scripts/predict.py - master - Dinky6/Assessment-Question-Difficulty-Classifier

scripts/predict.py
72 строки3 KB
Dinky6
create scripts/predict.py
28 дек 2025, 17:15
28 дек 2025, 17:15fcfc951
#!/usr/bin/env python3
"""
Скрипт для предсказания сложности новых вопросов
"""

import sys
import os
sys.path.append(os.path.join(os.path.dirname(__file__), '..'))

from src.model_training import DifficultyClassifier
from src.feature_extraction import TextFeatureExtractor
import pandas as pd
import numpy as np
import argparse

def predict_single_question(question_text, response_time, correct_rate):
    """Предсказание для одного вопроса"""
    # Загрузка модели
    classifier = DifficultyClassifier.load_model('models/question_classifier.pkl')
    
    # Извлечение признаков
    feature_extractor = classifier.feature_extractor
    text_features = feature_extractor.extract_features([question_text])[0]
    
    # Простые признаки
    simple_features = np.array([response_time, correct_rate, 
                               len(question_text), len(question_text.split()),
                               1 if '?' in question_text else 0])
    
    # Объединение признаков
    features = np.hstack([text_features, simple_features])
    
    # Масштабирование
    features_scaled = classifier.scaler.transform([features])
    
    # Предсказание
    prediction = classifier.predict(features_scaled)[0]
    probabilities = classifier.predict_proba(features_scaled)[0]
    
    # Декодирование
    difficulty = classifier.encoder.inverse_transform([prediction])[0]
    
    return difficulty, probabilities

def main():
    parser = argparse.ArgumentParser(description='Predict question difficulty')
    parser.add_argument('--text', required=True, help='Question text')
    parser.add_argument('--time', type=float, required=True, help='Response time in seconds')
    parser.add_argument('--rate', type=float, required=True, help='Correct answer rate (0-1)')
    
    args = parser.parse_args()
    
    print("Question Difficulty Prediction")
    print("=" * 50)
    print(f"Question: {args.text}")
    print(f"Response time: {args.time} sec")
    print(f"Correct rate: {args.rate:.2%}")
    
    difficulty, probabilities = predict_single_question(
        args.text, args.time, args.rate
    )
    
    print("\nPrediction Results:")
    print(f"Predicted difficulty: {difficulty.upper()}")
    print("\nProbabilities:")
    for cls, prob in zip(['Easy', 'Medium', 'Hard'], probabilities):
        print(f"  {cls}: {prob:.2%}")
    
    return difficulty

if __name__ == '__main__':
    main()