Исправлен .gitignore: пакет stylometry/features добавлен в репозиторий

lab1/stylometry/features/length.py

@@ -0,0 +1,48 @@
+"""Признаки длины и структуры реплики."""
+
+from __future__ import annotations
+
+import pandas as pd
+
+from .base import (
+    as_list,
+    finite,
+    mean,
+    median,
+    safe_div,
+    sentence_word_counts,
+    std,
+    word_lengths,
+    word_tokens,
+)
+
+
+def extract(row: pd.Series) -> dict[str, float]:
+    tokens = as_list(row, "tokens")
+    words = word_tokens(tokens)
+    lengths = word_lengths(words)
+    sentence_lengths = sentence_word_counts(row)
+    n_sentences = max(len(sentence_lengths), 1)
+    text = str(row.get("text", ""))
+
+    return {
+        "len_tokens": float(len(tokens)),
+        "len_words": float(len(words)),
+        "len_chars": float(len(text)),
+        "len_sentences": float(n_sentences),
+        "len_mean_sentence_words": finite(mean(sentence_lengths)),
+        "len_std_sentence_words": finite(std(sentence_lengths)),
+        "len_min_sentence_words": float(min(sentence_lengths)) if sentence_lengths else 0.0,
+        "len_max_sentence_words": float(max(sentence_lengths)) if sentence_lengths else 0.0,
+        "len_median_sentence_words": finite(median(sentence_lengths)),
+        "len_pct_sentences_lt5": safe_div(sum(1 for value in sentence_lengths if value < 5), n_sentences),
+        "len_pct_sentences_gt20": safe_div(sum(1 for value in sentence_lengths if value > 20), n_sentences),
+        "len_pct_one_word_sentences": safe_div(sum(1 for value in sentence_lengths if value == 1), n_sentences),
+        "len_mean_word_chars": finite(mean(lengths)),
+        "len_std_word_chars": finite(std(lengths)),
+        "len_median_word_chars": finite(median(lengths)),
+        "len_pct_words_ge7": safe_div(sum(1 for value in lengths if value >= 7), len(lengths)),
+        "len_pct_words_ge10": safe_div(sum(1 for value in lengths if value >= 10), len(lengths)),
+        "len_pct_words_ge13": safe_div(sum(1 for value in lengths if value >= 13), len(lengths)),
+        "len_pct_single_letter_words": safe_div(sum(1 for value in lengths if value == 1), len(lengths)),
+    }