I use the following code, saving the scaler and models from a model that predicts heart disease. I’m now trying to use the code to send a JSON via Postman, but I’m running into the following error: TypeError: Object of type ndarray is not JSON serializable. After doing some Googling, it looks like I would have to use json.dumps(cls=NumPyArrayEncoder) or pd.Series(your_array).to_json(orient=”values”). What I don’t understand is where to add this? What do I use this code on, the model before it’s saved as an h5? Do I modify the inputs for the model? Do I also use json dumps on the saved scalers?

from flask import Flask, request, jsonify
import numpy as np
from tensorflow.keras.models import load_model
import joblib

def return_prediction(model, scaler, sample_json):

    age = sample_json["age"]
    sex = sample_json["sex"]
    cp = sample_json['cp']
    trestbps = sample_json['trestbps']
    chol = sample_json['chol']
    fbs = sample_json['fbs']
    restecg = sample_json['restecg']
    thalach = sample_json['thalach']
    exang = sample_json['exang']
    oldpeak = sample_json['oldpeak']
    slope = sample_json['slope']
    ca = sample_json['ca']
    thal = sample_json['thal']

    heart = [[age, sex, cp, trestbps, chol, fbs, restecg, 
    thalach, exang, oldpeak, slope, ca, thal]]

    heart = scaler.transform(heart)

    classes = np.array(['0','1'])

    class_ind = model.predict(heart)

    return classes

app = Flask(__name__)

@app.route("/")
def index():
    return '<h1>FLASK RUNNING</h1>'

heart_model = load_model("heart_disease_model_bry.h5")
heart_scaler = joblib.load("heart_scaler_bry.pkl")

@app.route('/api/heart',methods=['POST'])
def prediction():
    content = request.json
    results = return_prediction(heart_model, heart_scaler, content)
    return jsonify(results)

if __name__=='__main__':
    app.run()

JSON using as a sample to request a prediction via Postman:

{"age":53,
"sex":1,
"cp":0,
"trestbps":125,
"chol":212,
"fbs":0,
"restecg":1,
"thalach":168,
"exang":0,
"oldpeak":1.0,
"slope":2,
"ca":2,
"thal":3}