speak_qt_async.py ダウンロード/コピー
speak_qt_async.py
speak_qt_async.py
1import sys
2import os
3import traceback
4import shutil
5import re
6import uuid
7import glob
8import time
9from typing import List, Optional, Dict, Tuple, Any
10
11try:
12 import chardet
13except ImportError:
14 print("Error: Missing library: chardet. Please run 'pip install chardet'")
15 sys.exit(1)
16
17try:
18 import tktts
19 from pydub import AudioSegment
20except ImportError:
21 print("Error: Missing libraries: tktts or pydub. Please run 'pip install tktts pydub'")
22 sys.exit(1)
23
24
25from PySide6.QtWidgets import (
26 QApplication, QWidget, QVBoxLayout, QHBoxLayout,
27 QTextEdit, QLineEdit, QPushButton, QFileDialog,
28 QSlider, QDoubleSpinBox, QComboBox, QLabel, QMessageBox,
29 QProgressBar, QGridLayout, QFrame, QTabWidget, QSizePolicy, QCheckBox
30)
31from PySide6.QtCore import Qt, QUrl, QThread, Signal, Slot, QRect, QDir
32from PySide6.QtMultimedia import QMediaPlayer, QAudioOutput, QMediaDevices
33
34
35# --- 定数とヘルパー関数(tkttsで利用される引数構造を維持) ---
36DEFAULT_ENGINE = "pyttsx3"
37DEFAULT_VOICEVOX_ENDPOINT = "http://127.0.0.1:50021"
38DEFAULT_QWEN_API_SERVER = "http://192.168.27.18:8004"
39DEFAULT_IRODORI_API_SERVER = "http://192.168.27.18:8002"
40DEFAULT_AQUESTALK_PATH = "AquesTalkPlayer.exe"
41DEFAULT_TEMP_DIR = "tts_temp_wavs"
42TEMP_WAV_PREFIX = "_tktts_tmp_"
43TEMP_WAV_EXT = ".wav"
44INI_FILE_NAME = os.path.splitext(__file__)[0] + ".ini"
45
46
47def detect_encoding(file_path):
48 """ファイルの文字コードを判定して開く"""
49 with open(file_path, 'rb') as f:
50 raw_data = f.read()
51 result = chardet.detect(raw_data)
52 return result['encoding']
53
54# グローバルな置換辞書とタイムスタンプキャッシュ
55_GLOBAL_REPLACE_DICT_CACHE = {}
56_GLOBAL_TIMESTAMP_CACHE = {}
57
58def load_replace_dict(ini_path, force_reload=False):
59 """
60 replace.iniを読み込んで辞書を作成(TOML風の簡易実装)。
61 タイムスタンプをチェックし、更新がなければキャッシュを使用。
62 """
63 if not ini_path or not os.path.isfile(ini_path):
64 return {}
65
66 try:
67 current_timestamp = os.path.getmtime(ini_path)
68 except OSError:
69 # ファイルが存在しない、またはアクセス権がない場合
70 return {}
71
72 # キャッシュチェック
73 if not force_reload and ini_path in _GLOBAL_REPLACE_DICT_CACHE and \
74 _GLOBAL_TIMESTAMP_CACHE.get(ini_path) == current_timestamp:
75 return _GLOBAL_REPLACE_DICT_CACHE[ini_path]
76
77 # ファイルの読み込みとパース
78 replace_dict = {}
79 try:
80 encoding = detect_encoding(ini_path)
81 if encoding is None: encoding = 'utf-8'
82 with open(ini_path, 'r', encoding=encoding) as f:
83 for line in f:
84 line = line.rstrip('\n')
85 if line.startswith('#') or '=' not in line:
86 continue
87
88 # キーと値を抽出する正規表現(キーはクォートあり/なしに対応)
89 match = re.match(r"""^(['"].+?['"]|[^=]+?)=(.*)$""", line)
90 if not match:
91 continue
92
93 raw_key, val = match.groups()
94 # キーからクォートを除去
95 key = raw_key[1:-1] if (raw_key.startswith("'") and raw_key.endswith("'")) or (raw_key.startswith('"') and raw_key.endswith('"')) else raw_key.strip()
96 replace_dict[key] = val.strip()
97
98 # キャッシュを更新
99 _GLOBAL_REPLACE_DICT_CACHE[ini_path] = replace_dict
100 _GLOBAL_TIMESTAMP_CACHE[ini_path] = current_timestamp
101 print(f"Status: Loaded/Reloaded INI file: {os.path.basename(ini_path)}")
102 return replace_dict
103
104 except Exception as e:
105 print(f" [skip] Failed to read/parse {ini_path}: {e}")
106 return {}
107
108def apply_replacements(text, replace_dict: Dict[str, str]):
109 """テキストに対して正規表現による置換を適用"""
110
111 replace_list = list(replace_dict.items())
112
113 for pattern, replacement in replace_list:
114 try:
115 # re.IGNORECASE (大文字小文字無視) と re.MULTILINE (複数行モード) を適用
116 text = re.sub(pattern, replacement, text, flags=re.IGNORECASE | re.MULTILINE)
117 except Exception as e:
118 print(f"re.sub error for [{pattern}]: {e}")
119 return text
120
121
122class ArgsStub:
123 """tkttsのヘルパー関数に引数を渡すためのスタブクラス"""
124 def __init__(self, **kwargs):
125 for k, v in kwargs.items():
126 setattr(self, k, v)
127
128
129class MyTTSWorker(QThread):
130 """音声生成を非同期で実行するワーカークラス (tkttsの内部ロジックを直接実行)"""
131 finished = Signal(str)
132 error = Signal(str)
133 progress = Signal(int)
134
135 def __init__(self, text, outfile, tts_engine, speed_rate, pitch, instruction, voice_name, tmp_files,
136 aquestalk_path, temp_dir, voicevox_endpoint,
137 qwen3_language="Japanese",
138 qwen3_model_id="Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice",
139 qwen3_device="auto",
140 qwen3_dtype="auto",
141 qwen3_instruct="",
142 qwen_api_server="",
143 qwen_api_timeout=1800.0,
144 irodori_caption="落ち着いた自然な声で、明瞭に読み上げる。",
145 irodori_ref_wav="",
146 irodori_model_id="Aratako/Irodori-TTS-v4.1-Small",
147 irodori_device="auto",
148 irodori_precision="auto",
149 irodori_num_steps=40,
150 irodori_duration_scale=1.0,
151 irodori_seed="0",
152 irodori_api_server=DEFAULT_IRODORI_API_SERVER,
153 irodori_api_timeout=1800.0,
154 irodori_api_sway=False,
155 irodori_api_sway_coeff=-1.0,
156 gemini_model="gemini-3.8-flash-tts",
157 edge_rate="+0%",
158 edge_volume="+0%",
159 edge_pitch="+0Hz",
160 parent=None):
161 super().__init__(parent)
162 self.text = text
163 self.user_outfile = outfile if outfile and outfile.strip() else None
164
165 self.tts_engine = tts_engine.lower()
166 self.speed_rate = speed_rate
167 self.pitch = pitch
168 self.instruction = instruction
169 self.voice_name = voice_name
170 self.tmp_files = tmp_files
171
172 self.aquestalk_path = aquestalk_path
173 self.temp_dir = temp_dir
174 self.voicevox_endpoint = voicevox_endpoint
175
176 self.qwen3_language = qwen3_language
177 self.qwen3_model_id = qwen3_model_id
178 self.qwen3_device = qwen3_device
179 self.qwen3_dtype = qwen3_dtype
180 self.qwen3_instruct = qwen3_instruct
181 self.qwen_api_server = qwen_api_server
182 self.qwen_api_timeout = qwen_api_timeout
183
184 self.irodori_caption = irodori_caption
185 self.irodori_ref_wav = irodori_ref_wav or None
186 self.irodori_model_id = irodori_model_id
187 self.irodori_device = irodori_device
188 self.irodori_precision = irodori_precision
189 self.irodori_num_steps = irodori_num_steps
190 self.irodori_duration_scale = irodori_duration_scale
191 self.irodori_seed = irodori_seed
192 self.irodori_api_server = irodori_api_server
193 self.irodori_api_timeout = irodori_api_timeout
194 self.irodori_api_sway = irodori_api_sway
195 self.irodori_api_sway_coeff = irodori_api_sway_coeff
196 self.gemini_model = gemini_model
197 self.edge_rate = edge_rate
198 self.edge_volume = edge_volume
199 self.edge_pitch = edge_pitch
200
201 # tkttsの引数スタブを更新
202 self.tktts_args = ArgsStub(
203 tts=self.tts_engine,
204 monologue=1,
205 voices=self.voice_name,
206 speak_rate=150,
207 fspeak_rate=self.speed_rate,
208 fspeak_pitch=self.pitch,
209 tinterval=0.5,
210 temp_dir=self.temp_dir,
211 outfile="",
212 instruction=self.instruction,
213 aquestalk_path=self.aquestalk_path,
214 endpoint=self.voicevox_endpoint,
215 elevenlabs_api_key=os.getenv("ELEVENLABS_API_KEY"),
216 gemini_api_key=os.getenv("GEMINI_API_KEY"),
217 gemini_model=self.gemini_model,
218 edge_rate=self.edge_rate,
219 edge_volume=self.edge_volume,
220 edge_pitch=self.edge_pitch,
221
222 # Qwen3-TTS
223 qwen3_language=self.qwen3_language,
224 qwen3_model_id=self.qwen3_model_id,
225 qwen3_device=self.qwen3_device,
226 qwen3_dtype=self.qwen3_dtype,
227 qwen3_instruct=(self.qwen3_instruct or None),
228 qwen_api_server=self.qwen_api_server,
229 qwen_api_timeout=self.qwen_api_timeout,
230
231 # Irodori-TTS
232 irodori_caption=self.irodori_caption,
233 irodori_ref_wav=self.irodori_ref_wav,
234 irodori_ref_wavs=None,
235 irodori_model_id=self.irodori_model_id,
236 irodori_device=self.irodori_device,
237 irodori_precision=self.irodori_precision,
238 irodori_codec_device=None,
239 irodori_codec_precision=None,
240 irodori_num_steps=self.irodori_num_steps,
241 irodori_cfg_scale_text=3.5,
242 irodori_cfg_scale_caption=3.0,
243 irodori_cfg_scale_speaker=5.0,
244 irodori_duration_scale=self.irodori_duration_scale,
245 irodori_seed=self.irodori_seed,
246 irodori_lora_adapter=None,
247
248 # Irodori-TTS Flask API
249 irodori_api_server=self.irodori_api_server,
250 irodori_api_timeout=self.irodori_api_timeout,
251 irodori_api_sway=self.irodori_api_sway,
252 irodori_api_sway_coeff=self.irodori_api_sway_coeff,
253 )
254
255 def run(self):
256 """tktts.py の統一インターフェースで音声を非同期生成する。"""
257
258 self.progress.emit(10)
259
260 try:
261 config = tktts.TTS_ENGINES.get(self.tts_engine)
262 if config is None:
263 self.error.emit(
264 f"TTSエンジン [{self.tts_engine}] の設定が見つかりません。"
265 )
266 return
267
268 if tktts.get_tts(self.tts_engine) is None:
269 self.error.emit(
270 f"TTSエンジン [{self.tts_engine}] のロードに失敗しました。"
271 )
272 return
273
274 lines = self.text.strip().split("\n")
275 dialogue = [(None, line.strip()) for line in lines if line.strip()]
276 if not dialogue:
277 self.error.emit("読み上げるテキストがありません。")
278 return
279
280 temp_dir = tktts.create_temp_dir(self.temp_dir)
281
282 if self.user_outfile:
283 output_path = os.path.abspath(self.user_outfile)
284 else:
285 # GUIでの再生互換性を優先し、一時出力は常にWAVにする。
286 temp_filebody = TEMP_WAV_PREFIX + uuid.uuid4().hex[:8]
287 output_path = os.path.abspath(
288 os.path.join(temp_dir, temp_filebody + "_merged.wav")
289 )
290
291 output_dir = os.path.dirname(output_path)
292 if output_dir:
293 os.makedirs(output_dir, exist_ok=True)
294
295 self.tktts_args.outfile = output_path
296 self.tktts_args.endpoint = self.voicevox_endpoint
297 if self.tts_engine == "gemini":
298 active_api_key = os.getenv("GEMINI_API_KEY")
299 self.tktts_args.gemini_api_key = active_api_key
300 else:
301 active_api_key = os.getenv("ELEVENLABS_API_KEY")
302 self.tktts_args.elevenlabs_api_key = active_api_key
303
304 print(f"Status: {self.tts_engine.upper()}の音声生成開始...")
305 # 音声名を文字列のまま渡すと、WinRTバックエンドでspeaker名として
306 # 正規化され、空白を含む ``Microsoft Ayumi ...`` が切れることがある。
307 # 値として保持する話者マップにして渡す。
308 selected_voice_map = {
309 None: self.voice_name,
310 "": self.voice_name,
311 0: self.voice_name,
312 }
313
314 result = tktts.speak_dialogue(
315 self.tktts_args,
316 dialogue,
317 voice_map=selected_voice_map,
318 replacements={},
319 endpoint=self.voicevox_endpoint,
320 api_key=active_api_key,
321 output_format=None, # 出力パスの拡張子から tktts.py が判定
322 )
323
324 if not result:
325 self.error.emit(
326 f"TTSエンジン [{self.tts_engine}] での音声生成に失敗しました。"
327 )
328 return
329
330 generated_path = result if isinstance(result, str) else output_path
331 if not os.path.isfile(generated_path):
332 self.error.emit(f"生成された音声ファイルが見つかりません: {generated_path}")
333 return
334
335 self.progress.emit(100)
336 self.finished.emit(generated_path)
337
338 if not self.user_outfile:
339 self.tmp_files.append(generated_path)
340
341 except Exception as e:
342 error_msg = f"音声生成またはファイル操作エラー: {type(e).__name__}: {e}"
343 print(error_msg)
344 traceback.print_exc()
345 self.error.emit(error_msg)
346
347
348class MyTTSApp(QWidget):
349 def __init__(self):
350 super().__init__()
351 self.setWindowTitle("統合TTS GUI (コンパクト・リサイズ可能)")
352
353 # メディア関連の初期化
354 self.worker_thread: Optional[MyTTSWorker] = None
355 self.audio_output = QAudioOutput(QMediaDevices.defaultAudioOutput())
356 self.player: QMediaPlayer = QMediaPlayer()
357 self.player.setAudioOutput(self.audio_output)
358 self.current_audio_path: Optional[str] = None
359
360 # 状態管理
361 self.tmp_files = []
362 self.slide_data: Dict[int, str] = {}
363 self.last_dir: Dict[str, str] = {}
364 self.is_converted_text_dirty: bool = True
365
366 self._load_settings()
367
368 # シグナルとスロットの接続
369 self.player.durationChanged.connect(self.set_slider_range)
370 self.player.positionChanged.connect(self.update_slider)
371 self.player.playbackStateChanged.connect(self.update_playback_buttons)
372 self.player.errorOccurred.connect(self.handle_media_player_error)
373
374 self.initUI()
375 self.setGeometry(self.settings.get('geometry', QRect(100, 100, 800, 650)))
376 self.setMinimumSize(400, 450)
377
378 # TTS設定UIの変更を監視し、ダーティフラグを立てる
379 self.text_input_converted.textChanged.connect(self.set_dirty)
380 self.speed_spin.valueChanged.connect(self.set_dirty)
381 self.pitch_spin.valueChanged.connect(self.set_dirty)
382 self.instruction_line.textChanged.connect(self.set_dirty)
383 self.engine_combo.currentIndexChanged.connect(self.update_voice_list) # update_voice_list内でもset_dirtyを呼ぶ
384 self.voice_combo.currentIndexChanged.connect(self.set_dirty) # ボイス変更時
385
386 # Qwen/Irodori settings
387 for w in (
388 self.qwen3_model_line, self.qwen3_language_line, self.qwen3_instruct_line,
389 self.qwen_api_server_line, self.qwen_api_timeout_line,
390 self.irodori_model_line, self.irodori_caption_line,
391 self.irodori_ref_wav_line, self.irodori_seed_line,
392 self.irodori_api_server_line, self.irodori_api_timeout_line,
393 ):
394 w.textChanged.connect(self.set_dirty)
395 self.qwen3_device_combo.currentIndexChanged.connect(self.set_dirty)
396 self.qwen3_dtype_combo.currentIndexChanged.connect(self.set_dirty)
397 self.irodori_device_combo.currentIndexChanged.connect(self.set_dirty)
398 self.irodori_precision_combo.currentIndexChanged.connect(self.set_dirty)
399 self.irodori_steps_spin.valueChanged.connect(self.set_dirty)
400 self.irodori_duration_spin.valueChanged.connect(self.set_dirty)
401 self.irodori_api_sway_check.toggled.connect(self.set_dirty)
402 self.irodori_api_sway_coeff_spin.valueChanged.connect(self.set_dirty)
403 self.gemini_model_line.textChanged.connect(self.set_dirty)
404 self.edge_rate_line.textChanged.connect(self.set_dirty)
405 self.edge_volume_line.textChanged.connect(self.set_dirty)
406 self.edge_pitch_line.textChanged.connect(self.set_dirty)
407
408 self.update_voice_list()
409
410 QApplication.instance().aboutToQuit.connect(self.cleanup_temp_files)
411 QApplication.instance().aboutToQuit.connect(self._save_settings)
412
413
414 # --- 状態管理 ---
415 @Slot()
416 def set_dirty(self):
417 """TTS出力に影響する設定が変更されたとき、フラグを立てる"""
418 self.is_converted_text_dirty = True
419 if self.player.playbackState() != QMediaPlayer.PlaybackState.PlayingState:
420 self.update_playback_buttons(self.player.playbackState())
421
422 def _load_settings(self):
423 # 設定のロードロジックは省略せずに残す
424 default_settings = {
425 'geometry': QRect(100, 100, 800, 650),
426 'temp_dir': DEFAULT_TEMP_DIR,
427 'aquestalk_path': DEFAULT_AQUESTALK_PATH,
428 'voicevox_endpoint': DEFAULT_VOICEVOX_ENDPOINT,
429 'input_file': "input.md",
430 'replace_file': "replace.ini",
431 'replace_file2': "user_replace.ini",
432 'output_file': "",
433 'qwen3_language': "Japanese",
434 'qwen3_model_id': "Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice",
435 'qwen3_device': "auto",
436 'qwen3_dtype': "auto",
437 'qwen3_instruct': "",
438 'qwen_api_server': os.environ.get("QWEN_TTS_API", DEFAULT_QWEN_API_SERVER),
439 'qwen_api_timeout': "1800",
440 'irodori_caption': "落ち着いた自然な声で、明瞭に読み上げる。",
441 'irodori_ref_wav': "",
442 'irodori_model_id': "Aratako/Irodori-TTS-v4.1-Small",
443 'irodori_device': "auto",
444 'irodori_precision': "auto",
445 'irodori_num_steps': "40",
446 'irodori_duration_scale': "1.0",
447 'irodori_seed': "0",
448 'irodori_api_server': os.environ.get("IRODORI_TTS_API", DEFAULT_IRODORI_API_SERVER),
449 'irodori_api_timeout': "1800",
450 'irodori_api_sway': "0",
451 'irodori_api_sway_coeff': "-1.0",
452 'gemini_model': "gemini-3.8-flash-tts",
453 'edge_rate': "+0%",
454 'edge_volume': "+0%",
455 'edge_pitch': "+0Hz",
456 }
457 self.settings: Dict[str, Any] = default_settings.copy()
458
459 # 環境変数に旧来の 198.168... の誤記がある場合も補正する。
460 if self.settings.get("qwen_api_server", "").startswith("http://198.168.27.18:8004"):
461 self.settings["qwen_api_server"] = DEFAULT_QWEN_API_SERVER
462
463 if not os.path.exists(INI_FILE_NAME):
464 return
465
466 try:
467 with open(INI_FILE_NAME, 'r', encoding='utf-8') as f:
468 content = f.read()
469 current_section = None
470 for line in content.splitlines():
471 line = line.strip()
472 if not line or line.startswith('#'):
473 continue
474 if line.startswith('[') and line.endswith(']'):
475 current_section = line[1:-1].strip()
476 elif '=' in line:
477 key, value = line.split('=', 1)
478 key = key.strip()
479 value = value.strip().strip('"')
480
481 if current_section == "window":
482 if key == "x": self.settings['x'] = int(value)
483 elif key == "y": self.settings['y'] = int(value)
484 elif key == "width": self.settings['width'] = int(value)
485 elif key == "height": self.settings['height'] = int(value)
486 elif current_section in ("tts_paths", "qwen3", "qwen_api", "irodori", "irodori_api", "gemini", "edge"):
487 if key in default_settings:
488 self.settings[key] = value
489
490 # 旧版に保存され得た 198.168... の誤記を補正する。
491 if self.settings.get("qwen_api_server", "").startswith("http://198.168.27.18:8004"):
492 self.settings["qwen_api_server"] = DEFAULT_QWEN_API_SERVER
493
494 if 'x' in self.settings:
495 self.settings['geometry'] = QRect(
496 self.settings['x'], self.settings['y'],
497 self.settings.get('width', 800), self.settings.get('height', 650)
498 )
499 except Exception as e:
500 print(f"設定ファイル読み込みエラー: {e}")
501 self.settings = default_settings.copy()
502
503
504 def _save_settings(self):
505 # 設定保存ロジックは省略せずに残す
506 geom = self.geometry()
507
508 current_settings = {
509 'temp_dir': self.temp_dir_line.text(),
510 'aquestalk_path': self.aquestalk_path_line.text(),
511 'voicevox_endpoint': self.voicevox_endpoint_line.text(),
512 'input_file': self.input_file_line.text(),
513 'replace_file': self.replace_ini_line.text(),
514 'replace_file2': self.replace_ini2_line.text(),
515 'output_file': self.output_line.text(),
516 'qwen3_language': self.qwen3_language_line.text(),
517 'qwen3_model_id': self.qwen3_model_line.text(),
518 'qwen3_device': self.qwen3_device_combo.currentText(),
519 'qwen3_dtype': self.qwen3_dtype_combo.currentText(),
520 'qwen3_instruct': self.qwen3_instruct_line.text(),
521 'qwen_api_server': self.qwen_api_server_line.text(),
522 'qwen_api_timeout': self.qwen_api_timeout_line.text(),
523 'irodori_caption': self.irodori_caption_line.text(),
524 'irodori_ref_wav': self.irodori_ref_wav_line.text(),
525 'irodori_model_id': self.irodori_model_line.text(),
526 'irodori_device': self.irodori_device_combo.currentText(),
527 'irodori_precision': self.irodori_precision_combo.currentText(),
528 'irodori_num_steps': str(self.irodori_steps_spin.value()),
529 'irodori_duration_scale': str(self.irodori_duration_spin.value()),
530 'irodori_seed': self.irodori_seed_line.text(),
531 'irodori_api_server': self.irodori_api_server_line.text(),
532 'irodori_api_timeout': self.irodori_api_timeout_line.text(),
533 'irodori_api_sway': "1" if self.irodori_api_sway_check.isChecked() else "0",
534 'irodori_api_sway_coeff': str(self.irodori_api_sway_coeff_spin.value()),
535 'gemini_model': self.gemini_model_line.text(),
536 'edge_rate': self.edge_rate_line.text(),
537 'edge_volume': self.edge_volume_line.text(),
538 'edge_pitch': self.edge_pitch_line.text(),
539 }
540
541 content = (
542 f'[window]\n'
543 f'x = {geom.x()}\n'
544 f'y = {geom.y()}\n'
545 f'width = {geom.width()}\n'
546 f'height = {geom.height()}\n'
547 f'\n'
548 f'[tts_paths]\n'
549 f'temp_dir = "{current_settings["temp_dir"]}"\n'
550 f'aquestalk_path = "{current_settings["aquestalk_path"]}"\n'
551 f'voicevox_endpoint = "{current_settings["voicevox_endpoint"]}"\n'
552 f'input_file = "{current_settings["input_file"]}"\n'
553 f'replace_file = "{current_settings["replace_file"]}"\n'
554 f'replace_file2 = "{current_settings["replace_file2"]}"\n'
555 f'output_file = "{current_settings["output_file"]}"\n'
556 f'\n'
557 f'[qwen3]\n'
558 f'qwen3_language = "{current_settings["qwen3_language"]}"\n'
559 f'qwen3_model_id = "{current_settings["qwen3_model_id"]}"\n'
560 f'qwen3_device = "{current_settings["qwen3_device"]}"\n'
561 f'qwen3_dtype = "{current_settings["qwen3_dtype"]}"\n'
562 f'qwen3_instruct = "{current_settings["qwen3_instruct"]}"\n'
563 f'\n'
564 f'[qwen_api]\n'
565 f'qwen_api_server = "{current_settings["qwen_api_server"]}"\n'
566 f'qwen_api_timeout = "{current_settings["qwen_api_timeout"]}"\n'
567 f'\n'
568 f'[irodori]\n'
569 f'irodori_caption = "{current_settings["irodori_caption"]}"\n'
570 f'irodori_ref_wav = "{current_settings["irodori_ref_wav"]}"\n'
571 f'irodori_model_id = "{current_settings["irodori_model_id"]}"\n'
572 f'irodori_device = "{current_settings["irodori_device"]}"\n'
573 f'irodori_precision = "{current_settings["irodori_precision"]}"\n'
574 f'irodori_num_steps = "{current_settings["irodori_num_steps"]}"\n'
575 f'irodori_duration_scale = "{current_settings["irodori_duration_scale"]}"\n'
576 f'irodori_seed = "{current_settings["irodori_seed"]}"\n'
577 f'\n'
578 f'[irodori_api]\n'
579 f'irodori_api_server = "{current_settings["irodori_api_server"]}"\n'
580 f'irodori_api_timeout = "{current_settings["irodori_api_timeout"]}"\n'
581 f'irodori_api_sway = "{current_settings["irodori_api_sway"]}"\n'
582 f'irodori_api_sway_coeff = "{current_settings["irodori_api_sway_coeff"]}"\n'
583 f'\n'
584 f'[gemini]\n'
585 f'gemini_model = "{current_settings["gemini_model"]}"\n'
586 f'\n'
587 f'[edge]\n'
588 f'edge_rate = "{current_settings["edge_rate"]}"\n'
589 f'edge_volume = "{current_settings["edge_volume"]}"\n'
590 f'edge_pitch = "{current_settings["edge_pitch"]}"\n'
591 )
592
593 try:
594 with open(INI_FILE_NAME, 'w', encoding='utf-8') as f:
595 f.write(content)
596 except Exception as e:
597 print(f"設定ファイル保存エラー: {e}")
598
599
600 def cleanup_temp_files(self):
601 # クリーンアップロジックは省略せずに残す
602 self.handle_stop()
603 self.player.setSource(QUrl())
604 if self.worker_thread and self.worker_thread.isRunning():
605 self.worker_thread.quit()
606 self.worker_thread.wait()
607
608 for f in self.tmp_files:
609 try:
610 if os.path.isfile(f):
611 os.remove(f)
612 except Exception as e:
613 print(f"一時ファイル削除エラー: {f}: {e}")
614
615 temp_dir = self.settings.get('temp_dir', DEFAULT_TEMP_DIR)
616 try:
617 if os.path.exists(temp_dir):
618 search_pattern = os.path.join(temp_dir, f"{TEMP_WAV_PREFIX}*")
619 for file_path in glob.glob(search_pattern):
620 try:
621 if os.path.isfile(file_path):
622 os.remove(file_path)
623 except:
624 pass
625 if not os.listdir(temp_dir):
626 os.rmdir(temp_dir)
627 except Exception as e:
628 print(f"一時ディレクトリのクリーンアップエラー: {e}")
629
630
631 # --- UI初期化 ---
632 def initUI(self):
633 main_layout = QVBoxLayout()
634 self.tabs = QTabWidget()
635
636 # --- TTS設定ウィジェットの事前初期化 ---
637 # これらのウィジェットはMain/Configタブ間で共有されるため、先に初期化する
638 self.engine_combo = QComboBox(self)
639 self.engine_combo.addItems([
640 "pyttsx3",
641 "winrt",
642 "voicevox",
643 "qwen3",
644 "qwen_api",
645 "irodori",
646 "irodori_api",
647 "openai",
648 "elevenlabs",
649 "aquestalkplayer",
650 "gemini",
651 "edge"
652 ])
653 self.engine_combo.setCurrentText(DEFAULT_ENGINE)
654 self.voice_combo = QComboBox(self)
655 self.speed_spin = QDoubleSpinBox(self)
656 self.speed_spin.setRange(0.1, 5.0)
657 self.speed_spin.setSingleStep(0.1)
658 self.speed_spin.setValue(1.0)
659 self.pitch_spin = QDoubleSpinBox(self)
660 self.pitch_spin.setRange(-10.0, 10.0)
661 self.pitch_spin.setSingleStep(0.1)
662 self.pitch_spin.setValue(0.0)
663 self.instruction_line = QLineEdit()
664 self.instruction_line.setPlaceholderText("OpenAI/Gemini instruction (Qwen/Irodoriは下の専用設定を使用)")
665
666 # --- Tab 1: Main Content (メイン操作) ---
667 main_page = QWidget()
668 main_page_layout = QVBoxLayout(main_page)
669
670 # 1. ファイル設定
671 file_slide_layout = QGridLayout()
672 # Input File
673 file_slide_layout.addWidget(QLabel("Input File (infile):"), 0, 0)
674 self.input_file_line = QLineEdit(self.settings.get('input_file'))
675 file_slide_layout.addWidget(self.input_file_line, 0, 1)
676 self.input_file_btn = QPushButton("Path")
677 self.input_file_btn.clicked.connect(self.select_input_file)
678 file_slide_layout.addWidget(self.input_file_btn, 0, 2)
679 # Replace INI 1 (Default)
680 file_slide_layout.addWidget(QLabel("Replace INI (Default):"), 1, 0)
681 self.replace_ini_line = QLineEdit(self.settings.get('replace_file'))
682 file_slide_layout.addWidget(self.replace_ini_line, 1, 1)
683 self.replace_ini_btn = QPushButton("Path")
684 self.replace_ini_btn.clicked.connect(lambda: self.select_ini_file(self.replace_ini_line, 'replace_file'))
685 file_slide_layout.addWidget(self.replace_ini_btn, 1, 2)
686 # Replace INI 2 (User)
687 file_slide_layout.addWidget(QLabel("Replace INI (User):"), 2, 0)
688 self.replace_ini2_line = QLineEdit(self.settings.get('replace_file2'))
689 file_slide_layout.addWidget(self.replace_ini2_line, 2, 1)
690 self.replace_ini2_btn = QPushButton("Path")
691 self.replace_ini2_btn.clicked.connect(lambda: self.select_ini_file(self.replace_ini2_line, 'replace_file2'))
692 file_slide_layout.addWidget(self.replace_ini2_btn, 2, 2)
693 # Output File
694 file_slide_layout.addWidget(QLabel("出力ファイル (wav/mp3):"), 3, 0)
695 self.output_line = QLineEdit(self)
696 self.output_line.setPlaceholderText("未指定の場合、一時ファイルを作成して再生します (推奨)")
697 self.output_line.setText(self.settings.get('output_file', ''))
698 file_slide_layout.addWidget(self.output_line, 3, 1)
699 self.output_btn = QPushButton("Path")
700 self.output_btn.clicked.connect(self.select_output_file)
701 file_slide_layout.addWidget(self.output_btn, 3, 2)
702 main_page_layout.addLayout(file_slide_layout)
703
704 # Slide Page Pulldown (位置変更)
705 slide_page_layout = QHBoxLayout()
706 slide_page_layout.addWidget(QLabel("Slide Page:"))
707 self.slide_page_combo = QComboBox(self)
708 self.slide_page_combo.addItem("1. No file loaded")
709 self.slide_page_combo.setCurrentIndex(0)
710 self.slide_page_combo.currentIndexChanged.connect(self.on_slide_page_changed)
711 slide_page_layout.addWidget(self.slide_page_combo)
712 main_page_layout.addLayout(slide_page_layout)
713
714 # 4. テキスト入力エリア (2分割 & 拡張可能に)
715 text_layout = QVBoxLayout()
716
717 # Original Text
718 text_layout.addWidget(QLabel("読み上げテキスト (Original):"))
719 self.text_input_original = QTextEdit(self)
720 self.text_input_original.setPlaceholderText("入力ファイルの内容(スライドページ)がここに表示されます。")
721 self.text_input_original.setMinimumHeight(80)
722 self.text_input_original.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Expanding)
723 text_layout.addWidget(self.text_input_original)
724
725 # Converted Text
726 text_layout.addWidget(QLabel("読み上げテキスト (Converted):"))
727 self.text_input_converted = QTextEdit(self)
728 self.text_input_converted.setPlaceholderText("置換ルール適用後のテキストがここに表示されます。Play/Generateボタンはこのテキストを読み上げます。")
729 self.text_input_converted.setMinimumHeight(80)
730 self.text_input_converted.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Expanding)
731 text_layout.addWidget(self.text_input_converted)
732 main_page_layout.addLayout(text_layout)
733
734 # 5. コントロール (Convertedテキストの直下に配置)
735 control_layout = QHBoxLayout()
736
737 # Convert Button
738 self.convert_btn = QPushButton("⚙️ Convert (Apply Rules)")
739 self.convert_btn.clicked.connect(self.handle_convert)
740 self.convert_btn.setStyleSheet("font-weight: bold; padding: 5px;")
741 control_layout.addWidget(self.convert_btn)
742
743 # 再生/生成ボタン
744 self.play_btn = QPushButton("▶ Play/Generate")
745 self.generate_btn = QPushButton("⚡ Generate (Force)")
746 self.pause_btn = QPushButton("⏸ Pause")
747 self.stop_btn = QPushButton("■ Stop")
748
749 self.play_btn.clicked.connect(lambda: self.handle_play(force_generate=False))
750 self.generate_btn.clicked.connect(lambda: self.handle_play(force_generate=True)) # 強制生成
751 self.pause_btn.clicked.connect(self.handle_pause)
752 self.stop_btn.clicked.connect(self.handle_stop)
753
754 control_layout.addWidget(self.play_btn)
755 control_layout.addWidget(self.generate_btn)
756 control_layout.addWidget(self.pause_btn)
757 control_layout.addWidget(self.stop_btn)
758
759 main_page_layout.addLayout(control_layout)
760
761
762 # --- Tab 2: Config (パス設定) ---
763 config_page = QWidget()
764 config_page_layout = QVBoxLayout(config_page)
765
766 # アプリケーションパス設定
767 app_path_layout = QGridLayout()
768 # AquesTalk Path
769 app_path_layout.addWidget(QLabel("AquesTalk Path:"), 0, 0)
770 self.aquestalk_path_line = QLineEdit(self.settings.get('aquestalk_path', DEFAULT_AQUESTALK_PATH))
771 app_path_layout.addWidget(self.aquestalk_path_line, 0, 1)
772 self.aquestalk_path_btn = QPushButton("Path")
773 self.aquestalk_path_btn.clicked.connect(self.select_aquestalk_path)
774 app_path_layout.addWidget(self.aquestalk_path_btn, 0, 2)
775 # Voicevox Endpoint
776 app_path_layout.addWidget(QLabel("Voicevox Endpoint:"), 1, 0)
777 self.voicevox_endpoint_line = QLineEdit(self.settings.get('voicevox_endpoint', DEFAULT_VOICEVOX_ENDPOINT))
778 app_path_layout.addWidget(self.voicevox_endpoint_line, 1, 1, 1, 2)
779 # Temp Dir
780 app_path_layout.addWidget(QLabel("Temp Dir:"), 2, 0)
781 self.temp_dir_line = QLineEdit(self.settings.get('temp_dir', DEFAULT_TEMP_DIR))
782 app_path_layout.addWidget(self.temp_dir_line, 2, 1, 1, 2)
783
784 config_page_layout.addLayout(app_path_layout)
785
786 # TTS設定(Configタブに配置)
787 tts_settings_layout_config = QGridLayout()
788 tts_settings_layout_config.addWidget(QLabel("Engine:"), 3, 0)
789 tts_settings_layout_config.addWidget(self.engine_combo, 3, 1)
790 tts_settings_layout_config.addWidget(QLabel("Voice:"), 3, 2)
791 tts_settings_layout_config.addWidget(self.voice_combo, 3, 3)
792 tts_settings_layout_config.addWidget(QLabel("Speed (fspeak_rate):"), 4, 0)
793 tts_settings_layout_config.addWidget(self.speed_spin, 4, 1)
794 tts_settings_layout_config.addWidget(QLabel("Pitch (ピッチ):"), 4, 2)
795 tts_settings_layout_config.addWidget(self.pitch_spin, 4, 3)
796 tts_settings_layout_config.addWidget(QLabel("Instruction (OpenAI/Gemini):"), 5, 0)
797 tts_settings_layout_config.addWidget(self.instruction_line, 5, 1, 1, 3)
798
799 config_page_layout.addLayout(tts_settings_layout_config)
800
801 # Gemini TTS settings. API キーは INI に保存せず GEMINI_API_KEY を使用する。
802 gemini_layout = QGridLayout()
803 gemini_layout.addWidget(QLabel("Gemini TTS model:"), 0, 0)
804 self.gemini_model_line = QLineEdit(
805 self.settings.get("gemini_model", "gemini-3.8-flash-tts")
806 )
807 gemini_layout.addWidget(self.gemini_model_line, 0, 1, 1, 3)
808 gemini_layout.addWidget(QLabel("API key:"), 1, 0)
809 gemini_layout.addWidget(QLabel("GEMINI_API_KEY environment variable"), 1, 1, 1, 3)
810 self.gemini_frame = QFrame()
811 self.gemini_frame.setFrameShape(QFrame.Shape.StyledPanel)
812 self.gemini_frame.setLayout(gemini_layout)
813 config_page_layout.addWidget(QLabel("Gemini Flash TTS"))
814 config_page_layout.addWidget(self.gemini_frame)
815
816 # Edge TTS settings. No API key is required, but Edge TTS uses the
817 # Microsoft online service when the voice list or audio is requested.
818 edge_layout = QGridLayout()
819 edge_layout.addWidget(QLabel("Rate:"), 0, 0)
820 self.edge_rate_line = QLineEdit(self.settings.get("edge_rate", "+0%"))
821 self.edge_rate_line.setPlaceholderText("+0% (e.g. +20%, -10%)")
822 edge_layout.addWidget(self.edge_rate_line, 0, 1)
823 edge_layout.addWidget(QLabel("Volume:"), 0, 2)
824 self.edge_volume_line = QLineEdit(self.settings.get("edge_volume", "+0%"))
825 self.edge_volume_line.setPlaceholderText("+0%")
826 edge_layout.addWidget(self.edge_volume_line, 0, 3)
827 edge_layout.addWidget(QLabel("Pitch:"), 1, 0)
828 self.edge_pitch_line = QLineEdit(self.settings.get("edge_pitch", "+0Hz"))
829 self.edge_pitch_line.setPlaceholderText("+0Hz (e.g. +5Hz, -5Hz)")
830 edge_layout.addWidget(self.edge_pitch_line, 1, 1)
831 edge_layout.addWidget(QLabel("Internet connection required; no API key."), 1, 2, 1, 2)
832 self.edge_frame = QFrame()
833 self.edge_frame.setFrameShape(QFrame.Shape.StyledPanel)
834 self.edge_frame.setLayout(edge_layout)
835 config_page_layout.addWidget(QLabel("Edge TTS"))
836 config_page_layout.addWidget(self.edge_frame)
837
838 # Qwen3-TTS settings
839 qwen_layout = QGridLayout()
840 qwen_layout.addWidget(QLabel("Qwen3 model:"), 0, 0)
841 self.qwen3_model_line = QLineEdit(self.settings.get("qwen3_model_id", "Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice"))
842 qwen_layout.addWidget(self.qwen3_model_line, 0, 1, 1, 3)
843
844 qwen_layout.addWidget(QLabel("Language:"), 1, 0)
845 self.qwen3_language_line = QLineEdit(self.settings.get("qwen3_language", "Japanese"))
846 qwen_layout.addWidget(self.qwen3_language_line, 1, 1)
847
848 qwen_layout.addWidget(QLabel("Device:"), 1, 2)
849 self.qwen3_device_combo = QComboBox()
850 self.qwen3_device_combo.addItems(["auto", "cuda:0", "cpu"])
851 self.qwen3_device_combo.setCurrentText(self.settings.get("qwen3_device", "auto"))
852 qwen_layout.addWidget(self.qwen3_device_combo, 1, 3)
853
854 qwen_layout.addWidget(QLabel("dtype:"), 2, 0)
855 self.qwen3_dtype_combo = QComboBox()
856 self.qwen3_dtype_combo.addItems(["auto", "bfloat16", "float16", "float32"])
857 self.qwen3_dtype_combo.setCurrentText(self.settings.get("qwen3_dtype", "auto"))
858 qwen_layout.addWidget(self.qwen3_dtype_combo, 2, 1)
859
860 qwen_layout.addWidget(QLabel("Qwen instruct:"), 2, 2)
861 self.qwen3_instruct_line = QLineEdit(self.settings.get("qwen3_instruct", ""))
862 qwen_layout.addWidget(self.qwen3_instruct_line, 2, 3)
863
864 qwen_layout.addWidget(QLabel("Qwen API server:"), 3, 0)
865 self.qwen_api_server_line = QLineEdit(
866 self.settings.get("qwen_api_server", os.environ.get("QWEN_TTS_API", ""))
867 )
868 self.qwen_api_server_line.setPlaceholderText("http://GPU-server:port")
869 qwen_layout.addWidget(self.qwen_api_server_line, 3, 1, 1, 3)
870
871 qwen_layout.addWidget(QLabel("API timeout (sec):"), 4, 0)
872 self.qwen_api_timeout_line = QLineEdit(self.settings.get("qwen_api_timeout", "1800"))
873 qwen_layout.addWidget(self.qwen_api_timeout_line, 4, 1)
874
875 self.qwen_frame = QFrame()
876 self.qwen_frame.setFrameShape(QFrame.Shape.StyledPanel)
877 self.qwen_frame.setLayout(qwen_layout)
878 config_page_layout.addWidget(QLabel("Qwen3-TTS"))
879 config_page_layout.addWidget(self.qwen_frame)
880
881 # Irodori-TTS settings
882 irodori_layout = QGridLayout()
883 irodori_layout.addWidget(QLabel("Irodori model:"), 0, 0)
884 self.irodori_model_line = QLineEdit(self.settings.get("irodori_model_id", "Aratako/Irodori-TTS-v4.1-Small"))
885 irodori_layout.addWidget(self.irodori_model_line, 0, 1, 1, 3)
886
887 irodori_layout.addWidget(QLabel("Caption:"), 1, 0)
888 self.irodori_caption_line = QLineEdit(self.settings.get("irodori_caption", "落ち着いた自然な声で、明瞭に読み上げる。"))
889 irodori_layout.addWidget(self.irodori_caption_line, 1, 1, 1, 3)
890
891 irodori_layout.addWidget(QLabel("Reference WAV:"), 2, 0)
892 self.irodori_ref_wav_line = QLineEdit(self.settings.get("irodori_ref_wav", ""))
893 irodori_layout.addWidget(self.irodori_ref_wav_line, 2, 1, 1, 2)
894 self.irodori_ref_wav_btn = QPushButton("Path")
895 self.irodori_ref_wav_btn.clicked.connect(self.select_irodori_ref_wav)
896 irodori_layout.addWidget(self.irodori_ref_wav_btn, 2, 3)
897
898 irodori_layout.addWidget(QLabel("Device:"), 3, 0)
899 self.irodori_device_combo = QComboBox()
900 self.irodori_device_combo.addItems(["auto", "cuda", "cuda:0", "cpu"])
901 self.irodori_device_combo.setCurrentText(self.settings.get("irodori_device", "auto"))
902 irodori_layout.addWidget(self.irodori_device_combo, 3, 1)
903
904 irodori_layout.addWidget(QLabel("Precision:"), 3, 2)
905 self.irodori_precision_combo = QComboBox()
906 self.irodori_precision_combo.addItems(["auto", "bf16", "fp32"])
907 self.irodori_precision_combo.setCurrentText(self.settings.get("irodori_precision", "auto"))
908 irodori_layout.addWidget(self.irodori_precision_combo, 3, 3)
909
910 irodori_layout.addWidget(QLabel("Steps:"), 4, 0)
911 self.irodori_steps_spin = QDoubleSpinBox()
912 self.irodori_steps_spin.setDecimals(0)
913 self.irodori_steps_spin.setRange(1, 200)
914 self.irodori_steps_spin.setSingleStep(1)
915 self.irodori_steps_spin.setValue(float(self.settings.get("irodori_num_steps", "40")))
916 irodori_layout.addWidget(self.irodori_steps_spin, 4, 1)
917
918 irodori_layout.addWidget(QLabel("Duration scale:"), 4, 2)
919 self.irodori_duration_spin = QDoubleSpinBox()
920 self.irodori_duration_spin.setRange(0.1, 3.0)
921 self.irodori_duration_spin.setSingleStep(0.05)
922 self.irodori_duration_spin.setValue(float(self.settings.get("irodori_duration_scale", "1.0")))
923 irodori_layout.addWidget(self.irodori_duration_spin, 4, 3)
924
925 irodori_layout.addWidget(QLabel("Seed:"), 5, 0)
926 self.irodori_seed_line = QLineEdit(self.settings.get("irodori_seed", "0"))
927 self.irodori_seed_line.setPlaceholderText("0 / random / none")
928 irodori_layout.addWidget(self.irodori_seed_line, 5, 1)
929
930 irodori_layout.addWidget(QLabel("API server:"), 6, 0)
931 self.irodori_api_server_line = QLineEdit(
932 self.settings.get("irodori_api_server", DEFAULT_IRODORI_API_SERVER)
933 )
934 self.irodori_api_server_line.setPlaceholderText(DEFAULT_IRODORI_API_SERVER)
935 irodori_layout.addWidget(self.irodori_api_server_line, 6, 1, 1, 3)
936
937 irodori_layout.addWidget(QLabel("API timeout (sec):"), 7, 0)
938 self.irodori_api_timeout_line = QLineEdit(
939 self.settings.get("irodori_api_timeout", "1800")
940 )
941 irodori_layout.addWidget(self.irodori_api_timeout_line, 7, 1)
942
943 self.irodori_api_sway_check = QCheckBox("sway sampling")
944 self.irodori_api_sway_check.setChecked(
945 self.settings.get("irodori_api_sway", "0") in ("1", "true", "True")
946 )
947 irodori_layout.addWidget(self.irodori_api_sway_check, 7, 2)
948
949 self.irodori_api_sway_coeff_spin = QDoubleSpinBox()
950 self.irodori_api_sway_coeff_spin.setRange(-10.0, 10.0)
951 self.irodori_api_sway_coeff_spin.setSingleStep(0.1)
952 self.irodori_api_sway_coeff_spin.setValue(
953 float(self.settings.get("irodori_api_sway_coeff", "-1.0"))
954 )
955 irodori_layout.addWidget(self.irodori_api_sway_coeff_spin, 7, 3)
956
957 self.irodori_frame = QFrame()
958 self.irodori_frame.setFrameShape(QFrame.Shape.StyledPanel)
959 self.irodori_frame.setLayout(irodori_layout)
960 config_page_layout.addWidget(QLabel("Irodori-TTS"))
961 config_page_layout.addWidget(self.irodori_frame)
962
963 config_page_layout.addStretch(1) # 残りのスペースを埋める
964
965 # Tab Widgetに追加
966 self.tabs.addTab(main_page, "Main")
967 self.tabs.addTab(config_page, "Config")
968 main_layout.addWidget(self.tabs)
969
970 # --- Tabの外の共通コントロール ---
971
972 # 6. プログレスバーと再生位置を1行に統合
973 progress_slider_layout = QHBoxLayout()
974 progress_slider_layout.addWidget(QLabel("Prog/Pos:"))
975
976 # プログレスバー
977 self.progress_bar = QProgressBar(self)
978 self.progress_bar.setRange(0, 100)
979 self.progress_bar.setValue(0)
980 self.progress_bar.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Preferred)
981 progress_slider_layout.addWidget(self.progress_bar)
982
983 # 再生位置スライダー
984 self.position_slider = QSlider(Qt.Orientation.Horizontal)
985 self.position_slider.setRange(0, 0)
986 self.position_slider.setTracking(False)
987 self.position_slider.sliderMoved.connect(self.seek_position)
988 self.position_slider.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Preferred)
989 progress_slider_layout.addWidget(self.position_slider)
990
991 main_layout.addLayout(progress_slider_layout)
992
993 # 7. ステータスラベル
994 self.status_label = QLabel("Status: Ready")
995 main_layout.addWidget(self.status_label)
996
997 self.setLayout(main_layout)
998 self.update_playback_buttons(self.player.playbackState())
999
1000 # 初期状態では選択中エンジン以外の専用設定を無効化
1001 self.qwen_frame.setEnabled(
1002 self.engine_combo.currentText().lower() in ("qwen3", "qwen", "qwen_api", "qwen-api")
1003 )
1004 self.irodori_frame.setEnabled(self.engine_combo.currentText().lower() in ("irodori", "irodori-tts"))
1005 self.gemini_frame.setEnabled(self.engine_combo.currentText().lower() == "gemini")
1006 self.edge_frame.setEnabled(self.engine_combo.currentText().lower() in ("edge", "edge_tts", "edge-tts"))
1007
1008 self.text_input_original.setText("TTSアプリへようこそ!\nInput Fileを選択すると、内容がここに表示され、Convertボタンで置換が適用されます。")
1009 self.text_input_converted.setText("Play/Generateボタンを押すと、このconvertedテキストが読み上げられます。")
1010
1011
1012 # --- ファイル選択/スロット群 ---
1013
1014 def _get_initial_dir(self, path_line: QLineEdit, key: str) -> str:
1015 """QLineEditの内容または記憶されたディレクトリを元に初期ディレクトリを返す"""
1016 current_path = path_line.text()
1017 if os.path.isfile(current_path):
1018 dir_name = os.path.dirname(current_path)
1019 self.last_dir[key] = dir_name
1020 return dir_name
1021 elif os.path.isdir(current_path):
1022 self.last_dir[key] = current_path
1023 return current_path
1024 elif key in self.last_dir and os.path.isdir(self.last_dir[key]):
1025 return self.last_dir[key]
1026 return QDir.currentPath()
1027
1028 @Slot()
1029 def select_input_file(self):
1030 """Input Fileを選択し、ファイルを読み込み、スライドをパースする"""
1031 key = 'input_file'
1032 initial_dir = self._get_initial_dir(self.input_file_line, key)
1033 file_path, _ = QFileDialog.getOpenFileName(
1034 self,
1035 "入力テキストファイルの選択",
1036 initial_dir,
1037 "テキストファイル (*.txt *.md);;全てのファイル (*)"
1038 )
1039 if file_path:
1040 self.input_file_line.setText(file_path)
1041 self.last_dir[key] = os.path.dirname(file_path)
1042 self.load_input_file(file_path)
1043
1044 @Slot()
1045 def select_output_file(self):
1046 """出力音声ファイルを選択し、パスを更新する"""
1047 key = 'output_file'
1048 initial_dir = self._get_initial_dir(self.output_line, key)
1049 file_path, _ = QFileDialog.getSaveFileName(
1050 self,
1051 "出力音声ファイルの選択",
1052 initial_dir,
1053 "Audio Files (*.wav *.mp3);;Wave Files (*.wav);;MP3 Files (*.mp3);;All Files (*)"
1054 )
1055 if file_path:
1056 root, ext = os.path.splitext(file_path)
1057 if not ext:
1058 engine = self.engine_combo.currentText().lower()
1059 default_ext = ".mp3" if engine in ("openai", "elevenlabs", "eleven", "edge", "edge_tts", "edge-tts") else ".wav"
1060 file_path += default_ext
1061 self.output_line.setText(file_path)
1062 self.last_dir[key] = os.path.dirname(file_path)
1063
1064 @Slot()
1065 def select_aquestalk_path(self):
1066 """AquesTalkPlayer.exeのファイルパスを選択するダイアログを開く"""
1067 key = 'aquestalk_path'
1068 initial_dir = self._get_initial_dir(self.aquestalk_path_line, key)
1069 file_path, _ = QFileDialog.getOpenFileName(
1070 self,
1071 "AquesTalkPlayer.exeの選択",
1072 initial_dir,
1073 "実行ファイル (*.exe);;全てのファイル (*)"
1074 )
1075 if file_path:
1076 self.aquestalk_path_line.setText(file_path)
1077 self.last_dir[key] = os.path.dirname(file_path)
1078
1079 @Slot()
1080 def select_irodori_ref_wav(self):
1081 key = "irodori_ref_wav"
1082 initial_dir = self._get_initial_dir(self.irodori_ref_wav_line, key)
1083 file_path, _ = QFileDialog.getOpenFileName(
1084 self,
1085 "Irodori-TTS Reference WAV",
1086 initial_dir,
1087 "Wave Files (*.wav);;Audio Files (*.wav *.mp3 *.flac);;All Files (*)"
1088 )
1089 if file_path:
1090 self.irodori_ref_wav_line.setText(file_path)
1091 self.last_dir[key] = os.path.dirname(file_path)
1092
1093 @Slot(QLineEdit, str)
1094 def select_ini_file(self, line_edit: QLineEdit, key: str):
1095 """INIファイルを選択するダイアログを開く"""
1096 initial_dir = self._get_initial_dir(line_edit, key)
1097 file_path, _ = QFileDialog.getOpenFileName(
1098 self,
1099 "INIファイル(置換ルール)の選択",
1100 initial_dir,
1101 "INIファイル (*.ini);;全てのファイル (*)"
1102 )
1103 if file_path:
1104 line_edit.setText(file_path)
1105 self.last_dir[key] = os.path.dirname(file_path)
1106
1107 def load_input_file(self, file_path: str):
1108 self.slide_data = {}
1109 self.current_infile_path = file_path
1110
1111 try:
1112 encoding = detect_encoding(file_path)
1113 if encoding is None: encoding = 'utf-8'
1114 with open(file_path, 'r', encoding=encoding) as f:
1115 full_text = f.read()
1116
1117 self.slide_data[0] = full_text.strip()
1118 current_slide_number = 1
1119
1120 # スライド区切りパターン: `# Slide` で始まる行、または `*1, *2` などの行
1121 slide_separator_pattern = re.compile(r"^\s*(#\s*Slide.*|\*\d+)\s*$", re.IGNORECASE | re.MULTILINE)
1122
1123 parts = slide_separator_pattern.split(full_text)
1124
1125 if len(parts) > 1:
1126
1127 # parts[0]は最初の区切りより前のテキスト
1128 if parts[0].strip():
1129 self.slide_data[1] = parts[0].strip()
1130 current_slide_number = 2
1131
1132 # parts[1]以降は区切りと区切りの間のテキスト
1133 for part in parts[1:]:
1134 part_content = part.strip()
1135 # 区切りパターンにマッチしたテキスト(空か、区切り文字自体)はスキップ
1136 if part_content and not slide_separator_pattern.match(part_content):
1137 self.slide_data[current_slide_number] = part_content
1138 current_slide_number += 1
1139
1140 # スライドが分割された場合は、改めてページ0(全文)を再構築
1141 self.slide_data[0] = full_text.strip()
1142
1143
1144 # 3. Slide pageプルダウンを更新
1145 self.slide_page_combo.clear()
1146
1147 is_slide_parsed = len(self.slide_data) > 1
1148
1149 self.slide_page_combo.addItem("0. (All Document)")
1150 if is_slide_parsed:
1151 for i in sorted([k for k in self.slide_data.keys() if k > 0]):
1152 self.slide_page_combo.addItem(f"{i}. Slide {i}")
1153 self.slide_page_combo.setEnabled(True)
1154 else:
1155 self.slide_page_combo.setEnabled(False)
1156
1157 self.slide_page_combo.setCurrentIndex(0)
1158
1159 except Exception as e:
1160 QMessageBox.critical(self, "ファイル読み込みエラー", f"ファイルの読み込みに失敗しました: {e}")
1161 self.slide_page_combo.clear()
1162 self.slide_page_combo.addItem("1. No file loaded")
1163 self.slide_page_combo.setEnabled(False)
1164 self.text_input_original.setText("")
1165 self.text_input_converted.setText("")
1166
1167 @Slot(int)
1168 def on_slide_page_changed(self, index):
1169 page_key = index
1170
1171 if page_key in self.slide_data:
1172 original_text = self.slide_data[page_key]
1173
1174 # Convertedテキストの変更を一時的に無効化
1175 self.text_input_converted.textChanged.disconnect(self.set_dirty)
1176
1177 self.text_input_original.setText(original_text)
1178 self.text_input_converted.setText("")
1179 self.is_converted_text_dirty = True # ページが変わったので変換が必要
1180
1181 # Convertedテキストの変更監視を再開
1182 self.text_input_converted.textChanged.connect(self.set_dirty)
1183
1184 self.status_label.setText(f"Status: Page {page_key} loaded. Ready to Convert.")
1185 else:
1186 self.text_input_original.setText("")
1187 self.text_input_converted.setText("")
1188 self.status_label.setText("Status: Error - Page content missing.")
1189
1190 @Slot()
1191 def handle_convert(self):
1192 """Convertボタンが押されたとき、置換処理を実行する"""
1193 original_text = self.text_input_original.toPlainText()
1194 if not original_text.strip():
1195 QMessageBox.warning(self, "Convertエラー", "Originalテキストが空です。ファイルを読み込むか、テキストを入力してください。")
1196 return
1197
1198 default_ini_path = self.replace_ini_line.text()
1199 user_ini_path = self.replace_ini2_line.text()
1200
1201 self.status_label.setText("Status: Loading/Checking replacement rules...")
1202 QApplication.processEvents()
1203
1204 # タイムスタンプチェックと再読み込み (force_reload=True)
1205 dict2 = load_replace_dict(user_ini_path, force_reload=True)
1206 dict1 = load_replace_dict(default_ini_path, force_reload=True)
1207
1208 dict1_filtered = {k: v for k, v in dict1.items() if k not in dict2}
1209
1210 final_replace_dict = {}
1211 final_replace_dict.update(dict1_filtered)
1212 final_replace_dict.update(dict2)
1213
1214 if not final_replace_dict:
1215 QMessageBox.information(self, "Convert情報", "有効な置換ルールがINIファイルから見つかりませんでした。")
1216 self.text_input_converted.setText(original_text)
1217
1218 # Convertedテキストの変更を一時的に無効化
1219 self.text_input_converted.textChanged.disconnect(self.set_dirty)
1220 self.is_converted_text_dirty = False
1221 self.text_input_converted.textChanged.connect(self.set_dirty)
1222
1223 self.status_label.setText("Status: No rules applied. Converted = Original.")
1224 return
1225
1226 self.status_label.setText("Status: Applying replacement rules...")
1227 QApplication.processEvents()
1228
1229 replaced_text = apply_replacements(original_text, final_replace_dict)
1230
1231 # 修正: `# Slide...` および `*N` 形式のマーカー行を完全に削除
1232 # 1. `# Slide...` 行の削除(行頭/行末に空白があっても良い)
1233 replaced_text = re.sub(r"^\s*#\s*Slide.*$", "\n", replaced_text, flags=re.IGNORECASE | re.MULTILINE)
1234
1235 # 2. `*N` (スライド番号) 行の削除(行頭/行末に空白があっても良い)
1236 replaced_text = re.sub(r"^\s*\*\d+\s*$", "\n", replaced_text, flags=re.IGNORECASE | re.MULTILINE)
1237
1238 # 空行のみの行を削除(連続する空行を一つにまとめる)
1239 replaced_text = re.sub(r'\n\s*\n', '\n\n', replaced_text).strip()
1240
1241 # Convertedテキストの変更を一時的に無効化してから更新
1242 self.text_input_converted.textChanged.disconnect(self.set_dirty)
1243 self.text_input_converted.setText(replaced_text)
1244 self.is_converted_text_dirty = False # 変換完了
1245 self.text_input_converted.textChanged.connect(self.set_dirty)
1246
1247 self.status_label.setText("Status: Conversion complete. Ready to Play.")
1248
1249
1250 def handle_play(self, force_generate: bool = False):
1251 """
1252 Play/Generateボタンの動作。
1253 force_generate=True の場合は、ダーティフラグにかかわらず強制的に生成。
1254 """
1255
1256 # 1. 既に再生中の場合は無視
1257 if self.player.playbackState() == QMediaPlayer.PlaybackState.PlayingState:
1258 return
1259
1260 # 2. ポーズ状態からの再開
1261 if self.player.playbackState() == QMediaPlayer.PlaybackState.PausedState:
1262 self.player.play()
1263 return
1264
1265 # 3. 強制生成が不要かつダーティでない場合 -> 再生
1266 is_ready_to_play = (
1267 not force_generate and
1268 not self.is_converted_text_dirty and
1269 self.current_audio_path and
1270 os.path.exists(self.current_audio_path)
1271 )
1272
1273 if is_ready_to_play:
1274 self.status_label.setText(f"Status: Playing existing audio: {os.path.basename(self.current_audio_path)}")
1275 try:
1276 audio_url = QUrl.fromLocalFile(self.current_audio_path)
1277 self.player.stop()
1278 self.player.setSource(audio_url)
1279 self.player.play()
1280 except Exception as e:
1281 QMessageBox.critical(self, "再生エラー", f"既存の音声ファイルの再生に失敗しました: {e}")
1282 self.current_audio_path = None
1283 return
1284
1285 # 4. 生成が必要な場合 (ダーティ or ファイルがない or 強制生成)
1286
1287 if self.worker_thread and self.worker_thread.isRunning():
1288 QMessageBox.warning(self, "処理中", "現在、音声生成が実行中です。完了をお待ちください。")
1289 return
1290
1291 text = self.text_input_converted.toPlainText()
1292 outfile = self.output_line.text().strip()
1293 tts_engine = self.engine_combo.currentText()
1294 speed_rate = self.speed_spin.value()
1295 voice_name = self.voice_combo.currentText()
1296 pitch = self.pitch_spin.value()
1297 instruction = self.instruction_line.text()
1298
1299 aquestalk_path = self.aquestalk_path_line.text().strip()
1300 temp_dir = self.temp_dir_line.text().strip()
1301 voicevox_endpoint = self.voicevox_endpoint_line.text().strip()
1302
1303 qwen3_language = self.qwen3_language_line.text().strip() or "Japanese"
1304 qwen3_model_id = self.qwen3_model_line.text().strip() or "Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice"
1305 qwen3_device = self.qwen3_device_combo.currentText()
1306 qwen3_dtype = self.qwen3_dtype_combo.currentText()
1307 qwen3_instruct = self.qwen3_instruct_line.text().strip()
1308 qwen_api_server = self.qwen_api_server_line.text().strip()
1309 try:
1310 qwen_api_timeout = float(self.qwen_api_timeout_line.text().strip() or "1800")
1311 except ValueError:
1312 QMessageBox.warning(self, "入力エラー", "Qwen API timeout は数値で指定してください。")
1313 return
1314
1315 irodori_caption = self.irodori_caption_line.text().strip()
1316 irodori_ref_wav = self.irodori_ref_wav_line.text().strip()
1317 irodori_model_id = self.irodori_model_line.text().strip() or "Aratako/Irodori-TTS-v4.1-Small"
1318 irodori_device = self.irodori_device_combo.currentText()
1319 irodori_precision = self.irodori_precision_combo.currentText()
1320 irodori_num_steps = int(self.irodori_steps_spin.value())
1321 irodori_duration_scale = float(self.irodori_duration_spin.value())
1322 irodori_seed = self.irodori_seed_line.text().strip() or "0"
1323 irodori_api_server = self.irodori_api_server_line.text().strip()
1324 try:
1325 irodori_api_timeout = float(self.irodori_api_timeout_line.text().strip() or "1800")
1326 except ValueError:
1327 QMessageBox.warning(self, "入力エラー", "Irodori API timeout は数値で指定してください。")
1328 return
1329 irodori_api_sway = self.irodori_api_sway_check.isChecked()
1330 irodori_api_sway_coeff = self.irodori_api_sway_coeff_spin.value()
1331 gemini_model = self.gemini_model_line.text().strip() or "gemini-3.8-flash-tts"
1332 edge_rate = self.edge_rate_line.text().strip() or "+0%"
1333 edge_volume = self.edge_volume_line.text().strip() or "+0%"
1334 edge_pitch = self.edge_pitch_line.text().strip() or "+0Hz"
1335
1336 if not text.strip():
1337 QMessageBox.warning(self, "入力エラー", "読み上げテキスト (Converted) が空です。")
1338 return
1339
1340 if tts_engine.lower() in ("qwen_api", "qwen-api") and not qwen_api_server:
1341 QMessageBox.warning(self, "入力エラー", "Qwen API server を指定してください。")
1342 return
1343 if tts_engine.lower() in ("irodori_api", "irodori-api") and not irodori_api_server:
1344 QMessageBox.warning(self, "入力エラー", "Irodori API server を指定してください。")
1345 return
1346 if tts_engine.lower() == "gemini" and not os.getenv("GEMINI_API_KEY"):
1347 QMessageBox.warning(self, "入力エラー", "GEMINI_API_KEY 環境変数を設定してください。")
1348 return
1349
1350 if not voice_name or "No voices found" in voice_name or "Error loading voices" in voice_name:
1351 QMessageBox.warning(self, "ボイス選択エラー", "有効なボイスが選択されていません。エンジンを確認してください。")
1352 return
1353
1354 self.status_label.setText("Status: 音声生成を開始しました... (非同期実行中)")
1355 self.update_playback_buttons(QMediaPlayer.PlaybackState.StoppedState, is_generating=True)
1356 self.progress_bar.setValue(0)
1357
1358 self.worker_thread = MyTTSWorker(
1359 text, outfile, tts_engine, speed_rate, pitch, instruction, voice_name, self.tmp_files,
1360 aquestalk_path, temp_dir, voicevox_endpoint,
1361 qwen3_language=qwen3_language,
1362 qwen3_model_id=qwen3_model_id,
1363 qwen3_device=qwen3_device,
1364 qwen3_dtype=qwen3_dtype,
1365 qwen3_instruct=qwen3_instruct,
1366 qwen_api_server=qwen_api_server,
1367 qwen_api_timeout=qwen_api_timeout,
1368 irodori_caption=irodori_caption,
1369 irodori_ref_wav=irodori_ref_wav,
1370 irodori_model_id=irodori_model_id,
1371 irodori_device=irodori_device,
1372 irodori_precision=irodori_precision,
1373 irodori_num_steps=irodori_num_steps,
1374 irodori_duration_scale=irodori_duration_scale,
1375 irodori_seed=irodori_seed,
1376 irodori_api_server=irodori_api_server,
1377 irodori_api_timeout=irodori_api_timeout,
1378 irodori_api_sway=irodori_api_sway,
1379 irodori_api_sway_coeff=irodori_api_sway_coeff,
1380 gemini_model=gemini_model,
1381 edge_rate=edge_rate,
1382 edge_volume=edge_volume,
1383 edge_pitch=edge_pitch,
1384 )
1385 self.worker_thread.finished.connect(self.on_synthesis_finished)
1386 self.worker_thread.error.connect(self.on_synthesis_error)
1387 self.worker_thread.progress.connect(self.progress_bar.setValue)
1388 self.worker_thread.start()
1389
1390
1391 @Slot(str)
1392 def on_synthesis_finished(self, file_path):
1393 self.current_audio_path = file_path
1394 self.is_converted_text_dirty = False
1395 self.status_label.setText(f"Status: 音声生成が完了しました: {os.path.basename(file_path)}")
1396 self.progress_bar.setValue(100)
1397
1398 self.update_playback_buttons(QMediaPlayer.PlaybackState.StoppedState, is_generating=False)
1399
1400 if not file_path or not os.path.exists(file_path):
1401 self.on_synthesis_error(f"音声ファイルが見つかりません: {file_path}")
1402 return
1403
1404 try:
1405 audio_url = QUrl.fromLocalFile(file_path)
1406 self.player.stop()
1407 self.player.setSource(audio_url)
1408 self.player.play()
1409 except Exception as e:
1410 self.on_synthesis_error(f"音声再生の開始に失敗しました: {e}")
1411
1412 @Slot(str)
1413 def on_synthesis_error(self, message):
1414 self.status_label.setText("Status: エラー発生")
1415 self.progress_bar.setValue(0)
1416 QMessageBox.critical(self, "エラー", message)
1417 self.current_audio_path = None
1418
1419 self.update_playback_buttons(QMediaPlayer.PlaybackState.StoppedState, is_generating=False)
1420 if self.worker_thread:
1421 self.worker_thread.quit()
1422 self.worker_thread.wait()
1423
1424 @Slot()
1425 def update_voice_list(self):
1426 """エンジンに応じて音声一覧と有効な設定項目を更新する。"""
1427
1428 engine = self.engine_combo.currentText()
1429 engine_lower = engine.lower()
1430 self.voice_combo.clear()
1431
1432 is_winrt = engine_lower in ("winrt", "onecore")
1433 is_voicevox = engine_lower == "voicevox"
1434 is_qwen_local = engine_lower in ("qwen3", "qwen")
1435 is_qwen_api = engine_lower in ("qwen_api", "qwen-api")
1436 is_qwen = is_qwen_local or is_qwen_api
1437 is_irodori_local = engine_lower in ("irodori", "irodori-tts")
1438 is_irodori_api = engine_lower in ("irodori_api", "irodori-api")
1439 is_irodori = is_irodori_local or is_irodori_api
1440 is_openai = engine_lower == "openai"
1441 is_elevenlabs = engine_lower in ("elevenlabs", "eleven")
1442 is_aqt = engine_lower in ("aquestalkplayer", "atp")
1443 is_gemini = engine_lower == "gemini"
1444 is_edge = engine_lower in ("edge", "edge_tts", "edge-tts")
1445
1446 # Pitchは現在 VoiceVox / AquesTalkPlayer のみ。
1447 self.pitch_spin.setEnabled(is_voicevox or is_aqt)
1448 # Speedは pyttsx3 / WinRT / ElevenLabsを含め、全エンジンで表示する。
1449 self.speed_spin.setEnabled(True)
1450 self.instruction_line.setEnabled(is_openai or is_gemini)
1451
1452 if hasattr(self, "qwen_frame"):
1453 self.qwen_frame.setEnabled(is_qwen)
1454 if hasattr(self, "qwen_api_server_line"):
1455 self.qwen_api_server_line.setEnabled(is_qwen_api)
1456 self.qwen_api_timeout_line.setEnabled(is_qwen_api)
1457 if hasattr(self, "irodori_frame"):
1458 self.irodori_frame.setEnabled(is_irodori)
1459 if hasattr(self, "gemini_frame"):
1460 self.gemini_frame.setEnabled(is_gemini)
1461 if hasattr(self, "edge_frame"):
1462 self.edge_frame.setEnabled(is_edge)
1463 if hasattr(self, "irodori_api_server_line"):
1464 self.irodori_api_server_line.setEnabled(is_irodori_api)
1465 self.irodori_api_timeout_line.setEnabled(is_irodori_api)
1466 self.irodori_api_sway_check.setEnabled(is_irodori_api)
1467 self.irodori_api_sway_coeff_spin.setEnabled(is_irodori_api)
1468 self.irodori_ref_wav_line.setEnabled(is_irodori_local)
1469 self.irodori_ref_wav_btn.setEnabled(is_irodori_local)
1470 self.irodori_model_line.setEnabled(is_irodori_local)
1471 self.irodori_device_combo.setEnabled(is_irodori_local)
1472 self.irodori_precision_combo.setEnabled(is_irodori_local)
1473 self.irodori_seed_line.setEnabled(is_irodori_local)
1474
1475 # Qwen/Irodori は speed/pitch の直接制御を行わない
1476 if is_qwen or is_irodori:
1477 self.speed_spin.setEnabled(False)
1478 self.pitch_spin.setEnabled(False)
1479
1480 if hasattr(self, "voicevox_endpoint_line"):
1481 self.voicevox_endpoint_line.setEnabled(is_voicevox)
1482 self.aquestalk_path_line.setEnabled(is_aqt)
1483
1484 endpoint = None
1485 if is_voicevox and hasattr(self, "voicevox_endpoint_line"):
1486 endpoint = self.voicevox_endpoint_line.text().strip()
1487 elif is_qwen_api and hasattr(self, "qwen_api_server_line"):
1488 endpoint = self.qwen_api_server_line.text().strip()
1489 elif is_irodori_api and hasattr(self, "irodori_api_server_line"):
1490 endpoint = self.irodori_api_server_line.text().strip()
1491
1492 api_key = os.getenv("ELEVENLABS_API_KEY") if is_elevenlabs else None
1493
1494 try:
1495 voices = tktts.get_available_voices(
1496 engine_lower,
1497 endpoint=endpoint,
1498 api_key=api_key,
1499 )
1500
1501 if voices:
1502 self.voice_combo.addItems([str(v) for v in voices])
1503
1504 default_voice = None
1505 if engine_lower == "pyttsx3":
1506 default_voice = getattr(tktts, "default_pyttsx3_voice", None)
1507 elif is_winrt:
1508 default_voice = getattr(tktts, "default_winrt_voice", None)
1509 elif is_openai:
1510 default_voice = getattr(
1511 tktts,
1512 "default_openai_voice",
1513 getattr(tktts, "default_optnai_voice", None),
1514 )
1515 elif is_elevenlabs:
1516 default_voice = getattr(tktts, "default_elevenlabs_voice", None)
1517 elif is_voicevox:
1518 default_voice = getattr(tktts, "default_voicevox_voice", None)
1519 elif is_qwen_api:
1520 default_voice = getattr(tktts, "default_qwen_api_voice", "Ono_Anna")
1521 elif is_qwen_local:
1522 default_voice = getattr(tktts, "default_qwen3_voice", "Ono_Anna")
1523 elif is_irodori_api:
1524 default_voice = getattr(tktts, "default_irodori_api_voice", "none")
1525 elif is_irodori_local:
1526 default_voice = getattr(tktts, "default_irodori_voice", "default")
1527 elif is_aqt:
1528 default_voice = getattr(tktts, "default_aqt_preset", None)
1529 elif is_gemini:
1530 default_voice = getattr(tktts, "default_gemini_voice", "Kore")
1531 elif is_edge:
1532 default_voice = getattr(tktts, "default_edge_voice", "Nanami")
1533
1534 if default_voice and default_voice in voices:
1535 self.voice_combo.setCurrentText(default_voice)
1536 else:
1537 self.voice_combo.setCurrentIndex(0)
1538
1539 self.status_label.setText(
1540 f"Status: {engine} voices loaded ({len(voices)})."
1541 )
1542 else:
1543 if is_elevenlabs and not api_key:
1544 self.voice_combo.addItem("ELEVENLABS_API_KEY is not set")
1545 else:
1546 self.voice_combo.addItem(f"No voices found for {engine}")
1547
1548 except Exception as e:
1549 error_msg = f"Error loading voices for {engine}: {e}"
1550 self.voice_combo.addItem(error_msg)
1551 print(error_msg)
1552 traceback.print_exc()
1553
1554 # エンジン変更または音声一覧の再取得は、必ず再生成対象にする。
1555 self.set_dirty()
1556
1557 def update_playback_buttons(self, state: QMediaPlayer.PlaybackState, is_generating: Optional[bool] = None):
1558 if is_generating is None:
1559 is_worker_running = self.worker_thread and self.worker_thread.isRunning()
1560 else:
1561 is_worker_running = is_generating
1562
1563 self.convert_btn.setEnabled(not is_worker_running)
1564 self.generate_btn.setEnabled(not is_worker_running)
1565
1566 if is_worker_running:
1567 self.play_btn.setText("▶ Generating...")
1568 self.play_btn.setEnabled(False)
1569 self.pause_btn.setEnabled(False)
1570 self.stop_btn.setEnabled(False)
1571 return
1572
1573 if state == QMediaPlayer.PlaybackState.PlayingState:
1574 self.play_btn.setText("▶ Playing...")
1575 self.play_btn.setEnabled(False)
1576 self.pause_btn.setEnabled(True)
1577 self.stop_btn.setEnabled(True)
1578 elif state == QMediaPlayer.PlaybackState.PausedState:
1579 self.play_btn.setText("▶ Resume")
1580 self.play_btn.setEnabled(True)
1581 self.pause_btn.setEnabled(False)
1582 self.stop_btn.setEnabled(True)
1583 elif state == QMediaPlayer.PlaybackState.StoppedState:
1584 if self.is_converted_text_dirty:
1585 self.play_btn.setText("▶ Generate (Text changed)")
1586 elif self.current_audio_path:
1587 self.play_btn.setText("▶ Play (Existing)")
1588 else:
1589 self.play_btn.setText("▶ Play/Generate")
1590
1591 self.play_btn.setEnabled(True)
1592 self.pause_btn.setEnabled(False)
1593 self.stop_btn.setEnabled(False)
1594 if not is_worker_running and "Status: エラー" not in self.status_label.text():
1595 if self.current_audio_path and not self.is_converted_text_dirty:
1596 self.status_label.setText("Status: Generated (Ready to play)")
1597 else:
1598 self.status_label.setText("Status: Ready")
1599
1600 # --- メディア制御系は変更なし ---
1601 def handle_pause(self):
1602 if self.player.playbackState() == QMediaPlayer.PlaybackState.PlayingState:
1603 self.player.pause()
1604
1605 def handle_stop(self):
1606 self.player.stop()
1607 self.position_slider.setValue(0)
1608 self.status_label.setText("Status: 停止")
1609 self.update_playback_buttons(QMediaPlayer.PlaybackState.StoppedState)
1610
1611 def handle_media_player_error(self, error: QMediaPlayer.Error, error_string: str):
1612 if error != QMediaPlayer.Error.NoError:
1613 print(f"--- QMediaPlayer Error ---")
1614 print(f"Code: {error.name}, Message: {error_string}")
1615 print(f"--------------------------")
1616 self.status_label.setText(f"Status: 再生エラー発生 ({error_string})")
1617 self.player.stop()
1618 self.update_playback_buttons(QMediaPlayer.PlaybackState.StoppedState)
1619
1620 def set_slider_range(self, duration):
1621 self.position_slider.setRange(0, duration)
1622
1623 def update_slider(self, position):
1624 self.position_slider.setValue(position)
1625 total_duration = self.player.duration()
1626 if total_duration > 0:
1627 current_sec = position // 1000
1628 total_sec = total_duration // 1000
1629 self.status_label.setText(f"Status: Playing ({current_sec} sec / {total_sec} sec)")
1630
1631 def seek_position(self, position):
1632 self.player.setPosition(position)
1633
1634
1635if __name__ == '__main__':
1636 app = QApplication(sys.argv)
1637 ex = MyTTSApp()
1638 ex.show()
1639 sys.exit(app.exec())