deploy the ml model to flask

.env.dev

@@ -0,0 +1,3 @@
+FLASK_APP=project/app.py
+FLASK_DEBUG=1
+FLASK_RUN_PORT=8002

.env.prod

@@ -0,0 +1,3 @@
+FLASK_APP=project/app.py
+FLASK_DEBUG=0
+FLASK_RUN_PORT=8002

Dockerfile.dev

@@ -0,0 +1,11 @@
+FROM python:3.10-slim-buster
+
+ENV PYTHONUNBUFFERED=1
+
+WORKDIR /carbonzero-ml-model-dir
+
+COPY ./carbonzero_ml_model/requirements.txt ./requirements.txt
+COPY ./carbonzero_ml_model/.env.dev ./.env.dev
+RUN pip3 install -r requirements.txt
+
+COPY ./carbonzero_ml_model/. .

Dockerfile.prod

@@ -0,0 +1,11 @@
+FROM python:3.10-slim-buster
+
+ENV PYTHONUNBUFFERED=1
+
+WORKDIR /carbonzero-ml-model-dir
+
+COPY ./carbonzero_ml_model/requirements.txt ./requirements.txt
+COPY ./carbonzero_ml_model/.env.prod ./.env.prod
+RUN pip3 install -r requirements.txt
+
+COPY ./carbonzero_ml_model/. .

manage.py

@@ -0,0 +1,8 @@
+from flask.cli import FlaskGroup
+from project.app import app
+
+cli = FlaskGroup(app)
+
+if __name__ == "__main__":
+    cli()
+

project/app.py

@@ -0,0 +1,29 @@
+from flask import Flask, jsonify, Blueprint, request
+from flask_restx import Api, Resource, fields, fields, abort
+from .model import model
+import numpy as np
+
+app_ins = Flask(__name__)
+
+blueprint = Blueprint('api', __name__, url_prefix='/api/ml-model')
+app = Api(blueprint, default="API", default_label="ML Model API")
+
+app_ins.register_blueprint(blueprint)
+
+carbon_emission_fields = app.model('Total Emission', {
+    'total_emission': fields.Integer,
+})
+@app.route("/carbon_emission_comparison")
+class MyResource(Resource):
+    @app.expect(carbon_emission_fields, validate=True)
+    def post(self):
+        try:
+            total_emission = float(request.json['total_emission'])
+            prediction = model.predict(np.array([[total_emission]]))
+
+            predicted_class = prediction.argmax()
+            return jsonify(result=int(predicted_class))
+        except ValueError:
+            abort(400, "Invalid total_emission value")
+        except Exception as e:
+            abort(400, str(e))

project/model.py

@@ -0,0 +1,36 @@
+import tensorflow as tf # for model arc
+from tensorflow.keras.models import Sequential
+from tensorflow.keras.layers import Dense, LSTM
+import pandas as pd # for split dataset
+from sklearn.model_selection import train_test_split
+import numpy as np # for predict or test the model
+import pickle
+import os
+
+dataset = pd.read_csv('project/datasetKgCO2.csv', delimiter=',', header=0)
+
+x = dataset.iloc[:,:-1].values
+y = dataset.iloc[:,1].values
+
+# Split dataset into train and validation
+x_train, x_val, y_train, y_val = train_test_split(x, y, test_size=0.2, random_state=0)
+
+# Model architecture
+model = Sequential([
+    Dense(64, activation='relu', input_shape=(None, 1)),
+    Dense(32, activation='relu'),
+    LSTM(4, activation='relu'),
+    Dense(32, activation='relu'),
+    Dense(4, activation='linear')])
+
+# Optimizer and loss for the model
+model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.001),
+              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
+              metrics=['acc'])
+
+# Training the model
+model.fit(tf.expand_dims(x_train, axis=-1), y_train, epochs=100, validation_data=(x_val, y_val), steps_per_epoch=10)
+
+pickle.dump(model, open('model.pkl', 'wb'))
+
+model = pickle.load(open('model.pkl', 'rb'))

requirements.txt

@@ -0,0 +1,7 @@
+Flask==2.3.2
+gunicorn==20.1.0
+flask-restx==1.1.0
+tensorflow==2.12.0
+numpy==1.23.5
+pandas==2.0.2
+scikit-learn==1.2.2