train_model.py

import os
import cv2
import glob
import numpy as np
import joblib
from sklearn.model_selection import train_test_split
from sklearn.svm import SVC
from sklearn.preprocessing import LabelEncoder, StandardScaler
from sklearn.pipeline import make_pipeline
from sklearn.metrics import classification_report

from landmarks import get_landmarks
from geometry import extract_features

DATASET_DIR = "../dataset"
MODEL_PATH = "face_shape_model.pkl"

def get_feature_vector(features):
    """Converts feature dictionary to a list for the model."""
    return [
        features["lw_ratio"],
        features["jaw_ratio"],
        features["forehead_ratio"],
        # Add more if geometry.py calculates them
    ]

def train():
    print("Starting training...")
    X = []
    y = []
    
    # Iterate through each subdirectory in the dataset
    # Assumes structure: dataset/ShapeName/image.jpg
    if not os.path.exists(DATASET_DIR):
        print(f"Dataset directory not found: {DATASET_DIR}")
        return

    classes = [d for d in os.listdir(DATASET_DIR) if os.path.isdir(os.path.join(DATASET_DIR, d))]
    print(f"Found classes: {classes}")

    for label in classes:
        class_dir = os.path.join(DATASET_DIR, label)
        image_files = glob.glob(os.path.join(class_dir, "*"))
        
        # Normalize label to capitalized format (e.g., "oval" -> "Oval") to match recommendation engine keys
        normalized_label = label.capitalize()
        
        print(f"Processing {label} (normalized to {normalized_label}): {len(image_files)} images")
        
        for img_path in image_files:
            try:
                landmarks = get_landmarks(img_path)
                feats = extract_features(landmarks)
                vector = get_feature_vector(feats)
                
                X.append(vector)
                y.append(normalized_label)
            except Exception as e:
                # print(f"Skipping {img_path}: {e}")
                pass

    if len(X) == 0:
        print("No valid data found. Check dataset and landmarks extraction.")
        return

    X = np.array(X)
    y = np.array(y)

    print(f"Training on {len(X)} samples...")

    # Train/Test split
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

    # Pipeline: Scale features -> SVM Classifier
    model = make_pipeline(StandardScaler(), SVC(kernel='rbf', probability=True))
    model.fit(X_train, y_train)

    # Evaluate
    print("Evaluating model...")
    predictions = model.predict(X_test)
    print(classification_report(y_test, predictions))

    # Save
    joblib.dump(model, MODEL_PATH)
    print(f"Model saved to {MODEL_PATH}")

if __name__ == "__main__":
    train()