| import pandas as pd |
| import zipfile |
| import base64 |
| import os |
| import requests |
| import re |
| from pydub import AudioSegment |
| import json |
|
|
| def remove_quotes(text): |
| |
| return text.replace('"', '') |
|
|
| def text_to_speech(input_file, selected_option): |
| |
| api_key = 'AIzaSyBqvy0wRyZVapYyZ_GfTlRpQwrwj9YqYaw' |
| data = pd.read_csv(input_file) |
| zip_path = 'output_audio_files.zip' |
| error_log = [] |
|
|
| with zipfile.ZipFile(zip_path, 'w') as z: |
| for idx, row in data.iterrows(): |
| try: |
| |
| script = row.get('script', '') |
| print("原稿", script) |
| if pd.isna(script): |
| error_log.append(f"行 {idx+1}: スクリプトが空です") |
| continue |
| if not isinstance(script, str): |
| script = str(script) |
| |
| |
| parts = re.split(r'(A:|B:)', script) |
| ssml_parts = [] |
|
|
| |
| for i in range(1, len(parts), 2): |
| if parts[i] == "A:": |
| voice_name = row["voiceA"] |
| elif parts[i] == "B:": |
| voice_name = row["voiceB"] |
| else: |
| continue |
| |
| text = parts[i + 1].strip() |
| text = remove_quotes(text) |
|
|
| |
| text = text.replace("a.m.", 'AM') |
| text = text.replace("p.m.", 'PM') |
| text = text.replace("U.S.", 'US') |
| text = text.replace("U.K.", 'UK') |
| text = text.replace("Mr.", 'Mister') |
| text = text.replace("Ms.", 'MIZ') |
| text = text.replace("Mrs.", 'Misiz') |
| text = text.replace("Dr.", 'Doctor') |
| text = text.replace("Mt.", 'Mount') |
| text = text.replace("Prof.", 'Professor') |
|
|
| |
| text = text.replace("\n", '<break time="1s"/>') |
| text = text.replace(".", '.<break time="500ms"/>') |
|
|
| if selected_option == "ブレイクタイム有": |
| |
| if row["eikenn"] in ["5級", "4級", "3級", "準2級", "2級", "準1級"]: |
| text = text.replace(",", '<break time="50ms"/>') |
| |
| ssml_parts.append(f'<voice name="{voice_name}"><prosody rate="{row["speed"]}"><p>{text}</p></prosody></voice>') |
|
|
| ssml = '<speak>' + ''.join(ssml_parts) |
| |
| if pd.notna(row.get('question')) and row['question'] != '': |
| ssml += f'<break time="1s"/><voice name="{row["voiceQuestion"]}"><prosody rate="{row["speed"]}"><p>Question</p><break time="1s"/><p>{row["question"]}</p></prosody></voice>' |
| if pd.notna(row.get('choices')) and row['choices'] != '': |
| choices_list = row['choices'].split('/') |
| choices_ssml = '<break time="1s"/>'.join(choices_list) |
| ssml += f'<break time="1s"/><voice name="{row["voiceB"]}"><prosody rate="{row["speed"]}"><p>{choices_ssml}</p></prosody></voice>' |
| ssml += '</speak>' |
|
|
| |
| body = { |
| "input": {"ssml": ssml}, |
| "voice": {"languageCode": "en-US"}, |
| "audioConfig": {"audioEncoding": "MP3"} |
| } |
| headers = { |
| "X-Goog-Api-Key": api_key, |
| "Content-Type": "application/json" |
| } |
| url = "https://texttospeech.googleapis.com/v1/text:synthesize" |
| |
| |
| try: |
| response = requests.post(url, headers=headers, json=body) |
| |
| |
| if response.status_code != 200: |
| error_message = f"行 {idx+1}: APIエラー - HTTP {response.status_code}" |
| try: |
| error_data = response.json() |
| if 'error' in error_data and 'message' in error_data['error']: |
| error_message += f" - {error_data['error']['message']}" |
| except: |
| error_message += f" - レスポンス: {response.text}" |
| |
| print(error_message) |
| error_log.append(error_message) |
| continue |
| |
| response_data = response.json() |
| |
| |
| if 'audioContent' not in response_data: |
| error_message = f"行 {idx+1}: 音声データがレスポンスに含まれていません" |
| print(error_message) |
| error_log.append(error_message) |
| continue |
| |
| |
| audio_content = base64.b64decode(response_data['audioContent']) |
| temp_file_name = f"{row['id']}_temp.mp3" |
| final_file_name = f"{row['id']}.mp3" |
| |
| |
| with open(temp_file_name, "wb") as out: |
| out.write(audio_content) |
| |
| |
| audio = AudioSegment.from_file(temp_file_name, format="mp3") |
| audio.export(final_file_name, format="mp3", bitrate="128k") |
| |
| |
| z.write(final_file_name) |
| |
| |
| os.remove(temp_file_name) |
| os.remove(final_file_name) |
| |
| print(f"行 {idx+1}: 処理成功 - ファイルID: {row['id']}") |
| |
| except requests.exceptions.RequestException as e: |
| error_message = f"行 {idx+1}: API接続エラー - {str(e)}" |
| print(error_message) |
| error_log.append(error_message) |
| |
| except Exception as e: |
| error_message = f"行 {idx+1}: 予期せぬエラー - {str(e)}" |
| print(error_message) |
| error_log.append(error_message) |
|
|
| |
| if error_log: |
| print("\n===== エラーログ =====") |
| for error in error_log: |
| print(error) |
| print(f"合計 {len(error_log)} 件のエラーが発生しました。") |
| |
| |
| with open('tts_error_log.txt', 'w', encoding='utf-8') as log_file: |
| log_file.write("===== Google Text-to-Speech API エラーログ =====\n") |
| log_file.write(f"処理日時: {pd.Timestamp.now()}\n\n") |
| for error in error_log: |
| log_file.write(f"{error}\n") |
| print("エラーログは 'tts_error_log.txt' に保存されました。") |
| else: |
| print("すべての行が正常に処理されました。") |
|
|
| return zip_path |