Update mainapp.py
Browse files- mainapp.py +742 -287
mainapp.py
CHANGED
|
@@ -1,309 +1,764 @@
|
|
| 1 |
import os
|
| 2 |
import re
|
| 3 |
-
import
|
| 4 |
-
import
|
| 5 |
-
from janome.tokenizer import Tokenizer
|
| 6 |
-
import jaconv
|
| 7 |
-
from typing import List, Optional, Dict
|
| 8 |
import logging
|
| 9 |
import json
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
import tkinter as tk
|
| 11 |
-
from tkinter import filedialog, messagebox
|
| 12 |
-
|
| 13 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
|
| 15 |
logger = logging.getLogger(__name__)
|
| 16 |
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
class
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
def
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
remove_symbols
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
|
| 43 |
-
|
| 44 |
-
def extract_hiragana(self, text: str) -> str:
|
| 45 |
-
"""
|
| 46 |
-
テキストデータからひらがなを抽出します。
|
| 47 |
-
|
| 48 |
-
Args:
|
| 49 |
-
text (str): 処理するテキスト。
|
| 50 |
-
|
| 51 |
-
Returns:
|
| 52 |
-
str: ひらがな抽出後のテキスト。
|
| 53 |
-
"""
|
| 54 |
-
hiragana_words: List[str] = []
|
| 55 |
-
for token in self.tokenizer.tokenize(text):
|
| 56 |
-
if token.reading != '*':
|
| 57 |
-
# 読みが存在する場合は、ひらがなに変換
|
| 58 |
-
reading = jaconv.h2z(token.reading, kana=True, ascii=False, digit=False)
|
| 59 |
-
hiragana_words.append(jaconv.kata2hira(reading))
|
| 60 |
-
elif all('\u3041' <= char <= '\u3096' for char in token.surface):
|
| 61 |
-
# ひらがなの場合はそのまま追加
|
| 62 |
-
hiragana_words.append(token.surface)
|
| 63 |
-
elif all('\u4e00' <= char <= '\u9fff' for char in token.surface) and self.convert_kanji:
|
| 64 |
-
# 漢字のみの場合は、変換を試みる
|
| 65 |
-
# (ここに漢字をひらがなに変換するロジックを追加)
|
| 66 |
-
# 今のところ、漢字はそのまま追加
|
| 67 |
-
hiragana_words.append(token.surface)
|
| 68 |
-
elif any('\u3041' <= char <= '\u3096' for char in token.surface) or any('\u4e00' <= char <= '\u9fff' for char in token.surface):
|
| 69 |
-
# ひらがなと漢字が混ざっている場合は、可能な限りひらがな化
|
| 70 |
-
mixed_word = ""
|
| 71 |
-
for char in token.surface:
|
| 72 |
-
if '\u3041' <= char <= '\u3096':
|
| 73 |
-
mixed_word += char
|
| 74 |
-
elif '\u4e00' <= char <= '\u9fff' and self.convert_kanji:
|
| 75 |
-
mixed_word += char
|
| 76 |
-
hiragana_words.append(mixed_word)
|
| 77 |
-
elif token.part_of_speech.startswith('記号') and not self.remove_symbols:
|
| 78 |
-
# 記号を削除しない設定の場合
|
| 79 |
-
hiragana_words.append(token.surface)
|
| 80 |
-
|
| 81 |
-
return "".join(hiragana_words)
|
| 82 |
-
|
| 83 |
-
def preprocess_line(self, text: str, to_lower: bool = False, remove_punctuation: bool = False, remove_numbers: bool = False) -> str:
|
| 84 |
-
"""
|
| 85 |
-
テキストデータの前処理を行います(行ごと処理)。
|
| 86 |
-
|
| 87 |
-
Args:
|
| 88 |
-
text (str): 処理するテキスト。
|
| 89 |
-
to_lower (bool): 小文字に変換するかどうか。デフォルトはFalse。
|
| 90 |
-
remove_punctuation (bool): 句読点を削除するかどうか。デフォルトはFalse。
|
| 91 |
-
remove_numbers (bool): 数字を削除するかどうか。デフォルトはFalse。
|
| 92 |
-
|
| 93 |
-
Returns:
|
| 94 |
-
str: 前処理後のテキスト。
|
| 95 |
-
"""
|
| 96 |
-
text = text.replace('\r\n', '\n').replace('\r', '\n').strip()
|
| 97 |
-
text = re.sub(r'\s+', ' ', text)
|
| 98 |
-
|
| 99 |
-
if to_lower:
|
| 100 |
-
text = text.lower()
|
| 101 |
-
|
| 102 |
-
if remove_punctuation:
|
| 103 |
-
text = re.sub(r'[!"#$%&\'()*+,-./:;<=>?@\[\\\]^_`{|}~]', '', text)
|
| 104 |
-
|
| 105 |
-
if remove_numbers:
|
| 106 |
-
text = re.sub(r'[0-9]', '', text)
|
| 107 |
-
|
| 108 |
-
katakana_words = []
|
| 109 |
-
for token in self.tokenizer.tokenize(text):
|
| 110 |
-
if token.reading == '*':
|
| 111 |
-
katakana_words.append(token.surface)
|
| 112 |
-
else:
|
| 113 |
-
katakana_words.append(token.reading)
|
| 114 |
-
|
| 115 |
-
return "".join(katakana_words)
|
| 116 |
-
|
| 117 |
-
def read_text_with_bom_removal(filepath: str, encoding: str = 'utf-8') -> str:
|
| 118 |
-
"""
|
| 119 |
-
BOM付きの可能性のあるテキストファイルを読み込み、BOMを取り除いて返します。
|
| 120 |
-
|
| 121 |
-
Args:
|
| 122 |
-
filepath (str): ファイルパス。
|
| 123 |
-
encoding (str): ファイルのエンコーディング。デフォルトは'utf-8'。
|
| 124 |
-
|
| 125 |
-
Returns:
|
| 126 |
-
str: BOMを取り除いたテキストデータ。
|
| 127 |
-
|
| 128 |
-
Raises:
|
| 129 |
-
FileProcessingError: ファイルの読み込みに失敗した場合。
|
| 130 |
-
"""
|
| 131 |
-
try:
|
| 132 |
-
with open(filepath, 'rb') as f:
|
| 133 |
-
raw_data = f.read()
|
| 134 |
-
|
| 135 |
-
if raw_data.startswith(b'\xef\xbb\xbf'):
|
| 136 |
-
return raw_data[3:].decode(encoding)
|
| 137 |
-
else:
|
| 138 |
-
return raw_data.decode(encoding)
|
| 139 |
-
except FileNotFoundError:
|
| 140 |
-
raise FileProcessingError(f"ファイル '{filepath}' が見つかりません。")
|
| 141 |
-
except UnicodeDecodeError as e:
|
| 142 |
-
raise FileProcessingError(f"ファイル '{filepath}' のデコードに失敗しました: {e}")
|
| 143 |
-
except Exception as e:
|
| 144 |
-
raise FileProcessingError(f"ファイル '{filepath}' の読み込み中に予期しないエラーが発生しました: {e}")
|
| 145 |
-
|
| 146 |
-
def output_comparison_data(filename: str, original_text: str, preprocessed_text: str, hiragana_text: str, output_folder: str, output_format: str = 'tsv') -> None:
|
| 147 |
-
"""
|
| 148 |
-
元のテキスト、カタカナ変換後のテキスト、ひらがな抽出後のテキストを行ごとに指定されたフォーマットでファイルに出力します。
|
| 149 |
-
TSVとJSONLの両方を出力するように変更。
|
| 150 |
-
"""
|
| 151 |
-
if not os.path.exists(output_folder):
|
| 152 |
-
# ディレクトリがない場合、作成するか確認
|
| 153 |
-
if messagebox.askyesno("ディレクトリ作成", f"出力ディレクトリ '{output_folder}' が存在しません。作成しますか?"):
|
| 154 |
try:
|
| 155 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 156 |
except Exception as e:
|
| 157 |
-
|
|
|
|
| 158 |
else:
|
| 159 |
-
|
| 160 |
-
|
| 161 |
-
|
| 162 |
-
|
| 163 |
-
|
| 164 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 165 |
try:
|
| 166 |
-
with open(
|
| 167 |
-
|
| 168 |
-
|
| 169 |
-
|
| 170 |
-
|
| 171 |
-
|
| 172 |
-
|
| 173 |
-
|
| 174 |
-
|
| 175 |
-
|
| 176 |
-
|
| 177 |
-
|
| 178 |
-
|
| 179 |
-
|
| 180 |
-
|
| 181 |
-
|
| 182 |
-
|
| 183 |
-
|
| 184 |
-
|
| 185 |
-
|
| 186 |
-
|
| 187 |
-
|
| 188 |
-
|
| 189 |
-
|
| 190 |
-
|
| 191 |
-
|
| 192 |
-
|
| 193 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 194 |
except Exception as e:
|
| 195 |
-
|
| 196 |
-
|
| 197 |
-
def process_file(filepath: str, output_folder: str, text_processor: TextProcessor, output_format: str) -> None:
|
| 198 |
-
"""
|
| 199 |
-
指定されたテキストファイルを読み込み、行ごとに前処理(カタカナ変換)とひらがな抽出を行い、比較データを出力します。
|
| 200 |
|
| 201 |
-
Args:
|
| 202 |
-
filepath (str): 処理するファイルのパス。
|
| 203 |
-
output_folder (str): 出力フォルダのパス。
|
| 204 |
-
text_processor (TextProcessor): テキスト処理を行うTextProcessorオブジェクト。
|
| 205 |
-
output_format (str): 出力フォーマット('tsv'または'csv')。
|
| 206 |
|
| 207 |
-
|
| 208 |
-
|
| 209 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 210 |
filename = os.path.basename(filepath)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 211 |
try:
|
| 212 |
-
|
| 213 |
-
|
| 214 |
-
preprocessed_lines: List[str] = []
|
| 215 |
-
hiragana_lines: List[str] = []
|
| 216 |
-
|
| 217 |
-
for line in original_lines:
|
| 218 |
-
if line.strip():
|
| 219 |
-
preprocessed_line = text_processor.preprocess_line(line)
|
| 220 |
-
hiragana_line = text_processor.extract_hiragana(line)
|
| 221 |
-
preprocessed_lines.append(preprocessed_line)
|
| 222 |
-
hiragana_lines.append(hiragana_line)
|
| 223 |
-
else:
|
| 224 |
-
preprocessed_lines.append("") # 空行を保持
|
| 225 |
-
hiragana_lines.append("") # 空行を保持
|
| 226 |
|
| 227 |
-
|
|
|
|
|
|
|
|
|
|
| 228 |
|
| 229 |
-
|
| 230 |
-
|
| 231 |
-
|
| 232 |
-
|
| 233 |
-
|
| 234 |
-
def select_files():
|
| 235 |
-
"""ファイルを選択するダイアログを表示します。"""
|
| 236 |
-
root = tk.Tk()
|
| 237 |
-
root.withdraw() # メインウィンドウを非表示
|
| 238 |
-
file_paths = filedialog.askopenfilenames(title="処理するテキストファイルを選択してください", filetypes=[("Text files", "*.txt")])
|
| 239 |
-
return list(file_paths)
|
| 240 |
-
|
| 241 |
-
def select_output_folder():
|
| 242 |
-
"""出力フォルダを選択するダイアログを表示します。"""
|
| 243 |
-
root = tk.Tk()
|
| 244 |
-
root.withdraw() # メインウィンドウを非表示
|
| 245 |
-
folder_path = filedialog.askdirectory(title="出���フォルダを選択してください")
|
| 246 |
-
return folder_path
|
| 247 |
-
|
| 248 |
-
def main():
|
| 249 |
-
"""
|
| 250 |
-
メイン関数。テキストファイルの前処理とひらがな抽出を行います。
|
| 251 |
-
"""
|
| 252 |
-
parser = argparse.ArgumentParser(description='テキストファイルの前処理とひらがな抽出を行います。')
|
| 253 |
-
parser.add_argument('--config', type=str, help='設定ファイルのパス')
|
| 254 |
-
parser.add_argument('--udic_path', type=str, help='udicのパス')
|
| 255 |
-
parser.add_argument('--file_extension', type=str, default='.txt', help='処理するファイルの拡張子')
|
| 256 |
-
parser.add_argument('--output_format', type=str, default='tsv', choices=['tsv', 'csv'], help='出力フォーマット(tsvまたはcsv)')
|
| 257 |
-
parser.add_argument('--remove_symbols', action='store_true', help='記号を削除する')
|
| 258 |
-
parser.add_argument('--convert_kanji', action='store_true', help='漢字をひらがなに変換する')
|
| 259 |
-
args = parser.parse_args()
|
| 260 |
-
|
| 261 |
-
config = configparser.ConfigParser()
|
| 262 |
-
if args.config and os.path.exists(args.config):
|
| 263 |
-
config.read(args.config)
|
| 264 |
-
output_folder = config.get('Paths', 'output_folder', fallback=None)
|
| 265 |
-
udic_path = config.get('Paths', 'udic_path', fallback=args.udic_path)
|
| 266 |
-
file_extension = config.get('Settings', 'file_extension', fallback=args.file_extension)
|
| 267 |
-
output_format = config.get('Settings', 'output_format', fallback=args.output_format)
|
| 268 |
-
remove_symbols = config.getboolean('Settings', 'remove_symbols', fallback=not args.remove_symbols)
|
| 269 |
-
convert_kanji = config.getboolean('Settings', 'convert_kanji', fallback=args.convert_kanji)
|
| 270 |
-
else:
|
| 271 |
-
output_folder = None
|
| 272 |
-
udic_path = args.udic_path
|
| 273 |
-
file_extension = args.file_extension
|
| 274 |
-
output_format = args.output_format
|
| 275 |
-
remove_symbols = not args.remove_symbols
|
| 276 |
-
convert_kanji = args.convert_kanji
|
| 277 |
-
|
| 278 |
-
# ファイル選択ダイアログを表示
|
| 279 |
-
file_paths = select_files()
|
| 280 |
-
if not file_paths:
|
| 281 |
-
logger.info("ファイルが選択されなかったため、処理を中止します。")
|
| 282 |
-
return
|
| 283 |
-
|
| 284 |
-
# 出力フォルダ選択ダイアログを表示
|
| 285 |
-
if output_folder is None:
|
| 286 |
-
output_folder = select_output_folder()
|
| 287 |
-
if not output_folder:
|
| 288 |
-
logger.info("出力フォルダが選択されなかったため、処理を中止します。")
|
| 289 |
-
return
|
| 290 |
-
|
| 291 |
-
# パスの正規化
|
| 292 |
-
output_folder = os.path.normpath(output_folder)
|
| 293 |
-
|
| 294 |
-
text_processor = TextProcessor(udic_path, remove_symbols, convert_kanji)
|
| 295 |
|
| 296 |
-
try:
|
| 297 |
-
total_files = len(file_paths)
|
| 298 |
-
for i, filepath in enumerate(file_paths):
|
| 299 |
-
logger.info(f"処理中のファイル ({i+1}/{total_files}): {filepath}")
|
| 300 |
try:
|
| 301 |
-
|
| 302 |
-
|
| 303 |
-
|
| 304 |
-
|
| 305 |
-
|
| 306 |
-
|
| 307 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 308 |
if __name__ == "__main__":
|
| 309 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
import os
|
| 2 |
import re
|
| 3 |
+
import unicodedata
|
| 4 |
+
from typing import List, Optional, Dict, Any, Tuple, Literal
|
|
|
|
|
|
|
|
|
|
| 5 |
import logging
|
| 6 |
import json
|
| 7 |
+
import sys
|
| 8 |
+
import csv
|
| 9 |
+
import yaml # PyYAMLが必要: pip install PyYAML
|
| 10 |
+
import jaconv # jaconvが必要: pip install jaconv
|
| 11 |
+
import threading
|
| 12 |
import tkinter as tk
|
| 13 |
+
from tkinter import ttk, filedialog, messagebox, scrolledtext
|
| 14 |
+
|
| 15 |
+
# --- 依存ライブラリのチェックとインポート ---
|
| 16 |
+
try:
|
| 17 |
+
from janome.tokenizer import Tokenizer, Token
|
| 18 |
+
from janome.analyzer import Analyzer
|
| 19 |
+
from janome.charfilter import UnicodeNormalizeCharFilter
|
| 20 |
+
# from janome.tokenfilter import POSKeepFilter, CompoundNounFilter # 必要なら使う
|
| 21 |
+
JANOME_AVAILABLE = True
|
| 22 |
+
except ImportError:
|
| 23 |
+
JANOME_AVAILABLE = False
|
| 24 |
+
|
| 25 |
+
try:
|
| 26 |
+
import pyopenjtalk
|
| 27 |
+
PYOPENJTALK_AVAILABLE = True
|
| 28 |
+
except ImportError:
|
| 29 |
+
PYOPENJTALK_AVAILABLE = False
|
| 30 |
+
except Exception as e: # pyopenjtalkはImportError以外も発生しうる
|
| 31 |
+
PYOPENJTALK_AVAILABLE = False
|
| 32 |
+
print(f"警告: pyopenjtalkのインポート中にエラーが発生しました: {e}", file=sys.stderr)
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
# --- GUI機能が利用可能か ---
|
| 36 |
+
# ENABLE_GUI = True # tkinterのimport成功時にTrueになる想定
|
| 37 |
+
try:
|
| 38 |
+
# tkinterは標準ライブラリだが、最小環境でない場合もある
|
| 39 |
+
import tkinter as tk
|
| 40 |
+
from tkinter import ttk, filedialog, messagebox, scrolledtext
|
| 41 |
+
ENABLE_GUI = True
|
| 42 |
+
except ImportError:
|
| 43 |
+
ENABLE_GUI = False
|
| 44 |
+
|
| 45 |
+
# ロギングの設定 (GUI表示用にレベルはINFO推奨)
|
| 46 |
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
|
| 47 |
logger = logging.getLogger(__name__)
|
| 48 |
|
| 49 |
+
# --- 例外クラス ---
|
| 50 |
+
class TextProcessingError(Exception): pass
|
| 51 |
+
class FileProcessingError(TextProcessingError): pass
|
| 52 |
+
class InitializationError(TextProcessingError): pass
|
| 53 |
+
class NormalizationError(TextProcessingError): pass
|
| 54 |
+
|
| 55 |
+
# --- テキスト正規化クラス ---
|
| 56 |
+
class TextNormalizer:
|
| 57 |
+
def __init__(self, rule_filepath: Optional[str] = None):
|
| 58 |
+
self.rules = self._load_rules(rule_filepath)
|
| 59 |
+
logger.info(f"正規化ルールをロードしました (ファイル: {rule_filepath or 'デフォルト'})")
|
| 60 |
+
self.number_converter = self._create_number_converter()
|
| 61 |
+
|
| 62 |
+
def _load_rules(self, filepath: Optional[str]) -> Dict[str, Any]:
|
| 63 |
+
# デフォルトルール (必要最低限)
|
| 64 |
+
default_rules = {
|
| 65 |
+
'unicode_normalize': 'NFKC',
|
| 66 |
+
'remove_whitespace': True,
|
| 67 |
+
'replace_symbols': {'&': ' アンド ', '%': ' パーセント '},
|
| 68 |
+
'remove_symbols': r'[「」『』【】[]()<>‘’“”・※→←↑↓*#〜]',
|
| 69 |
+
'punctuation': {'remove': False, 'replacement': ' <pau> ', 'target': '、。?!'},
|
| 70 |
+
'number_conversion': {'enabled': True, 'target': r'\d+([\.,]\d+)?'},
|
| 71 |
+
'alphabet_conversion': {'enabled': True, 'rule': 'spellout', 'dictionary': {'AI': 'エーアイ', 'USB': 'ユーエスビー'}},
|
| 72 |
+
'custom_replacements_pre': [],
|
| 73 |
+
'custom_replacements_post': [],
|
| 74 |
+
}
|
| 75 |
+
if filepath and os.path.exists(filepath):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 76 |
try:
|
| 77 |
+
with open(filepath, 'r', encoding='utf-8') as f:
|
| 78 |
+
loaded_rules = yaml.safe_load(f)
|
| 79 |
+
# より安全なマージ (ネストした辞書も考慮するなら再帰的なマージが必要)
|
| 80 |
+
for key, value in loaded_rules.items():
|
| 81 |
+
if isinstance(value, dict) and isinstance(default_rules.get(key), dict):
|
| 82 |
+
default_rules[key].update(value)
|
| 83 |
+
else:
|
| 84 |
+
default_rules[key] = value
|
| 85 |
+
logger.info(f"カスタム正規化ルール '{filepath}' をロードしました。")
|
| 86 |
+
return default_rules
|
| 87 |
except Exception as e:
|
| 88 |
+
logger.warning(f"正規化ルールファイル '{filepath}' の読み込みに失敗: {e}。デフォルトルールを使用。")
|
| 89 |
+
return default_rules
|
| 90 |
else:
|
| 91 |
+
if filepath: logger.warning(f"正規化ルールファイル '{filepath}' が見つかりません。デフォルトルールを使用。")
|
| 92 |
+
else: logger.info("正規化ルールファイル未指定。デフォルトルールを使用。")
|
| 93 |
+
return default_rules
|
| 94 |
+
|
| 95 |
+
def _create_number_converter(self):
|
| 96 |
+
# 数字 -> 日本語読み 変換クラス (要実装)
|
| 97 |
+
class SimpleNumConverter:
|
| 98 |
+
def convert(self, num_str: str) -> str:
|
| 99 |
+
# ここに詳細な変換ロジックを実装する
|
| 100 |
+
# 例: 123 -> ひゃくにじゅうさん, 10.5 -> じゅってんご
|
| 101 |
+
return num_str # プレースホルダー
|
| 102 |
+
return SimpleNumConverter()
|
| 103 |
+
|
| 104 |
+
def normalize(self, text: str) -> str:
|
| 105 |
+
if not isinstance(text, str): return ""
|
| 106 |
+
try:
|
| 107 |
+
# 1. Unicode正規化
|
| 108 |
+
norm_type = self.rules.get('unicode_normalize')
|
| 109 |
+
if norm_type: text = unicodedata.normalize(norm_type, text)
|
| 110 |
+
|
| 111 |
+
# 2. 前処理カスタム置換
|
| 112 |
+
for rule in self.rules.get('custom_replacements_pre', []):
|
| 113 |
+
text = re.sub(rule['pattern'], rule['replacement'], text)
|
| 114 |
+
|
| 115 |
+
# 3. 空白正規化 (早期に行う方が後続処理が楽な場合も)
|
| 116 |
+
if self.rules.get('remove_whitespace', True):
|
| 117 |
+
text = re.sub(r'\s+', ' ', text).strip()
|
| 118 |
+
|
| 119 |
+
# 4. 数字変換 (ルールの target に基づいて検索・置換)
|
| 120 |
+
num_conf = self.rules.get('number_conversion', {})
|
| 121 |
+
if num_conf.get('enabled'):
|
| 122 |
+
pattern = num_conf.get('target', r'\d+([\.,]\d+)?')
|
| 123 |
+
try:
|
| 124 |
+
text = re.sub(pattern, lambda m: self.number_converter.convert(m.group(0)), text)
|
| 125 |
+
except Exception as e_num: logger.warning(f"数字変換エラー: {e_num}")
|
| 126 |
+
|
| 127 |
+
# 5. アルファベット変換 (辞書 + ルール)
|
| 128 |
+
alpha_conf = self.rules.get('alphabet_conversion', {})
|
| 129 |
+
if alpha_conf.get('enabled'):
|
| 130 |
+
for word, reading in alpha_conf.get('dictionary', {}).items():
|
| 131 |
+
# 大文字小文字区別など考慮
|
| 132 |
+
text = text.replace(word, f' {reading} ')
|
| 133 |
+
if alpha_conf.get('rule') == 'spellout':
|
| 134 |
+
# ここにアルファベットを1文字ずつ読む処理 (A->エー) を実装
|
| 135 |
+
pass
|
| 136 |
+
|
| 137 |
+
# 6. 記号読み置換
|
| 138 |
+
for symbol, reading in self.rules.get('replace_symbols', {}).items():
|
| 139 |
+
text = text.replace(symbol, reading)
|
| 140 |
+
|
| 141 |
+
# 7. 記号削除
|
| 142 |
+
remove_pattern = self.rules.get('remove_symbols')
|
| 143 |
+
if remove_pattern:
|
| 144 |
+
text = re.sub(remove_pattern, '', text)
|
| 145 |
+
|
| 146 |
+
# 8. 句読点処理
|
| 147 |
+
punct_conf = self.rules.get('punctuation', {})
|
| 148 |
+
target_punct = punct_conf.get('target', '、。?!')
|
| 149 |
+
punct_pattern = f'[{re.escape(target_punct)}]'
|
| 150 |
+
if punct_conf.get('remove', True):
|
| 151 |
+
text = re.sub(punct_pattern, '', text)
|
| 152 |
+
else:
|
| 153 |
+
replacement = punct_conf.get('replacement', '<pau>')
|
| 154 |
+
text = re.sub(punct_pattern, replacement, text)
|
| 155 |
+
|
| 156 |
+
# 9. 後処理カスタム置換
|
| 157 |
+
for rule in self.rules.get('custom_replacements_post', []):
|
| 158 |
+
text = re.sub(rule['pattern'], rule['replacement'], text)
|
| 159 |
+
|
| 160 |
+
# 10. 最終空白処理
|
| 161 |
+
if self.rules.get('remove_whitespace', True):
|
| 162 |
+
text = re.sub(r'\s+', ' ', text).strip()
|
| 163 |
+
|
| 164 |
+
return text
|
| 165 |
+
except Exception as e:
|
| 166 |
+
raise NormalizationError(f"正規化失敗 ({text[:20]}...): {e}")
|
| 167 |
+
|
| 168 |
+
|
| 169 |
+
# --- 読み・アクセント推定クラス ---
|
| 170 |
+
class PronunciationEstimator:
|
| 171 |
+
def __init__(
|
| 172 |
+
self,
|
| 173 |
+
engine: Literal['janome', 'pyopenjtalk'] = 'pyopenjtalk',
|
| 174 |
+
janome_udic_path: Optional[str] = None,
|
| 175 |
+
jtalk_dic_path: Optional[str] = None,
|
| 176 |
+
jtalk_user_dic_path: Optional[str] = None
|
| 177 |
+
):
|
| 178 |
+
self.engine = engine
|
| 179 |
+
self.jtalk_user_dic_path = jtalk_user_dic_path # ユーザー辞書パスを保持
|
| 180 |
+
self.unknown_words = set()
|
| 181 |
+
self.janome_analyzer = None # Janomeインスタンス用
|
| 182 |
+
self.jtalk_opts = [] # pyopenjtalkオプション用
|
| 183 |
+
|
| 184 |
+
if self.engine == 'janome':
|
| 185 |
+
if not JANOME_AVAILABLE: raise InitializationError("Janome利用不可")
|
| 186 |
+
try:
|
| 187 |
+
logger.info(f"Janome初期化 (ユーザー辞書: {janome_udic_path})")
|
| 188 |
+
char_filters = [UnicodeNormalizeCharFilter()]
|
| 189 |
+
# ユーザー辞書はTokenizerに渡す
|
| 190 |
+
tokenizer_kwargs = {"udic": janome_udic_path, "udic_enc": "utf8"} if janome_udic_path else {}
|
| 191 |
+
self.janome_analyzer = Analyzer(
|
| 192 |
+
char_filters=char_filters,
|
| 193 |
+
tokenizer=Tokenizer(**tokenizer_kwargs),
|
| 194 |
+
token_filters=[] # 必要なら追加
|
| 195 |
+
)
|
| 196 |
+
logger.info("Janome初期化完了")
|
| 197 |
+
except Exception as e: raise InitializationError(f"Janome初期化失敗: {e}")
|
| 198 |
+
|
| 199 |
+
elif self.engine == 'pyopenjtalk':
|
| 200 |
+
if not PYOPENJTALK_AVAILABLE: raise InitializationError("pyopenjtalk利用不可")
|
| 201 |
+
try:
|
| 202 |
+
logger.info("pyopenjtalk初期化...")
|
| 203 |
+
if jtalk_dic_path:
|
| 204 |
+
logger.info(f"jtalkシステム辞書パス設定: {jtalk_dic_path}")
|
| 205 |
+
os.environ['PYOPENJTALK_DICT_PATH'] = jtalk_dic_path
|
| 206 |
+
# pyopenjtalkのユーザー辞書指定はrun_frontendでは直接できないため注意喚起
|
| 207 |
+
if self.jtalk_user_dic_path:
|
| 208 |
+
logger.warning(f"jtalkユーザー辞書 '{self.jtalk_user_dic_path}' はrun_frontendでは直接使用されません。システム辞書への統合等を検討してください。g2pには影響する可能性があります。")
|
| 209 |
+
# g2p に渡すオプションなど、より詳細な制御が必要なら実装
|
| 210 |
+
|
| 211 |
+
_ = pyopenjtalk.g2p('テスト') # 動作確認
|
| 212 |
+
logger.info("pyopenjtalk初期化完了")
|
| 213 |
+
except Exception as e: raise InitializationError(f"pyopenjtalk初期化失敗: {e}")
|
| 214 |
+
else:
|
| 215 |
+
raise InitializationError(f"無効なエンジン: {self.engine}")
|
| 216 |
+
|
| 217 |
+
def _get_janome_reading(self, text: str) -> Tuple[str, List[Dict[str, Any]]]:
|
| 218 |
+
tokens_details = []
|
| 219 |
+
katakana_parts = []
|
| 220 |
+
try:
|
| 221 |
+
tokens = self.janome_analyzer.analyze(text)
|
| 222 |
+
for token in tokens:
|
| 223 |
+
surf = token.surface
|
| 224 |
+
pos = token.part_of_speech # 詳細な品詞
|
| 225 |
+
reading = token.reading if token.reading != '*' else None
|
| 226 |
+
pron = token.pronunciation if hasattr(token, 'pronunciation') and token.pronunciation != '*' else reading
|
| 227 |
+
|
| 228 |
+
reading_used = reading or pron # 読み or 発音
|
| 229 |
+
if reading_used:
|
| 230 |
+
katakana_parts.append(reading_used)
|
| 231 |
+
else:
|
| 232 |
+
katakana_parts.append(jaconv.hira2kata(surf)) # 表層形をカタカナ化
|
| 233 |
+
if not re.fullmatch(r'[ぁ-んァ-ンー\s]+', surf) and '記号' not in pos:
|
| 234 |
+
self.unknown_words.add(surf)
|
| 235 |
+
|
| 236 |
+
tokens_details.append({
|
| 237 |
+
"surface": surf, "pos": pos, "reading": reading, "pronunciation": pron
|
| 238 |
+
})
|
| 239 |
+
return "".join(katakana_parts), tokens_details
|
| 240 |
+
except Exception as e:
|
| 241 |
+
logger.error(f"Janome読み推定エラー: {e}", exc_info=True)
|
| 242 |
+
return "[READING ERROR]", []
|
| 243 |
+
|
| 244 |
+
def _get_jtalk_pronunciation(self, text: str) -> Tuple[str, List[Dict[str, Any]], List[str]]:
|
| 245 |
+
katakana_reading = "[READING ERROR]"
|
| 246 |
+
tokens_details = []
|
| 247 |
+
full_context_labels = []
|
| 248 |
+
try:
|
| 249 |
+
# 読みはg2pで取得するのが確実
|
| 250 |
+
# ユーザー辞書はg2pには影響するはず (-u オプション?) 要調査
|
| 251 |
+
g2p_opts = ['-u', self.jtalk_user_dic_path] if self.jtalk_user_dic_path and os.path.exists(self.jtalk_user_dic_path) else []
|
| 252 |
+
# katakana_reading = pyopenjtalk.g2p(text, kana=True, opts=g2p_opts) # opts引数は存在しない?
|
| 253 |
+
katakana_reading = pyopenjtalk.g2p(text, kana=True) # 現状 opts なしで実行
|
| 254 |
+
|
| 255 |
+
# LABデータはrun_frontendで取得
|
| 256 |
+
# run_frontendにはユーザー辞書オプションがない
|
| 257 |
+
full_context_labels = pyopenjtalk.run_frontend(text, self.jtalk_opts)
|
| 258 |
+
|
| 259 |
+
# 簡易的なトークン情報(本来はMeCabの結果をパースすべき)
|
| 260 |
+
tokens_details = [{"surface": text, "reading": katakana_reading}]
|
| 261 |
+
|
| 262 |
+
# 未知語判定(読みが表層形と同じでカタカナのみでない場合)
|
| 263 |
+
if katakana_reading == text and not re.fullmatch(r'[ァ-ンヴー\s]+', text):
|
| 264 |
+
self.unknown_words.add(text)
|
| 265 |
+
|
| 266 |
+
return katakana_reading, tokens_details, full_context_labels
|
| 267 |
+
except Exception as e:
|
| 268 |
+
logger.error(f"pyopenjtalk読み推定エラー: {e}", exc_info=True)
|
| 269 |
+
return "[READING ERROR]", [], []
|
| 270 |
+
|
| 271 |
+
def get_pronunciation(self, text: str, output_format: Literal['hiragana', 'katakana'] = 'katakana') -> Tuple[str, List[Dict[str, Any]], List[str]]:
|
| 272 |
+
reading = "[READING ERROR]"
|
| 273 |
+
tokens_details = []
|
| 274 |
+
lab_data = []
|
| 275 |
+
|
| 276 |
+
if not text: return "", [], []
|
| 277 |
+
|
| 278 |
+
if self.engine == 'janome':
|
| 279 |
+
reading, tokens_details = self._get_janome_reading(text)
|
| 280 |
+
elif self.engine == 'pyopenjtalk':
|
| 281 |
+
reading, tokens_details, lab_data = self._get_jtalk_pronunciation(text)
|
| 282 |
+
|
| 283 |
+
# 出力形式変換
|
| 284 |
+
if output_format == 'hiragana' and "[ERROR]" not in reading:
|
| 285 |
+
try: reading = jaconv.kata2hira(reading)
|
| 286 |
+
except Exception: logger.warning("ひらがな変換失敗", exc_info=True)
|
| 287 |
+
|
| 288 |
+
return reading, tokens_details, lab_data
|
| 289 |
+
|
| 290 |
+
|
| 291 |
+
# --- ファイルIO & 出力関数 ---
|
| 292 |
+
def read_text_with_bom_removal(filepath: str, encoding: str = 'utf-8') -> List[str]:
|
| 293 |
try:
|
| 294 |
+
with open(filepath, 'rb') as f: raw_data = f.read()
|
| 295 |
+
if raw_data.startswith(b'\xef\xbb\xbf'): text = raw_data[3:].decode(encoding, errors='replace')
|
| 296 |
+
elif raw_data.startswith((b'\xff\xfe', b'\xfe\xff')): text = raw_data.decode('utf-16', errors='replace')
|
| 297 |
+
else: text = raw_data.decode(encoding, errors='replace')
|
| 298 |
+
return text.splitlines()
|
| 299 |
+
except FileNotFoundError: raise FileProcessingError(f"ファイルが見つかりません: {filepath}")
|
| 300 |
+
except Exception as e: raise FileProcessingError(f"ファイル読み込みエラー ({filepath}): {e}")
|
| 301 |
+
|
| 302 |
+
def output_corpus_data(output_filepath_base: str, data: List[Dict[str, Any]], output_formats: List[Literal['tsv', 'jsonl']]) -> None:
|
| 303 |
+
output_successful = []
|
| 304 |
+
if not data: logger.warning(f"{output_filepath_base}*: 出力データなし"); return
|
| 305 |
+
|
| 306 |
+
# TSV Output
|
| 307 |
+
if 'tsv' in output_formats:
|
| 308 |
+
fpath = f"{output_filepath_base}.tsv"
|
| 309 |
+
try:
|
| 310 |
+
header = ["id", "text", "reading"] # 基本
|
| 311 |
+
# 必要なら他のキーも追加
|
| 312 |
+
with open(fpath, 'w', encoding='utf-8', newline='') as f:
|
| 313 |
+
writer = csv.DictWriter(f, fieldnames=header, delimiter='\t', quoting=csv.QUOTE_MINIMAL, extrasaction='ignore')
|
| 314 |
+
writer.writeheader()
|
| 315 |
+
writer.writerows(data)
|
| 316 |
+
logger.info(f"TSV出力完了: {fpath}")
|
| 317 |
+
output_successful.append('tsv')
|
| 318 |
+
except Exception as e: logger.error(f"TSV出力失敗 ({fpath}): {e}")
|
| 319 |
+
|
| 320 |
+
# JSONL Output
|
| 321 |
+
if 'jsonl' in output_formats:
|
| 322 |
+
fpath = f"{output_filepath_base}.jsonl"
|
| 323 |
+
try:
|
| 324 |
+
with open(fpath, 'w', encoding='utf-8') as f:
|
| 325 |
+
for item in data:
|
| 326 |
+
item_copy = {k: v for k, v in item.items() if k != 'lab_data'} # LABは除く
|
| 327 |
+
f.write(json.dumps(item_copy, ensure_ascii=False) + '\n')
|
| 328 |
+
logger.info(f"JSONL出力完了: {fpath}")
|
| 329 |
+
output_successful.append('jsonl')
|
| 330 |
+
except Exception as e: logger.error(f"JSONL出力失敗 ({fpath}): {e}")
|
| 331 |
+
|
| 332 |
+
if not output_successful and output_formats:
|
| 333 |
+
raise FileProcessingError("指定形式で���出力に成功せず")
|
| 334 |
+
|
| 335 |
+
def output_lab_data(output_filepath_base: str, all_lab_data: Dict[str, List[str]]) -> None:
|
| 336 |
+
lab_dir = f"{output_filepath_base}_lab"
|
| 337 |
+
try:
|
| 338 |
+
os.makedirs(lab_dir, exist_ok=True)
|
| 339 |
+
count = 0
|
| 340 |
+
for file_id, lab_lines in all_lab_data.items():
|
| 341 |
+
if lab_lines:
|
| 342 |
+
fpath = os.path.join(lab_dir, f"{file_id}.lab")
|
| 343 |
+
with open(fpath, 'w', encoding='utf-8') as f:
|
| 344 |
+
f.write("\n".join(lab_lines))
|
| 345 |
+
count += 1
|
| 346 |
+
logger.info(f"LABデータ出力完了: {count}ファイル -> {lab_dir}")
|
| 347 |
+
except Exception as e: logger.error(f"LABデータ出力失敗: {e}")
|
| 348 |
+
|
| 349 |
+
def output_comparison_data(output_filepath_base: str, comparison_data: List[Dict[str, str]]) -> None:
|
| 350 |
+
"""オプション: 元テキスト、正規化テキスト、読みの比較データをTSVで出力"""
|
| 351 |
+
fpath = f"{output_filepath_base}_comparison.tsv"
|
| 352 |
+
try:
|
| 353 |
+
if not comparison_data: return
|
| 354 |
+
header = ["id", "original", "normalized", "reading"]
|
| 355 |
+
with open(fpath, 'w', encoding='utf-8', errors='replace', newline='') as tsvfile:
|
| 356 |
+
writer = csv.DictWriter(tsvfile, fieldnames=header, delimiter='\t', quoting=csv.QUOTE_MINIMAL)
|
| 357 |
+
writer.writeheader()
|
| 358 |
+
writer.writerows(comparison_data)
|
| 359 |
+
logger.info(f"比較データ出力完了: {fpath}")
|
| 360 |
except Exception as e:
|
| 361 |
+
logger.error(f"比較データ出力失敗 ({fpath}): {e}")
|
|
|
|
|
|
|
|
|
|
|
|
|
| 362 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 363 |
|
| 364 |
+
# --- ファイル処理関数 ---
|
| 365 |
+
def process_file(
|
| 366 |
+
filepath: str, output_folder: str, normalizer: TextNormalizer, pron_estimator: PronunciationEstimator,
|
| 367 |
+
output_formats: List[Literal['tsv', 'jsonl']], lab_output_enabled: bool, reading_output_format: Literal['hiragana', 'katakana'],
|
| 368 |
+
encoding: str, output_comparison: bool = False,
|
| 369 |
+
) -> Tuple[int, int]:
|
| 370 |
+
"""ファイルを処理し、コーパスデータを出力"""
|
| 371 |
filename = os.path.basename(filepath)
|
| 372 |
+
output_base = os.path.join(output_folder, os.path.splitext(filename)[0])
|
| 373 |
+
processed_data: List[Dict[str, Any]] = []
|
| 374 |
+
comparison_items: List[Dict[str, str]] = []
|
| 375 |
+
all_lab_data: Dict[str, List[str]] = {}
|
| 376 |
+
success_count = 0
|
| 377 |
+
error_count = 0
|
| 378 |
+
|
| 379 |
try:
|
| 380 |
+
original_lines = read_text_with_bom_removal(filepath, encoding)
|
| 381 |
+
logger.debug(f"'{filename}'読み込み完了、{len(original_lines)}行")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 382 |
|
| 383 |
+
for i, line in enumerate(original_lines):
|
| 384 |
+
line_id = f"{os.path.splitext(filename)[0]}_{i:04d}"
|
| 385 |
+
original_line = line.strip()
|
| 386 |
+
if not original_line: continue
|
| 387 |
|
| 388 |
+
normalized_line = "[NORM ERROR]"
|
| 389 |
+
reading = "[READING ERROR]"
|
| 390 |
+
lab_data = []
|
| 391 |
+
line_success = False
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 392 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 393 |
try:
|
| 394 |
+
normalized_line = normalizer.normalize(original_line)
|
| 395 |
+
if not normalized_line: logger.warning(f"{line_id}: 正規化後空"); continue
|
| 396 |
+
|
| 397 |
+
reading, _, lab_data = pron_estimator.get_pronunciation(normalized_line, reading_output_format)
|
| 398 |
+
if "[ERROR]" not in reading: line_success = True
|
| 399 |
+
|
| 400 |
+
except NormalizationError as e_norm: logger.error(f"{line_id}: 正規化エラー: {e_norm}")
|
| 401 |
+
except Exception as e_line: logger.error(f"{line_id}: 行処理エラー: {e_line}", exc_info=True)
|
| 402 |
+
|
| 403 |
+
# 結果格納
|
| 404 |
+
if line_success: success_count += 1
|
| 405 |
+
else: error_count += 1
|
| 406 |
+
|
| 407 |
+
corpus_item = {"id": line_id, "text": normalized_line, "reading": reading}
|
| 408 |
+
processed_data.append(corpus_item)
|
| 409 |
+
if lab_output_enabled and lab_data: all_lab_data[line_id] = lab_data
|
| 410 |
+
if output_comparison:
|
| 411 |
+
comparison_items.append({"id": line_id, "original": original_line, "normalized": normalized_line, "reading": reading})
|
| 412 |
+
|
| 413 |
+
# ファイル単位出力
|
| 414 |
+
if processed_data: output_corpus_data(output_base, processed_data, output_formats)
|
| 415 |
+
if lab_output_enabled and all_lab_data: output_lab_data(output_base, all_lab_data)
|
| 416 |
+
if output_comparison and comparison_items: output_comparison_data(output_base, comparison_items)
|
| 417 |
+
|
| 418 |
+
logger.debug(f"'{filename}'処理完了 S:{success_count}, E:{error_count}")
|
| 419 |
+
return success_count, error_count
|
| 420 |
+
|
| 421 |
+
except FileProcessingError as e_fp: raise e_fp # 再throw
|
| 422 |
+
except Exception as e_f: raise FileProcessingError(f"ファイル処理中エラー ({filepath}): {e_f}")
|
| 423 |
+
|
| 424 |
+
|
| 425 |
+
# --- GUI アプリケーションクラス ---
|
| 426 |
+
class TextProcessorGUI(tk.Tk):
|
| 427 |
+
def __init__(self):
|
| 428 |
+
super().__init__()
|
| 429 |
+
# --- ウィンドウ設定 ---
|
| 430 |
+
self.title("音声合成コーパス前処理ツール")
|
| 431 |
+
self.geometry("850x700")
|
| 432 |
+
style = ttk.Style(self)
|
| 433 |
+
try: style.theme_use('clam') # or 'alt', 'default', 'classic'
|
| 434 |
+
except tk.TclError: pass # テーマが存在しない場合
|
| 435 |
+
|
| 436 |
+
# --- 変数定義 ---
|
| 437 |
+
self.input_files_var = tk.StringVar(value="")
|
| 438 |
+
self.output_folder_var = tk.StringVar(value="")
|
| 439 |
+
self.norm_rules_var = tk.StringVar(value="normalization_rules.yaml") # デフォルトファイル名
|
| 440 |
+
self.janome_udic_var = tk.StringVar(value="")
|
| 441 |
+
self.jtalk_dic_var = tk.StringVar(value="")
|
| 442 |
+
self.jtalk_user_dic_var = tk.StringVar(value="")
|
| 443 |
+
# 利用可能なエンジンのみ選択肢に含める
|
| 444 |
+
engine_choices = [eng for eng in ['janome', 'pyopenjtalk'] if globals()[f"{eng.upper()}_AVAILABLE"]]
|
| 445 |
+
default_engine = 'pyopenjtalk' if PYOPENJTALK_AVAILABLE else ('janome' if JANOME_AVAILABLE else '')
|
| 446 |
+
self.engine_var = tk.StringVar(value=default_engine)
|
| 447 |
+
self.output_format_vars = {"tsv": tk.BooleanVar(value=False), "jsonl": tk.BooleanVar(value=True)}
|
| 448 |
+
self.reading_format_var = tk.StringVar(value="katakana")
|
| 449 |
+
self.output_lab_var = tk.BooleanVar(value=False)
|
| 450 |
+
self.output_comp_var = tk.BooleanVar(value=False)
|
| 451 |
+
self.encoding_var = tk.StringVar(value="utf-8")
|
| 452 |
+
self.processing_active = False # 処理中フラグ
|
| 453 |
+
|
| 454 |
+
# --- ウィジェット作成 & ロギング設定 ---
|
| 455 |
+
self._create_widgets()
|
| 456 |
+
self._setup_logging()
|
| 457 |
+
|
| 458 |
+
# --- エンジン利用不可の場合の警告 ---
|
| 459 |
+
if not PYOPENJTALK_AVAILABLE:
|
| 460 |
+
logger.warning("pyopenjtalkが見つからないか初期化に失敗しました。pyopenjtalkエンジンは選択できません。")
|
| 461 |
+
if not JANOME_AVAILABLE:
|
| 462 |
+
logger.warning("Janomeが見つかりません。Janomeエンジンは選択できません。")
|
| 463 |
+
if not engine_choices:
|
| 464 |
+
messagebox.showerror("致命的エラー", "利用可能な読み推定エンジン(Janome/pyopenjtalk)がありません。\nライブラリのインストールを確認してください。")
|
| 465 |
+
self.destroy() # アプリケーション終了
|
| 466 |
+
|
| 467 |
+
def _create_widgets(self):
|
| 468 |
+
main_frame = ttk.Frame(self, padding="10")
|
| 469 |
+
main_frame.pack(expand=True, fill=tk.BOTH)
|
| 470 |
+
main_frame.columnconfigure(1, weight=1) # 2列目が伸びるように
|
| 471 |
+
|
| 472 |
+
row_idx = 0
|
| 473 |
+
|
| 474 |
+
# --- 入出力パス ---
|
| 475 |
+
io_frame = ttk.LabelFrame(main_frame, text=" 入出力 ", padding="10")
|
| 476 |
+
io_frame.grid(row=row_idx, column=0, columnspan=3, sticky="ew", padx=5, pady=(0,5))
|
| 477 |
+
io_frame.columnconfigure(1, weight=1)
|
| 478 |
+
ttk.Label(io_frame, text="入力ファイル:").grid(row=0, column=0, sticky=tk.W, padx=5, pady=2)
|
| 479 |
+
ttk.Entry(io_frame, textvariable=self.input_files_var).grid(row=0, column=1, sticky="ew", padx=5, pady=2)
|
| 480 |
+
ttk.Button(io_frame, text="選択...", width=8, command=self._select_input_files).grid(row=0, column=2, padx=(0,5), pady=2)
|
| 481 |
+
ttk.Label(io_frame, text="出力フォルダ:").grid(row=1, column=0, sticky=tk.W, padx=5, pady=2)
|
| 482 |
+
ttk.Entry(io_frame, textvariable=self.output_folder_var).grid(row=1, column=1, sticky="ew", padx=5, pady=2)
|
| 483 |
+
ttk.Button(io_frame, text="選択...", width=8, command=self._select_output_folder).grid(row=1, column=2, padx=(0,5), pady=2)
|
| 484 |
+
row_idx += 1
|
| 485 |
+
|
| 486 |
+
# --- 設定ファイル ---
|
| 487 |
+
config_frame = ttk.LabelFrame(main_frame, text=" 設定ファイル ", padding="10")
|
| 488 |
+
config_frame.grid(row=row_idx, column=0, columnspan=3, sticky="ew", padx=5, pady=5)
|
| 489 |
+
config_frame.columnconfigure(1, weight=1)
|
| 490 |
+
ttk.Label(config_frame, text="正規化ルール:").grid(row=0, column=0, sticky=tk.W, padx=5, pady=2)
|
| 491 |
+
ttk.Entry(config_frame, textvariable=self.norm_rules_var).grid(row=0, column=1, sticky="ew", padx=5, pady=2)
|
| 492 |
+
ttk.Button(config_frame, text="選択...", width=8, command=self._select_norm_rules).grid(row=0, column=2, padx=(0,5), pady=2)
|
| 493 |
+
row_idx += 1
|
| 494 |
+
|
| 495 |
+
# --- エンジンと辞書 ---
|
| 496 |
+
engine_frame = ttk.LabelFrame(main_frame, text=" 読み推定エンジンと辞書 ", padding="10")
|
| 497 |
+
engine_frame.grid(row=row_idx, column=0, columnspan=3, sticky="ew", padx=5, pady=5)
|
| 498 |
+
engine_frame.columnconfigure(1, weight=1)
|
| 499 |
+
engine_choices = [eng for eng in ['janome', 'pyopenjtalk'] if globals()[f"{eng.upper()}_AVAILABLE"]]
|
| 500 |
+
ttk.Label(engine_frame, text="エンジン:").grid(row=0, column=0, sticky=tk.W, padx=5, pady=2)
|
| 501 |
+
engine_combo = ttk.Combobox(engine_frame, textvariable=self.engine_var, values=engine_choices, state="readonly", width=15)
|
| 502 |
+
engine_combo.grid(row=0, column=1, sticky=tk.W, padx=5, pady=2)
|
| 503 |
+
|
| 504 |
+
ttk.Label(engine_frame, text="Janomeユーザー辞書:").grid(row=1, column=0, sticky=tk.W, padx=5, pady=2)
|
| 505 |
+
ttk.Entry(engine_frame, textvariable=self.janome_udic_var).grid(row=1, column=1, sticky="ew", padx=5, pady=2)
|
| 506 |
+
ttk.Button(engine_frame, text="選択...", width=8, command=lambda: self._select_file(self.janome_udic_var, "Janome辞書", [("CSV", "*.csv")])).grid(row=1, column=2, padx=(0,5), pady=2)
|
| 507 |
+
|
| 508 |
+
ttk.Label(engine_frame, text="JTalkシステム辞書:").grid(row=2, column=0, sticky=tk.W, padx=5, pady=2)
|
| 509 |
+
ttk.Entry(engine_frame, textvariable=self.jtalk_dic_var).grid(row=2, column=1, sticky="ew", padx=5, pady=2)
|
| 510 |
+
ttk.Button(engine_frame, text="選択...", width=8, command=self._select_jtalk_dic).grid(row=2, column=2, padx=(0,5), pady=2)
|
| 511 |
+
|
| 512 |
+
ttk.Label(engine_frame, text="JTalkユーザー辞書:").grid(row=3, column=0, sticky=tk.W, padx=5, pady=2)
|
| 513 |
+
ttk.Entry(engine_frame, textvariable=self.jtalk_user_dic_var).grid(row=3, column=1, sticky="ew", padx=5, pady=2)
|
| 514 |
+
ttk.Button(engine_frame, text="選択...", width=8, command=lambda: self._select_file(self.jtalk_user_dic_var, "JTalkユーザー辞書", [("CSV/DIC", "*.csv *.dic")])).grid(row=3, column=2, padx=(0,5), pady=2)
|
| 515 |
+
row_idx += 1
|
| 516 |
+
|
| 517 |
+
# --- 出力オプション ---
|
| 518 |
+
output_frame = ttk.LabelFrame(main_frame, text=" 出力オプション ", padding="10")
|
| 519 |
+
output_frame.grid(row=row_idx, column=0, columnspan=3, sticky="ew", padx=5, pady=5)
|
| 520 |
+
output_frame.columnconfigure(1, weight=1) # 右側のスペースを確保
|
| 521 |
+
|
| 522 |
+
# コーパス形式 (横並び)
|
| 523 |
+
cf_frame = ttk.Frame(output_frame)
|
| 524 |
+
cf_frame.grid(row=0, column=0, columnspan=4, sticky=tk.W, pady=2)
|
| 525 |
+
ttk.Label(cf_frame, text="コーパス形式:").pack(side=tk.LEFT, padx=(5,10))
|
| 526 |
+
ttk.Checkbutton(cf_frame, text="TSV", variable=self.output_format_vars["tsv"]).pack(side=tk.LEFT, padx=5)
|
| 527 |
+
ttk.Checkbutton(cf_frame, text="JSONL", variable=self.output_format_vars["jsonl"]).pack(side=tk.LEFT, padx=5)
|
| 528 |
+
|
| 529 |
+
# 読みの形式
|
| 530 |
+
rf_frame = ttk.Frame(output_frame)
|
| 531 |
+
rf_frame.grid(row=1, column=0, columnspan=4, sticky=tk.W, pady=2)
|
| 532 |
+
ttk.Label(rf_frame, text="読みの形式:").pack(side=tk.LEFT, padx=(5,10))
|
| 533 |
+
reading_combo = ttk.Combobox(rf_frame, textvariable=self.reading_format_var, values=['katakana', 'hiragana'], state="readonly", width=10)
|
| 534 |
+
reading_combo.pack(side=tk.LEFT, padx=5)
|
| 535 |
+
|
| 536 |
+
# その他チェックボックス (横並び)
|
| 537 |
+
oc_frame = ttk.Frame(output_frame)
|
| 538 |
+
oc_frame.grid(row=2, column=0, columnspan=4, sticky=tk.W, pady=2)
|
| 539 |
+
ttk.Checkbutton(oc_frame, text="LAB出力 (JTalk)", variable=self.output_lab_var).pack(side=tk.LEFT, padx=5)
|
| 540 |
+
ttk.Checkbutton(oc_frame, text="比較データ出力", variable=self.output_comp_var).pack(side=tk.LEFT, padx=5)
|
| 541 |
+
|
| 542 |
+
# エンコーディング
|
| 543 |
+
enc_frame = ttk.Frame(output_frame)
|
| 544 |
+
enc_frame.grid(row=3, column=0, columnspan=4, sticky=tk.W, pady=2)
|
| 545 |
+
ttk.Label(enc_frame, text="入力エンコーディング:").pack(side=tk.LEFT, padx=(5, 10))
|
| 546 |
+
encoding_entry = ttk.Entry(enc_frame, textvariable=self.encoding_var, width=12)
|
| 547 |
+
encoding_entry.pack(side=tk.LEFT, padx=5)
|
| 548 |
+
row_idx += 1
|
| 549 |
+
|
| 550 |
+
# --- 実行ボタン ---
|
| 551 |
+
self.run_button = ttk.Button(main_frame, text="処理実行", command=self._start_processing, style="Accent.TButton") # スタイル適用例
|
| 552 |
+
self.run_button.grid(row=row_idx, column=0, columnspan=3, pady=15)
|
| 553 |
+
row_idx += 1
|
| 554 |
+
|
| 555 |
+
# --- ログ表示 ---
|
| 556 |
+
log_frame = ttk.LabelFrame(main_frame, text=" ログ ", padding="10")
|
| 557 |
+
log_frame.grid(row=row_idx, column=0, columnspan=3, sticky="nsew", padx=5, pady=(5,0))
|
| 558 |
+
main_frame.rowconfigure(row_idx, weight=1) # ログエリアが伸縮するように
|
| 559 |
+
log_frame.grid_rowconfigure(0, weight=1)
|
| 560 |
+
log_frame.grid_columnconfigure(0, weight=1)
|
| 561 |
+
self.log_text = scrolledtext.ScrolledText(log_frame, wrap=tk.WORD, height=10, state=tk.DISABLED, bd=0, relief=tk.FLAT) # ボーダー調整
|
| 562 |
+
self.log_text.grid(row=0, column=0, sticky="nsew")
|
| 563 |
+
|
| 564 |
+
# --- ファイル/フォルダ選択メソッド ---
|
| 565 |
+
def _select_input_files(self):
|
| 566 |
+
fpaths = filedialog.askopenfilenames(title="入力ファイル選択 (複数可)", filetypes=[("テキストファイル", "*.txt"), ("全ファイル", "*.*")])
|
| 567 |
+
if fpaths: self.input_files_var.set(";".join(fpaths)) # 区切り文字で結合
|
| 568 |
+
|
| 569 |
+
def _select_output_folder(self):
|
| 570 |
+
fpath = filedialog.askdirectory(title="出力フォルダ選択")
|
| 571 |
+
if fpath: self.output_folder_var.set(fpath)
|
| 572 |
+
|
| 573 |
+
def _select_file(self, var, title, filetypes):
|
| 574 |
+
fpath = filedialog.askopenfilename(title=title, filetypes=filetypes)
|
| 575 |
+
if fpath: var.set(fpath)
|
| 576 |
+
|
| 577 |
+
def _select_norm_rules(self):
|
| 578 |
+
self._select_file(self.norm_rules_var, "正規化ルール選択", [("YAML", "*.yaml *.yml"), ("全ファイル", "*.*")])
|
| 579 |
+
|
| 580 |
+
def _select_jtalk_dic(self):
|
| 581 |
+
fpath = filedialog.askdirectory(title="JTalkシステム辞書フォルダ選択 (例: .../open_jtalk_dic_utf_8-1.11)")
|
| 582 |
+
if fpath: self.jtalk_dic_var.set(fpath)
|
| 583 |
+
|
| 584 |
+
# --- ロギング設定 ---
|
| 585 |
+
def _setup_logging(self):
|
| 586 |
+
class TextHandler(logging.Handler):
|
| 587 |
+
def __init__(self, text_widget):
|
| 588 |
+
logging.Handler.__init__(self)
|
| 589 |
+
self.text_widget = text_widget
|
| 590 |
+
self.queue = [] # メッセージキュー
|
| 591 |
+
self.processing = False
|
| 592 |
+
|
| 593 |
+
def schedule_update(self):
|
| 594 |
+
# 複数メッセージをまとめて更新
|
| 595 |
+
if not self.processing:
|
| 596 |
+
self.processing = True
|
| 597 |
+
self.text_widget.after(50, self._update_widget) # 50ms後に更新
|
| 598 |
+
|
| 599 |
+
def _update_widget(self):
|
| 600 |
+
self.processing = False
|
| 601 |
+
if not self.text_widget.winfo_exists(): return
|
| 602 |
+
|
| 603 |
+
self.text_widget.configure(state='normal')
|
| 604 |
+
while self.queue:
|
| 605 |
+
record = self.queue.pop(0)
|
| 606 |
+
msg = self.format(record)
|
| 607 |
+
self.text_widget.insert(tk.END, msg + '\n')
|
| 608 |
+
self.text_widget.configure(state='disabled')
|
| 609 |
+
self.text_widget.yview(tk.END)
|
| 610 |
+
|
| 611 |
+
def emit(self, record):
|
| 612 |
+
self.queue.append(record)
|
| 613 |
+
self.schedule_update()
|
| 614 |
+
|
| 615 |
+
gui_handler = TextHandler(self.log_text)
|
| 616 |
+
gui_handler.setFormatter(logging.Formatter('%(asctime)s - %(levelname)s - %(message)s', datefmt='%H:%M:%S'))
|
| 617 |
+
logging.getLogger().addHandler(gui_handler)
|
| 618 |
+
logging.getLogger().setLevel(logging.INFO) # INFOレベル以上をGUIに表示
|
| 619 |
+
|
| 620 |
+
# --- 処理実行メソッド ---
|
| 621 |
+
def _start_processing(self):
|
| 622 |
+
if self.processing_active: logger.warning("処理中"); return
|
| 623 |
+
|
| 624 |
+
# --- 設定値取得 & バリデーション ---
|
| 625 |
+
self.params = {} # パラメータを格納する辞書
|
| 626 |
+
try:
|
| 627 |
+
self.params['input_files'] = [f.strip() for f in self.input_files_var.get().split(';') if f.strip()]
|
| 628 |
+
self.params['output_folder'] = self.output_folder_var.get()
|
| 629 |
+
self.params['norm_rules'] = self.norm_rules_var.get()
|
| 630 |
+
self.params['engine'] = self.engine_var.get()
|
| 631 |
+
self.params['janome_udic'] = self.janome_udic_var.get() or None
|
| 632 |
+
self.params['jtalk_dic'] = self.jtalk_dic_var.get() or None
|
| 633 |
+
self.params['jtalk_user_dic'] = self.jtalk_user_dic_var.get() or None
|
| 634 |
+
self.params['output_format'] = [fmt for fmt, var in self.output_format_vars.items() if var.get()]
|
| 635 |
+
self.params['reading_format'] = self.reading_format_var.get()
|
| 636 |
+
self.params['output_lab'] = self.output_lab_var.get()
|
| 637 |
+
self.params['output_comparison'] = self.output_comp_var.get()
|
| 638 |
+
self.params['encoding'] = self.encoding_var.get().strip() or 'utf-8'
|
| 639 |
+
|
| 640 |
+
if not self.params['input_files']: raise ValueError("入力ファイル未選択")
|
| 641 |
+
if not self.params['output_folder']: raise ValueError("出力フォルダ未選択")
|
| 642 |
+
if not self.params['output_format']: raise ValueError("出力コーパス形式未選択")
|
| 643 |
+
if self.params['output_lab'] and self.params['engine'] != 'pyopenjtalk':
|
| 644 |
+
raise ValueError("LAB出力はpyopenjtalkエンジン選択時のみ可能")
|
| 645 |
+
if not os.path.exists(self.params['norm_rules']):
|
| 646 |
+
logger.warning(f"正規化ルールファイル '{self.params['norm_rules']}' が存在しません。デフォルトルールを使用します。")
|
| 647 |
+
# デフォルトを使用する場合、ファイルパスはNoneとして扱う方が良いかも
|
| 648 |
+
# self.params['norm_rules'] = None # またはそのまま存在しないパスを渡す
|
| 649 |
+
|
| 650 |
+
except ValueError as ve:
|
| 651 |
+
messagebox.showerror("入力エラー", str(ve)); return
|
| 652 |
+
except Exception as e:
|
| 653 |
+
messagebox.showerror("設定エラー", f"設定値の取得中にエラー: {e}"); return
|
| 654 |
+
|
| 655 |
+
# 処理開始
|
| 656 |
+
self.processing_active = True
|
| 657 |
+
self.run_button.config(state=tk.DISABLED, text="処理中...")
|
| 658 |
+
self.log_text.configure(state='normal'); self.log_text.delete('1.0', tk.END); self.log_text.configure(state='disabled')
|
| 659 |
+
logger.info("処理を開始します...")
|
| 660 |
+
|
| 661 |
+
# 別スレッドで実行
|
| 662 |
+
self.processing_thread = threading.Thread(target=self._run_processing_logic, daemon=True)
|
| 663 |
+
self.processing_thread.start()
|
| 664 |
+
self.after(100, self._check_thread) # スレッド監視開始
|
| 665 |
+
|
| 666 |
+
def _check_thread(self):
|
| 667 |
+
if self.processing_thread.is_alive():
|
| 668 |
+
self.after(100, self._check_thread)
|
| 669 |
+
# スレッド終了後の処理は _run_processing_logic の最後で after(0, ...) を使う
|
| 670 |
+
|
| 671 |
+
def _run_processing_logic(self):
|
| 672 |
+
"""実際の処理 (別スレッドで実行)"""
|
| 673 |
+
exit_code = 1
|
| 674 |
+
final_message = "処理中にエラーが発生しました。"
|
| 675 |
+
message_type = "error"
|
| 676 |
+
p = self.params # 短縮名
|
| 677 |
+
|
| 678 |
+
try:
|
| 679 |
+
# --- ログ表示 (主要パラメータ) ---
|
| 680 |
+
logger.info(f"入力ファイル数: {len(p['input_files'])}")
|
| 681 |
+
logger.info(f"出力フォルダ: {p['output_folder']}")
|
| 682 |
+
logger.info(f"正規化ルール: {p['norm_rules'] if os.path.exists(p['norm_rules']) else 'デフォルト'}")
|
| 683 |
+
logger.info(f"エンジン: {p['engine']}")
|
| 684 |
+
logger.info(f"出力形式: {p['output_format']}")
|
| 685 |
+
logger.info(f"LAB出力: {p['output_lab']}")
|
| 686 |
+
|
| 687 |
+
# --- 初期化 ---
|
| 688 |
+
normalizer = TextNormalizer(p['norm_rules'])
|
| 689 |
+
pron_estimator = PronunciationEstimator(
|
| 690 |
+
engine=p['engine'], janome_udic_path=p['janome_udic'],
|
| 691 |
+
jtalk_dic_path=p['jtalk_dic'], jtalk_user_dic_path=p['jtalk_user_dic']
|
| 692 |
+
)
|
| 693 |
+
|
| 694 |
+
# --- ファイル処理ループ ---
|
| 695 |
+
total_success_lines, total_error_lines, file_error_count, processed_files = 0, 0, 0, 0
|
| 696 |
+
for i, filepath in enumerate(p['input_files']):
|
| 697 |
+
logger.info(f"--- 処理中 ({i+1}/{len(p['input_files'])}): {os.path.basename(filepath)} ---")
|
| 698 |
+
try:
|
| 699 |
+
norm_filepath = os.path.normpath(filepath)
|
| 700 |
+
if not os.path.isfile(norm_filepath):
|
| 701 |
+
logger.warning(f"スキップ (非ファイル): {norm_filepath}"); file_error_count += 1; continue
|
| 702 |
+
s, e = process_file( # process_fileは (成功数, エラー数) を返す想定
|
| 703 |
+
norm_filepath, p['output_folder'], normalizer, pron_estimator,
|
| 704 |
+
p['output_format'], p['output_lab'], p['reading_format'],
|
| 705 |
+
p['encoding'], p['output_comparison']
|
| 706 |
+
)
|
| 707 |
+
total_success_lines += s; total_error_lines += e
|
| 708 |
+
if e > 0: file_error_count += 1
|
| 709 |
+
processed_files += 1
|
| 710 |
+
except FileProcessingError as e_fp: logger.error(f"{os.path.basename(filepath)} ファイルエラー: {e_fp}"); file_error_count += 1
|
| 711 |
+
except Exception as e_f: logger.error(f"{os.path.basename(filepath)} 予期せぬエラー: {e_f}", exc_info=True); file_error_count += 1
|
| 712 |
+
# logger.info(f"--- 処理完了 ({i+1}/{len(p['input_files'])}): {os.path.basename(filepath)} ---") # ログが冗長ならコメントアウト
|
| 713 |
+
|
| 714 |
+
# --- 未知語出力 ---
|
| 715 |
+
if pron_estimator.unknown_words:
|
| 716 |
+
unknown_file = os.path.join(p['output_folder'], "unknown_words.txt")
|
| 717 |
+
try:
|
| 718 |
+
with open(unknown_file, 'w', encoding='utf-8') as f:
|
| 719 |
+
f.write("\n".join(sorted(list(pron_estimator.unknown_words))))
|
| 720 |
+
logger.warning(f"未知語リスト出力: {unknown_file}")
|
| 721 |
+
except Exception as e_unk: logger.error(f"未知語リスト出力失敗: {e_unk}")
|
| 722 |
+
|
| 723 |
+
# --- 最終結果判定 ---
|
| 724 |
+
logger.info("=" * 30 + " 処理結果 " + "=" * 30)
|
| 725 |
+
logger.info(f"処理ファイル数: {processed_files} / {len(p['input_files'])}")
|
| 726 |
+
logger.info(f"エラー発生ファイル数: {file_error_count}")
|
| 727 |
+
logger.info(f"総処理行数: {total_success_lines + total_error_lines}")
|
| 728 |
+
logger.info(f" 正常処理行数: {total_success_lines}")
|
| 729 |
+
logger.info(f" エラー発生行数: {total_error_lines}")
|
| 730 |
+
logger.info("=" * 68)
|
| 731 |
+
|
| 732 |
+
if file_error_count > 0 or total_error_lines > 0:
|
| 733 |
+
final_message = "処理完了(エラーあり)。詳細はログを確認してください。"
|
| 734 |
+
message_type = "warning"
|
| 735 |
+
exit_code = 1
|
| 736 |
+
else:
|
| 737 |
+
final_message = "すべての処理が正常に完了しました。"
|
| 738 |
+
message_type = "info"
|
| 739 |
+
exit_code = 0
|
| 740 |
+
|
| 741 |
+
except InitializationError as e_init: final_message = f"初期化エラー: {e_init}"; message_type = "error"
|
| 742 |
+
except Exception as e_main: final_message = f"予期せぬエラー: {e_main}"; message_type = "error"; logger.error("予期せぬエラー", exc_info=True)
|
| 743 |
+
|
| 744 |
+
# --- GUI要素の更新 (メインスレッドで実行) ---
|
| 745 |
+
def update_gui():
|
| 746 |
+
self.processing_active = False
|
| 747 |
+
if self.winfo_exists(): # ウィンドウが閉じられていないか確認
|
| 748 |
+
self.run_button.config(state=tk.NORMAL, text="処理実行")
|
| 749 |
+
if message_type == "info": messagebox.showinfo("完了", final_message)
|
| 750 |
+
elif message_type == "warning": messagebox.showwarning("完了(一部エラー)", final_message)
|
| 751 |
+
else: messagebox.showerror("エラー", final_message)
|
| 752 |
+
self.after(0, update_gui)
|
| 753 |
+
# スレッド終了
|
| 754 |
+
|
| 755 |
+
|
| 756 |
+
# --- アプリケーション起動 ---
|
| 757 |
if __name__ == "__main__":
|
| 758 |
+
if not ENABLE_GUI:
|
| 759 |
+
print("エラー: GUI表示に必要な tkinter が見つかりません。", file=sys.stderr)
|
| 760 |
+
sys.exit(1)
|
| 761 |
+
|
| 762 |
+
# 利用可能なエンジンがない場合のチェックはGUIクラスの __init__ で行う
|
| 763 |
+
app = TextProcessorGUI()
|
| 764 |
+
app.mainloop()
|