Upload app.py

app.py

@@ -3,8 +3,6 @@ import pandas as pd
 import seaborn as sns
 import matplotlib.pyplot as plt
 import joblib
-from sklearn.tree import DecisionTreeClassifier, XGBClassifier                   #using sklearn decisiontreeclassifier
-from sklearn.model_selection import train_test_split
 
 import os
 import shutil
@@ -22,7 +20,6 @@ else:
 
 data=pd.read_csv('data/heart.xls')
 
-
 data.info()   #checking the info
 
 data_corr=data.corr()
@@ -30,41 +27,7 @@ data_corr=data.corr()
 plt.figure(figsize=(20,20))
 sns.heatmap(data=data_corr,annot=True)
 #Heatmap for data
-"""
-# Get the Data
-X_train, y_train, X_val, y_val = train_test_split()
-classifier = XGBClassifier()
-# Training the Model
-classifier = classifier.fit(X_train, y_train)
-# Trained Model Evaluation on Validation Dataset
-confidence = classifier.score(X_val, y_val)
-# Validation Data Prediction
-y_pred = classifier.predict(X_val)
-# Model Validation Accuracy
-accuracy = accuracy_score(y_val, y_pred)
-# Model Confusion Matrix
-conf_mat = confusion_matrix(y_val, y_pred)
-# Model Classification Report
-clf_report = classification_report(y_val, y_pred)
-# Model Cross Validation Score
-score = cross_val_score(classifier, X_val, y_val, cv=3)
-
-try:
-    # Load Trained Model
-    clf = load(str(self.model_save_path + saved_model_name + ".joblib"))
-except Exception as e:
-    print("Model not found...")
-
-if test_data is not None:
-    result = clf.predict(test_data)
-    print(result)
-else:
-    result = clf.predict(self.test_features)
-accuracy = accuracy_score(self.test_labels, result)
-clf_report = classification_report(self.test_labels, result)
-print(accuracy, clf_report)
-"""
-####################
+
 feature_value=np.array(data_corr['output'])
 for i in range(len(feature_value)):
     if feature_value[i]<0:
@@ -82,6 +45,10 @@ feature_selected     #selected features which are very much correalated
 
 clean_data=data[feature_selected]
 
+from xgboost import XGBClassifier
+from sklearn.tree import DecisionTreeClassifier                  #using sklearn decisiontreeclassifier
+from sklearn.model_selection import train_test_split
+
 #making input and output dataset
 X=clean_data.iloc[:,1:]
 Y=clean_data['output']
@@ -97,7 +64,7 @@ x_train=sc.fit_transform(x_train)
 x_test=sc.transform(x_test)
 
 #training our model
-dt=XGBClassifier(criterion='entropy',max_depth=6)
+dt=XGBClassifier(max_depth=6)
 dt.fit(x_train,y_train)
 #dt.compile(x_trqin)
 
@@ -113,10 +80,9 @@ print("\nThe accuracy of decisiontreelassifier on Heart disease prediction datas
 
 joblib.dump(dt, 'heart_disease_dt_model.pkl')
 
-from concrete.ml.sklearn import DecisionTreeClassifier as ConcreteDecisionTreeClassifier
-from concrete.ml.sklearn import XGBClassifier as ConcreteXGBClassifier
+from concrete.ml.sklearn.tree import XGBClassifier as ConcreteXGBClassifier
 
-fhe_compatible = ConcreteXGBClassifier.from_sklearn_model(dt, x_train, n_bits = 10) #de FHE
+fhe_compatible = ConcreteXGBClassifier.from_sklearn_model(dt, x_train, n_bits = 10)
 fhe_compatible.compile(x_train)
 
 
@@ -134,3 +100,51 @@ dev.save()
 # Setup the server
 server = FHEModelServer(path_dir=fhe_directory)
 server.load()
+
+
+
+
+
+
+
+####### client
+
+from concrete.ml.deployment import FHEModelDev, FHEModelClient, FHEModelServer
+
+# Setup the client
+client = FHEModelClient(path_dir=fhe_directory, key_dir="/tmp/keys_client")
+serialized_evaluation_keys = client.get_serialized_evaluation_keys()
+
+
+# Load the dataset and select the relevant features
+data = pd.read_csv('data/heart.xls')
+
+# Perform the correlation analysis
+data_corr = data.corr()
+
+# Select features based on correlation with 'output'
+feature_value = np.array(data_corr['output'])
+for i in range(len(feature_value)):
+    if feature_value[i] < 0:
+        feature_value[i] = -feature_value[i]
+
+features_corr = pd.DataFrame(feature_value, index=data_corr['output'].index, columns=['correlation'])
+feature_sorted = features_corr.sort_values(by=['correlation'], ascending=False)
+feature_selected = feature_sorted.index
+
+# Clean the data by selecting the most correlated features
+clean_data = data[feature_selected]
+
+# Extract the first row of feature data for prediction (excluding 'output' column)
+sample_data = clean_data.iloc[0, 1:].values.reshape(1, -1)  # Reshape to 2D array for model input
+
+encrypted_data = client.quantize_encrypt_serialize(sample_data)
+
+
+
+##### end client
+
+encrypted_result = server.run(encrypted_data, serialized_evaluation_keys)
+
+result = client.deserialize_decrypt_dequantize(encrypted_result)
+print(result)