Feature/quiz4

.github/workflows/actions-demo.yml

@@ -20,9 +20,5 @@ jobs:
       - run: echo " This job's status is ${{ job.status }}."
       - name: Install dep
         run:  pip3 install --no-cache-dir -r requirements.txt
-      - name: Run pytest cases
-        run: pytest
-      - name: Run bash script to build and run docker image
-        run: |
-            chmod +x ./docker_run.sh
-            ./docker_run.sh
+      - name: Run Docker image
+        run: ./docker_run.sh

Dockerfile

@@ -0,0 +1,11 @@
+FROM python:3.8
+WORKDIR /app
+COPY app.py /app/
+#COPY utils.py /app/
+COPY requirements.txt /app/
+COPY saved_model /app/saved_model/
+RUN apt-get install libsm6 libxext6  -y
+RUN pip install -r requirements.txt
+VOLUME /app/saved_model
+EXPOSE 80
+CMD ["python", "app.py"]

api/app.py

@@ -1,10 +1,11 @@
 from flask import Flask,request
-
+from joblib import  load
+import numpy as np
 app = Flask(__name__)
 
 @app.route("/")
 def hello_world():
-    return "<p>Hello World!</p>"
+    return "<p>Hello, World!</p>"
 
 @app.route('/user/<username>')
 def profile(username):
@@ -20,6 +21,21 @@ def sum_two_numbers(x,y):
 @app.route("/predict", methods = ['POST'])
 def predict_fn():
     input_data = request.get_json()
-    x = input_data['x']
-    y = input_data['y']
-    return f"sum of {x} and {y} is {x + y}"
+    img1 = input_data['img1']
+    img2 = input_data['img2']
+
+    img1 = list(map(float, img1))
+    img2 = list(map(float, img2))
+
+    img1 = np.array(img1).reshape(1,-1)
+    img2 = np.array(img2).reshape(1,-1)
+
+    model = load("models/svm_gamma:0.01_C:0.1.joblib")
+    predicted1 = model.predict(img1)
+    predicted2 = model.predict(img2)
+    if predicted1[0] == predicted2[0]:
+        print("Both images are of same digit")
+        return "True"
+    else:
+        print("Both images are not of same digit")
+        return "False"

api/app_2.py

@@ -0,0 +1,41 @@
+from flask import Flask,request
+from joblib import  load
+import numpy as np
+app = Flask(__name__)
+
+@app.route("/")
+def hello_world():
+    return "<p>Hello, World!</p>"
+
+@app.route('/user/<username>')
+def profile(username):
+    return f'{username}\'s profile'
+
+
+@app.route("/sum/<x>/<y>")
+def sum_two_numbers(x,y):
+    res = int(x) + int(y)
+    return f"sum of {x} and {y} is {res}"
+
+
+@app.route("/predict", methods = ['POST'])
+def predict_fn():
+    input_data = request.get_json()
+    img1 = input_data['img1']
+    img2 = input_data['img2']
+
+    img1 = list(map(float, img1))
+    img2 = list(map(float, img2))
+
+    img1 = np.array(img1).reshape(1,-1)
+    img2 = np.array(img2).reshape(1,-1)
+
+    model = load("models/svm_gamma:0.01_C:0.1.joblib")
+    predicted1 = model.predict(img1)
+    predicted2 = model.predict(img2)
+    if predicted1[0] == predicted2[0]:
+        print("Both images are of same digit")
+        return "True"
+    else:
+        print("Both images are not of same digit")
+        return "False"

app.py

@@ -0,0 +1,46 @@
+import pickle
+from flask import Flask, request, jsonify
+import numpy as np
+from PIL import Image
+import json
+
+app = Flask("assigmnet6")
+
+with open("saved_model/mnist_classification_svm.pkl", "rb") as f:
+    model = pickle.load(f)
+
+def preprocess(image_file):
+    with Image.open(image_file) as img:
+        img = img.convert('L')
+        img = img.resize((8, 8), Image.Resampling.LANCZOS)
+        image_array = np.array(img, dtype=np.float64)
+        image_array = image_array.reshape(1, -1)
+        image_array *= (16 / 255) 
+    return image_array
+
+def predict_digit(image):
+    prediction = model.predict(image)
+    return prediction
+
+@app.route('/run_test', methods=['GET'])
+def test():
+    return jsonify({'run_test': 'OK'})
+
+@app.route('/predict', methods=['POST'])
+def check_same_digit():
+    try:
+        print(request.files.keys())
+        image1 = request.files['image1']
+        image2 = request.files['image2']
+
+        digit1 = predict_digit(preprocess(image1))
+        digit2 = predict_digit(preprocess(image2))
+
+        same_digit = digit1[0] == digit2[0]
+
+        return json.dumps({'digit match': bool(same_digit)})
+    except Exception as e:
+        return jsonify({'error': str(e)})
+
+if __name__ == '__main__':
+    app.run(host='0.0.0.0', port=80, debug=True)

docker/Dockerfile

@@ -25,6 +25,8 @@ WORKDIR /digits
 #create volume to mount it on host
 VOLUME /digits/models
 #run pytest 
-CMD ["pytest"]
+#CMD ["pytest"]
+ENV FLASK_APP=api/app2.py
+CMD ["flask","run","--host=0.0.0.0"]
 #run python script to train model
-ENTRYPOINT ["python","exp.py"]
+# ENTRYPOINT ["python","exp.py"]

docker_run.sh

@@ -1,12 +1,3 @@
-docker build -t mlops_assignment_4:v1 -f docker/Dockerfile .
-echo "=======================================BEFORE EXECUTING CONTAINERS RUN================================================"
-ls -lh models
-echo "======================================================================================================================"
-
-docker run -v /mnt/c/Users/soura/Desktop/ML-OPS/Digits-Classification/models:/digits/models  mlops_assignment_4:v1 --total_run 1 --dev_size 0.3 --test_size 0.2 --model_type 'svm' 
-docker run -v /mnt/c/Users/soura/Desktop/ML-OPS/Digits-Classification/models:/digits/models  mlops_assignment_4:v1 --total_run 1 --dev_size 0.3 --test_size 0.2 --model_type 'tree'
-
-echo "=======================================AFTER EXECUTING CONTAINERS RUN================================================"
-ls -lh models
-echo "======================================================================================================================"
+docker build -t flask_app . --no-cache
 
+docker run --mount source=saved_model,destination=/app/saved_model flask_app

flask_app.py

@@ -0,0 +1,41 @@
+import pickle
+from flask import Flask, request, jsonify
+import numpy as np
+from PIL import Image
+import json
+
+app = Flask("quiz4")
+
+with open("saved_model/mnist_classification_svm.pkl", "rb") as f:
+    model = pickle.load(f)
+
+def preprocess(image_file):
+    with Image.open(image_file) as img:
+        img = img.convert('L')
+        img = img.resize((8, 8), Image.Resampling.LANCZOS)
+        image_array = np.array(img, dtype=np.float64)
+        image_array = image_array.reshape(1, -1)
+        image_array *= (16 / 255) 
+    return image_array
+
+def predict_digit(image):
+    prediction = model.predict(image)
+    return prediction
+
+@app.route('/test', methods=['GET', 'POST'])
+def test():
+    print("test")
+    return jsonify({'test': 'hit'})
+
+@app.route('/predict', methods=['POST'])
+def predict():
+    try:
+        image = request.files['image']
+        digit = predict_digit(preprocess(image))[0]
+
+        return json.dumps({'prediction': int(digit)})
+    except Exception as e:
+        return jsonify({'error': str(e)})
+
+if __name__ == '__main__':
+    app.run(host='0.0.0.0', port=5000, debug=True)

mnist_classification.py

@@ -0,0 +1,31 @@
+import numpy as np
+from sklearn import datasets
+from sklearn.model_selection import train_test_split
+from sklearn import svm
+from sklearn import metrics
+import pickle
+
+digits = datasets.load_digits()
+
+n_samples = len(digits.images)
+data = digits.images.reshape((n_samples, -1))
+
+X_train, X_test, y_train, y_test = train_test_split(data, digits.target, test_size=0.2, random_state=42)
+
+clf = svm.SVC(kernel='linear')
+
+clf.fit(X_train, y_train)
+
+y_pred = clf.predict(X_test)
+
+accuracy = metrics.accuracy_score(y_test, y_pred)
+print(f"Accuracy: {accuracy * 100:.2f}%")
+
+print("Classification Report:")
+print(metrics.classification_report(y_test, y_pred))
+
+print("Confusion Matrix:")
+print(metrics.confusion_matrix(y_test, y_pred))
+
+with open("./saved_model/mnist_classification_svm.pkl", "wb") as f:
+    pickle.dump(clf, f)

requirements.txt

@@ -1,5 +1,6 @@
-Flask==3.0.0
+matplotlib==3.7.2
+pandas==2.1.1
 matplotlib==3.7.2
 scikit-learn==1.3.0
 pytest==7.4.2
-pandas==2.1.1
+flask==3.0.0

test_flask_app.py

@@ -0,0 +1,39 @@
+import unittest
+from flask import Flask
+from flask_app import app  # Import your Flask app instance
+import json
+import requests
+import io
+import logging
+import tensorflow as tf
+from PIL import Image
+
+
+app = app.test_client()
+
+def convert_to_image(array):
+    image = Image.fromarray((array * 255).astype('uint8'), 'L')
+    img_byte_arr = io.BytesIO()
+    image.save(img_byte_arr, format='PNG')
+    img_byte_arr = img_byte_arr.getvalue()
+    return img_byte_arr
+
+def test_post_predict():
+
+    mnist = tf.keras.datasets.mnist
+    (_, _), (x_test, y_test) = mnist.load_data()
+
+    for digit in range(10):
+        index = list(y_test).index(digit)
+        sample_data = x_test[index]
+
+        image_data = convert_to_image(sample_data)
+
+        response = app.post("/predict", data={'image': (io.BytesIO(image_data), f'{digit}.png')})
+        predicted_digit = json.loads(response.text)['prediction']
+
+        assert response.status_code == 200
+        try:
+            assert predicted_digit == digit
+        except Exception as e:
+            print(f"predicted_digit: {predicted_digit}\t digit: {digit}")