#!/usr/bin/env python3 """ Aliyun TTS - Smart Voice Selection 智能语言检测和音色选择，避免 418 错误 解决问题： 1. 自动检测文本语言（中文/英文/混合） 2. 根据语言智能选择合适音色 3. 避免音色语言不匹配导致的 418 错误 4. 文本长度检查和分段支持 """ import sys import os import re # 导入原有的 TTS 函数 sys.path.insert(0, os.path.dirname(__file__)) from aliyun_tts_fixed import synthesize_speech, get_aliyun_token def detect_language(text): """ 检测文本主要语言 返回: 'zh': 主要是中文 'en': 主要是英文 'mixed': 中英混合 """ # 统计中文字符（包括标点） chinese_chars = len(re.findall(r'[\u4e00-\u9fff]', text)) # 统计英文字母 english_chars = len(re.findall(r'[a-zA-Z]', text)) # 总字符数（去除空格和标点） total_content_chars = chinese_chars + english_chars if total_content_chars == 0: return 'zh' # 默认中文 # 计算比例 chinese_ratio = chinese_chars / total_content_chars english_ratio = english_chars / total_content_chars # 判断逻辑 if chinese_ratio >= 0.8: return 'zh' elif english_ratio >= 0.8: return 'en' else: return 'mixed' def get_voice_info(): """ 获取阿里云音色信息 返回音色字典，包含语言支持信息 """ return { # 中文音色（仅支持中文） 'zh_voices': { 'Zhiqi': {'gender': 'female', 'description': '智琪 - 清晰标准（推荐）'}, 'Xiaoyun': {'gender': 'female', 'description': '小云 - 亲和力'}, 'Aibao': {'gender': 'male', 'description': '艾宝 - 男声通用'}, 'Aitong': {'gender': 'child', 'description': '艾彤 - 儿童音'}, 'Aixia': {'gender': 'female', 'description': '艾霞 - 粤语'}, 'Aishuo': {'gender': 'male', 'description': '艾硕 - 四川话'}, }, # 英文音色（仅支持英文） 'en_voices': { 'Catherine': {'gender': 'female', 'description': '凯瑟琳 - 标准美音（推荐）'}, 'Kenny': {'gender': 'male', 'description': '肯尼 - 标准美音'}, 'Rosa': {'gender': 'female', 'description': '罗莎 - 温柔'}, }, # 多语言音色（支持中英混合） 'mixed_voices': { 'Aida': {'gender': 'female', 'description': '艾达 - 中英混合（推荐）'}, } } def select_voice_for_language(language, preferred_voice=None, gender_preference=None): """ 根据语言选择合适的音色 Args: language: 'zh', 'en', 'mixed' preferred_voice: 用户指定的音色（可选） gender_preference: 性别偏好 'male'/'female'（可选） Returns: 选定的音色名称 """ voice_info = get_voice_info() # 检查用户指定的音色是否兼容 if preferred_voice: # 检查音色在哪个类别 for category, voices in voice_info.items(): if preferred_voice in voices: # 检查兼容性 if language == 'zh' and category == 'zh_voices': return preferred_voice elif language == 'en' and category == 'en_voices': return preferred_voice elif language == 'mixed' and category in ['zh_voices', 'mixed_voices']: return preferred_voice else: print(f"⚠️ Voice '{preferred_voice}' not compatible with {language} text", file=sys.stderr) print(f" Voice category: {category}, Text language: {language}", file=sys.stderr) break # 自动选择音色 if language == 'zh': # 中文：优先 Zhiqi if gender_preference == 'male': return 'Aibao' return 'Zhiqi' elif language == 'en': # 英文：优先 Catherine if gender_preference == 'male': return 'Kenny' return 'Catherine' else: # mixed # 混合：使用 Aida return 'Aida' def check_text_length(text, max_length=300): """ 检查文本长度是否超限 Args: text: 文本内容 max_length: 最大长度限制（默认 300） Returns: (is_valid, message) """ text_length = len(text) if text_length <= max_length: return True, f"Text length OK: {text_length} chars" else: return False, f"Text too long: {text_length} chars (max {max_length})" def synthesize_speech_smart(text, output_file, voice=None, speech_rate=0, auto_select=True): """ 智能 TTS：自动检测语言并选择音色 Args: text: 文本内容 output_file: 输出文件路径 voice: 指定音色（可选） speech_rate: 语速 (-500 to 500) auto_select: 是否启用自动音色选择（默认 True） Returns: 成功返回 True，失败返回 False """ print(f"🎤 Aliyun TTS Smart Mode", file=sys.stderr) print(f" Text length: {len(text)} chars", file=sys.stderr) # 1. 检测语言 language = detect_language(text) language_names = {'zh': '中文', 'en': 'English', 'mixed': '中英混合'} print(f"🔍 Detected language: {language_names.get(language, language)}", file=sys.stderr) # 2. 检查文本长度 is_valid, length_msg = check_text_length(text) if not is_valid: print(f"⚠️ {length_msg}", file=sys.stderr) print(f" Hint: Consider splitting text into segments", file=sys.stderr) # 不直接返回失败，让 TTS API 处理（可能支持更长文本） # 3. 选择音色 if auto_select: final_voice = select_voice_for_language(language, voice) if final_voice != voice: print(f"🎵 Auto-selected voice: {final_voice} (was: {voice or 'none'})", file=sys.stderr) else: print(f"🎵 Using voice: {final_voice}", file=sys.stderr) else: final_voice = voice or 'Zhiqi' print(f"🎵 Using specified voice: {final_voice}", file=sys.stderr) # 显示音色信息 voice_info = get_voice_info() for category, voices in voice_info.items(): if final_voice in voices: print(f" {voices[final_voice]['description']}", file=sys.stderr) break # 4. 调用 TTS try: success = synthesize_speech(text, output_file, final_voice, speech_rate) if success: print(f"✅ TTS generation successful", file=sys.stderr) else: print(f"❌ TTS generation failed", file=sys.stderr) # 提供故障排除建议 print(f"\n💡 Troubleshooting tips:", file=sys.stderr) if language == 'mixed': print(f" - Try voice 'Aida' for mixed language text", file=sys.stderr) elif language == 'zh': print(f" - For Chinese, try: Zhiqi, Xiaoyun, or Aibao", file=sys.stderr) elif language == 'en': print(f" - For English, try: Catherine, Kenny, or Rosa", file=sys.stderr) print(f" - Check text content for sensitive words", file=sys.stderr) print(f" - Try splitting text if it's too long", file=sys.stderr) return success except Exception as e: print(f"❌ Exception: {str(e)}", file=sys.stderr) import traceback traceback.print_exc(file=sys.stderr) return False def main(): """命令行入口""" if len(sys.argv) < 3: print("Usage: aliyun_tts_smart.py [voice] [speed] [--no-auto-select]", file=sys.stderr) print("", file=sys.stderr) print("Options:", file=sys.stderr) print(" --no-auto-select Disable auto voice selection", file=sys.stderr) print("", file=sys.stderr) print("Examples:", file=sys.stderr) print(" # Auto-select voice based on language", file=sys.stderr) print(" python3 aliyun_tts_smart.py '你好世界' output.mp3", file=sys.stderr) print("", file=sys.stderr) print(" # Specify voice (will auto-switch if incompatible)", file=sys.stderr) print(" python3 aliyun_tts_smart.py 'Hello World' output.mp3 Aibao", file=sys.stderr) print("", file=sys.stderr) print(" # Disable auto-select", file=sys.stderr) print(" python3 aliyun_tts_smart.py 'Test' output.mp3 Zhiqi 0 --no-auto-select", file=sys.stderr) sys.exit(1) text = sys.argv[1] output_file = sys.argv[2] voice = sys.argv[3] if len(sys.argv) > 3 else None # 解析速度参数 speed = 1.0 auto_select = True for i in range(4, len(sys.argv)): arg = sys.argv[i] if arg == '--no-auto-select': auto_select = False else: try: speed = float(arg) except ValueError: pass # 转换速度格式 speech_rate = int((speed - 1.0) * 500) speech_rate = max(-500, min(500, speech_rate)) # 调用智能 TTS success = synthesize_speech_smart(text, output_file, voice, speech_rate, auto_select) sys.exit(0 if success else 1) if __name__ == "__main__": main()