# -*- coding: utf-8 -*-
"""Workshop-Linear kernel.ipynb

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1PlQo6At9kkwbacgmJdJ7FuN93WVLFoEY
"""

import numpy as np

from sklearn import datasets

from sklearn.model_selection import train_test_split

from sklearn.preprocessing import StandardScaler

from sklearn.svm import SVC

from sklearn.metrics import accuracy_score, classification_report, confusion_matrix

# Breast cancer dataset: target = malignant (0) / benign (1)

data = datasets.load_breast_cancer()

X = data.data      # features

y = data.target    # labels



print("Feature matrix shape:", X.shape)

print("Labels shape:", y.shape)

X_train, X_test, y_train, y_test = train_test_split(

    X, y, test_size=0.3, random_state=42, stratify=y

)

print("Train size:", X_train.shape[0], "Test size:", X_test.shape[0])

scaler = StandardScaler()

X_train_scaled = scaler.fit_transform(X_train)

X_test_scaled = scaler.transform(X_test)

linear_svm = SVC(kernel='linear', C=1.0, random_state=42)

linear_svm.fit(X_train_scaled, y_train)



y_pred_linear = linear_svm.predict(X_test_scaled)

acc_linear = accuracy_score(y_test, y_pred_linear)



print("\n=== Linear Kernel SVM (Dot Product) ===")

print("Accuracy:", acc_linear)

print("Confusion matrix:\n", confusion_matrix(y_test, y_pred_linear))

print("Classification report:\n", classification_report(y_test, y_pred_linear))

rbf_svm = SVC(kernel='rbf', C=1.0, gamma='scale', random_state=42)

rbf_svm.fit(X_train_scaled, y_train)



y_pred_rbf = rbf_svm.predict(X_test_scaled)

acc_rbf = accuracy_score(y_test, y_pred_rbf)



print("\n=== RBF Kernel SVM ===")

print("Accuracy:", acc_rbf)

print("Confusion matrix:\n", confusion_matrix(y_test, y_pred_rbf))

print("Classification report:\n", classification_report(y_test, y_pred_rbf))

X_small = X_train_scaled[:5]



# Dot product kernel matrix: K_ij = x_i · x_j

K_dot = X_small @ X_small.T



# RBF kernel matrix by hand

def rbf_kernel(X1, X2, gamma):

    # ||x - x'||^2 = (x^2 summed) + (x'^2 summed) - 2 x·x'

    X1_sq = np.sum(X1 ** 2, axis=1).reshape(-1, 1)

    X2_sq = np.sum(X2 ** 2, axis=1).reshape(1, -1)

    sq_dist = X1_sq + X2_sq - 2 * X1 @ X2.T

    return np.exp(-gamma * sq_dist)



gamma_value = 0.1

K_rbf = rbf_kernel(X_small, X_small, gamma=gamma_value)



print("\n=== Example Kernel Matrices on 5 Training Samples ===")

print("\nDot-product kernel matrix (linear):\n", K_dot)

print("\nRBF kernel matrix (gamma = 0.1):\n", K_rbf)

import gradio as gr

def predict_cancer(feature1, feature2, feature3, feature4, feature5, feature6, feature7, feature8, feature9, feature10, feature11, feature12, feature13, feature14, feature15, feature16, feature17, feature18, feature19, feature20, feature21, feature22, feature23, feature24, feature25, feature26, feature27, feature28, feature29, feature30):
    # Collect all features into a numpy array
    features = np.array([
        feature1, feature2, feature3, feature4, feature5, feature6, feature7, feature8, feature9, feature10,
        feature11, feature12, feature13, feature14, feature15, feature16, feature17, feature18, feature19, feature20,
        feature21, feature22, feature23, feature24, feature25, feature26, feature27, feature28, feature29, feature30
    ]).reshape(1, -1)

    # Scale the input features using the pre-trained scaler
    features_scaled = scaler.transform(features)

    # Get predictions from both models
    prediction_linear = linear_svm.predict(features_scaled)[0]
    prediction_rbf = rbf_svm.predict(features_scaled)[0]

    # Map predictions to human-readable labels
    label_map = {0: "Malignant", 1: "Benign"}
    output_linear = label_map[prediction_linear]
    output_rbf = label_map[prediction_rbf]

    return output_linear, output_rbf

# Create Gradio input components for 30 features
inputs = [
    gr.Number(label=f"Feature {i+1}") for i in range(30)
]

# Create Gradio output components for both models' predictions
outputs = [
    gr.Textbox(label="Linear SVM Prediction"),
    gr.Textbox(label="RBF SVM Prediction")
]

# Create and launch the Gradio interface
iface = gr.Interface(
    fn=predict_cancer,
    inputs=inputs,
    outputs=outputs,
    title="Breast Cancer Prediction using SVM",
    description="Enter 30 features to predict if a tumor is Malignant (0) or Benign (1) using Linear and RBF Kernel SVMs."
)

iface.launch()