Add PyAudio alternative installation and improve genie-tts setup process

installer.py

@@ -10,6 +10,88 @@ import logging
 logger = logging.getLogger(__name__)
 
 
+def install_pyaudio_alternative():
+    """尝试安装PyAudio的替代方案或创建mock模块"""
+    try:
+        # 首先尝试正常安装PyAudio
+        subprocess.check_call([
+            sys.executable, "-m", "pip", "install", "pyaudio"
+        ], timeout=120)
+        logger.info("✓ PyAudio 安装成功")
+        return True
+    except Exception as e:
+        logger.warning(f"PyAudio 安装失败: {e}")
+        
+        # 如果PyAudio安装失败，创建一个mock模块
+        try:
+            import os
+            import site
+            
+            # 获取site-packages路径
+            site_packages = site.getsitepackages()[0]
+            pyaudio_mock_path = os.path.join(site_packages, "pyaudio.py")
+            
+            # 创建mock pyaudio模块
+            mock_pyaudio_code = '''"""
+Mock PyAudio module for environments without audio support
+"""
+
+class PyAudioError(Exception):
+    pass
+
+class Stream:
+    def __init__(self, *args, **kwargs):
+        raise PyAudioError("音频流在此环境中不可用")
+    
+    def read(self, *args, **kwargs):
+        raise PyAudioError("音频读取在此环境中不可用")
+    
+    def write(self, *args, **kwargs):
+        raise PyAudioError("音频写入在此环境中不可用")
+    
+    def start_stream(self):
+        pass
+    
+    def stop_stream(self):
+        pass
+    
+    def close(self):
+        pass
+
+class PyAudio:
+    paInt16 = 8
+    paFloat32 = 1
+    
+    def __init__(self):
+        logger.warning("使用Mock PyAudio - 音频功能不可用")
+    
+    def open(self, *args, **kwargs):
+        return Stream()
+    
+    def get_device_count(self):
+        return 0
+    
+    def get_device_info_by_index(self, index):
+        return {"name": "Mock Device", "maxInputChannels": 0, "maxOutputChannels": 0}
+    
+    def terminate(self):
+        pass
+
+# 导出主要类和常量
+__all__ = ['PyAudio', 'Stream', 'PyAudioError']
+'''
+            
+            with open(pyaudio_mock_path, 'w', encoding='utf-8') as f:
+                f.write(mock_pyaudio_code)
+            
+            logger.info("✓ 创建 PyAudio mock 模块成功")
+            return True
+            
+        except Exception as mock_error:
+            logger.error(f"创建 PyAudio mock 模块失败: {mock_error}")
+            return False
+
+
 def install_genie_tts():
     """尝试安装genie-tts包，处理Hugging Face Spaces的限制"""
     try:
@@ -24,7 +106,10 @@ def install_genie_tts():
                 "onnxruntime>=1.16.0",  # 最关键：没有它TTS完全无法工作
                 "numpy>=1.21.0",        # 基础依赖
                 "soundfile>=0.12.0",    # 音频处理
-                "huggingface-hub>=0.17.0"  # 模型下载
+                "huggingface-hub>=0.17.0",  # 模型下载
+                "scipy>=1.9.0",         # 科学计算
+                "pyyaml>=6.0",         # 配置文件解析
+                "rich>=12.0.0"         # 终端输出美化
             ]
             
             logger.info("正在安装关键依赖...")
@@ -38,34 +123,41 @@ def install_genie_tts():
                     logger.error(f"✗ 关键依赖安装失败: {dep} - {e}")
                     return False, f"关键依赖 {dep} 安装失败: {str(e)}"
             
-            # 尝试安装genie-tts（不包含依赖，避免PyAudio问题）
-            logger.info("正在安装 genie-tts...")
+            # 处理PyAudio依赖
+            logger.info("正在处理PyAudio依赖...")
+            pyaudio_success = install_pyaudio_alternative()
+            if not pyaudio_success:
+                logger.warning("PyAudio处理失败，但将继续安装...")
+            
+            # 尝试安装可选的日语处理依赖
+            try:
+                subprocess.check_call([
+                    sys.executable, "-m", "pip", "install", "pyopenjtalk"
+                ], timeout=120)
+                logger.info("✓ 成功安装 pyopenjtalk")
+            except Exception as e:
+                logger.warning(f"⚠ pyopenjtalk 安装失败: {e}")
+            
+            # 尝试安装可选的音频重采样依赖
+            try:
+                subprocess.check_call([
+                    sys.executable, "-m", "pip", "install", "soxr"
+                ], timeout=120)
+                logger.info("✓ 成功安装 soxr")
+            except Exception as e:
+                logger.warning(f"⚠ soxr 安装失败: {e}")
+            
+            # 最后安装genie-tts主包
+            logger.info("正在安装 genie-tts 主包...")
             subprocess.check_call([
                 sys.executable, "-m", "pip", "install", 
-                "genie-tts", "--no-deps", "--upgrade"
+                "genie-tts", "--upgrade"
             ], timeout=300)
             
-            # 安装其他可选依赖
-            optional_deps = [
-                "scipy>=1.9.0", 
-                "rich>=12.0.0",
-                "pyopenjtalk"  # 可能因为C扩展编译失败
-            ]
-            
-            logger.info("正在安装可选依赖...")
-            for dep in optional_deps:
-                try:
-                    subprocess.check_call([
-                        sys.executable, "-m", "pip", "install", dep
-                    ], timeout=120)
-                    logger.info(f"✓ 成功安装可选依赖: {dep}")
-                except Exception as e:
-                    logger.warning(f"⚠ 可选依赖安装失败: {dep} - {e}")
-            
             # 验证安装
             import genie_tts
             logger.info("✅ genie-tts安装成功")
-            return True, None
+            return True, "安装成功，但某些音频功能可能不可用（PyAudio限制）"
             
         except subprocess.TimeoutExpired:
             error_msg = "安装超时：Hugging Face Spaces 环境可能不支持某些依赖"
@@ -74,14 +166,21 @@ def install_genie_tts():
             
         except Exception as e:
             error_msg = str(e)
-            if "portaudio" in error_msg.lower():
-                error_msg = ("PyAudio编译失败：Hugging Face Spaces环境缺少系统级音频依赖。"
-                           "这是已知的限制，请在本地环境运行或使用替代方案。")
-            elif "onnxruntime" in error_msg.lower():
-                error_msg = ("ONNX Runtime安装失败：这是Genie TTS的核心依赖，"
-                           "没有它无法运行任何TTS功能。请检查网络连接和环境配置。")
-            logger.error(f"安装genie-tts失败: {error_msg}")
-            return False, error_msg
+            
+            # 如果仍然是PyAudio相关错误，提供更详细的解决建议
+            if "pyaudio" in error_msg.lower():
+                error_msg = (
+                    "PyAudio依赖问题：即使创建了mock模块，genie-tts仍无法正常导入。"
+                    "建议解决方案：\n"
+                    "1. 在本地环境运行（安装了音频驱动）\n"
+                    "2. 使用替代的TTS库（如 pyttsx3 或 espeak）\n" 
+                    "3. 联系Hugging Face支持团队关于音频依赖问题"
+                )
+                logger.error(error_msg)
+                return False, error_msg
+            else:
+                logger.error(f"安装genie-tts失败: {error_msg}")
+                return False, error_msg
 
 
 def setup_genie_import():
@@ -92,9 +191,20 @@ def setup_genie_import():
         try:
             import genie_tts as genie
             logger.info("Genie TTS导入成功")
-            return genie, None
+            return genie, install_error  # 返回警告信息（如果有）
         except ImportError as e:
-            logger.error(f"导入Genie TTS失败: {e}")
-            return None, f"导入失败: {str(e)}"
+            error_msg = f"导入失败: {str(e)}"
+            
+            # 如果导入仍然失败，提供备用方案的建议
+            if "pyaudio" in error_msg.lower():
+                error_msg += (
+                    "\n\n建议的备用方案：\n"
+                    "1. 使用在线TTS服务（Google TTS, Azure Speech等）\n"
+                    "2. 在本地环境部署Genie TTS\n"
+                    "3. 使用其他兼容Hugging Face Spaces的TTS模型"
+                )
+            
+            logger.error(error_msg)
+            return None, error_msg
     else:
         return None, install_error