evaluation.py

# evaluation.py - дополнительные метрики
import numpy as np
from collections import Counter

def evaluate_text_quality(text):
    """Расширенная оценка качества текста"""
    metrics = {}
    
    # 1. Длина текста
    metrics['length'] = len(text)
    
    # 2. Разнообразие слов
    words = text.lower().split()
    unique_words = set(words)
    metrics['lexical_diversity'] = len(unique_words) / len(words) if words else 0
    
    # 3. Средняя длина предложения
    sentences = text.replace('!', '.').replace('?', '.').split('.')
    sentences = [s.strip() for s in sentences if s.strip()]
    if sentences:
        avg_sentence_len = np.mean([len(s.split()) for s in sentences])
        metrics['avg_sentence_len'] = avg_sentence_len
    else:
        metrics['avg_sentence_len'] = 0
    
    # 4. Повторы (n-граммы)
    def get_ngrams(text, n):
        words = text.lower().split()
        return [' '.join(words[i:i+n]) for i in range(len(words)-n+1)]
    
    bigrams = get_ngrams(text, 2)
    if bigrams:
        bigram_counts = Counter(bigrams)
        most_common = bigram_counts.most_common(1)[0][1] if bigram_counts else 0
        metrics['repetition_score'] = 1 - (most_common / len(bigrams))
    else:
        metrics['repetition_score'] = 0
    
    # Итоговая оценка
    total_score = (
        min(metrics['length'] / 100, 1) * 0.3 +
        metrics['lexical_diversity'] * 0.3 +
        min(metrics['avg_sentence_len'] / 20, 1) * 0.2 +
        metrics['repetition_score'] * 0.2
    ) * 10
    
    metrics['total_score'] = round(total_score, 2)
    return metrics