
#
#     return jsonify(results)


if __name__ == '__main__':
    app.run(host=API_URL, port=API_PORT)