SVM/SVM_C.py

import os
import pickle
import pandas as pd

class SVM_Classifier:
    def __init__(self):
        
        self.weight = 70  # default weight in kg
        self.height = 1.75  # default height in m
        self.gender = "Male"  # default gender
        self.duration = 1.0  # default duration in hours
        self.fat = 25  # default fat percentage
        self.freq = 3  # default workouts per week
        self.experience = 1  # default experience level
        self.workout = "Cardio"  # default workout type

        # Add debug info dictionary
        self.debug_info = {}
        
        # Load the model and required data
        try:
            # Load the SVM model
            model_file = os.path.join("data", "svm_model.pkl")
            with open(model_file, 'rb') as f:
                self.svm_model = pickle.load(f)
            
            # Load the column names
            cols_file = os.path.join("data", "column_names.csv")
            with open(cols_file, 'r') as f:
                self.column_names = [line.strip() for line in f]
            
            # Load normalization parameters
            mean_file = os.path.join("data", "SVM_train_mean.csv")
            self.df_mean = pd.read_csv(mean_file, index_col=0)
            
            std_file = os.path.join("data", "SVM_train_std.csv")
            self.df_std = pd.read_csv(std_file, index_col=0)
            
        except Exception as e:
            print(f"Error loading model files: {str(e)}")
            raise

    def make_prediction(self):
        try:
            num = [self.weight, self.height, self.duration, self.fat, self.freq, self.experience]
            self.debug_info['original_values'] = dict(zip(
                ['weight', 'height', 'duration', 'fat', 'freq', 'experience'], 
                num
            ))

            m_norm = self.df_mean.values.flatten().tolist()
            s_norm = self.df_std.values.flatten().tolist()
            self.debug_info['normalization'] = {
                'means': m_norm,
                'stds': s_norm
            }

            norm = [(x-y)/z for x, y, z in zip(num, m_norm, s_norm)]
            self.debug_info['normalized_values'] = dict(zip(
                ['weight', 'height', 'duration', 'fat', 'freq', 'experience'],
                norm
            ))

            if self.gender == 'Female':
                norm.extend([1,0])
                self.debug_info['gender_encoding'] = 'Female: [1, 0]'
            else:
                norm.extend([0,1])
                self.debug_info['gender_encoding'] = 'Male: [0, 1]'

            # Add one-hot encoded workout type
            workout_encoding = {
                'Cardio': [1, 0, 0, 0],
                'HIIT': [0, 1, 0, 0],
                'Strength': [0, 0, 1, 0],
                'Yoga': [0, 0, 0, 1]
            }
            norm.extend(workout_encoding[self.workout])
            self.debug_info['workout_encoding'] = f'{self.workout}: {workout_encoding[self.workout]}'

            X = pd.DataFrame([norm],columns=self.column_names)
            self.debug_info['final_feature_vector'] = X.to_dict('records')[0]

            prediction = self.svm_model.predict(X)
            self.debug_info['prediction'] = prediction[0]

            return prediction[0]

        except Exception as e:
            self.debug_info['error'] = str(e)
            return f"Error: {str(e)}"

    def get_debug_info(self):
        """Returns formatted debug information"""
        debug_text = "=== DEBUG INFORMATION ===\n\n"
        
        # Original values
        debug_text += "Original Values:\n"
        for key, value in self.debug_info['original_values'].items():
            debug_text += f"{key}: {value}\n"
        
        # Normalization parameters
        debug_text += "\nNormalization Parameters:\n"
        for i, (mean, std) in enumerate(zip(
            self.debug_info['normalization']['means'],
            self.debug_info['normalization']['stds']
        )):
            debug_text += f"Feature {i}: mean={mean:.4f}, std={std:.4f}\n"
        
        # Normalized values
        debug_text += "\nNormalized Values:\n"
        for key, value in self.debug_info['normalized_values'].items():
            debug_text += f"{key}: {value:.4f}\n"
        
        # Encodings
        debug_text += f"\nGender Encoding: {self.debug_info['gender_encoding']}\n"
        debug_text += f"Workout Encoding: {self.debug_info['workout_encoding']}\n"

        # Final vector
        debug_text += f"\nVector: {self.debug_info['final_feature_vector']}\n"
        
        # Final prediction
        debug_text += f"\nFinal Prediction: {self.debug_info['prediction']}\n"
        
        return debug_text