Remove voice manager tests and update Dockerfiles for improved dependency management and user permissions

api/tests/test_voice_manager.py

@@ -1,134 +0,0 @@
-import pytest
-from unittest.mock import AsyncMock, patch, MagicMock
-import torch
-from pathlib import Path
-
-from ..src.inference.voice_manager import VoiceManager
-from ..src.structures.model_schemas import VoiceConfig
-
-
-@pytest.fixture
-def mock_voice_tensor():
-    return torch.randn(10, 10)  # Dummy tensor
-
-
-@pytest.fixture
-def voice_manager():
-    return VoiceManager(VoiceConfig())
-
-
-@pytest.mark.asyncio
-async def test_load_voice(voice_manager, mock_voice_tensor):
-    """Test loading a single voice"""
-    with patch("api.src.core.paths.load_voice_tensor", new_callable=AsyncMock) as mock_load:
-        mock_load.return_value = mock_voice_tensor
-        with patch("os.path.exists", return_value=True):
-            voice = await voice_manager.load_voice("af_bella", "cpu")
-            assert torch.equal(voice, mock_voice_tensor)
-
-
-@pytest.mark.asyncio
-async def test_load_voice_not_found(voice_manager):
-    """Test loading non-existent voice"""
-    with patch("os.path.exists", return_value=False):
-        with pytest.raises(RuntimeError, match="Voice not found: invalid_voice"):
-            await voice_manager.load_voice("invalid_voice", "cpu")
-
-
-@pytest.mark.skip(reason="Local saving is optional and not critical to functionality")
-@pytest.mark.asyncio
-async def test_combine_voices_with_saving(voice_manager, mock_voice_tensor):
-    """Test combining voices with local saving enabled"""
-    pass
-
-
-@pytest.mark.asyncio
-async def test_combine_voices_without_saving(voice_manager, mock_voice_tensor):
-    """Test combining voices without local saving"""
-    with patch("api.src.core.paths.load_voice_tensor", new_callable=AsyncMock) as mock_load, \
-         patch("torch.save") as mock_save, \
-         patch("os.makedirs"), \
-         patch("os.path.exists", return_value=True):
-        
-        # Setup mocks
-        mock_load.return_value = mock_voice_tensor
-        
-        # Mock settings
-        with patch("api.src.core.config.settings") as mock_settings:
-            mock_settings.allow_local_voice_saving = False
-            mock_settings.voices_dir = "/mock/voices"
-            
-            # Combine voices
-            combined = await voice_manager.combine_voices(["af_bella", "af_sarah"], "cpu")
-            assert combined == "af_bella+af_sarah"  # Note: using + separator
-            
-            # Verify voice was not saved
-            mock_save.assert_not_called()
-
-
-@pytest.mark.asyncio
-async def test_combine_voices_single_voice(voice_manager):
-    """Test combining with single voice"""
-    with pytest.raises(ValueError, match="At least 2 voices are required"):
-        await voice_manager.combine_voices(["af_bella"], "cpu")
-
-
-@pytest.mark.asyncio
-async def test_list_voices(voice_manager):
-    """Test listing available voices"""
-    with patch("os.listdir", return_value=["af_bella.pt", "af_sarah.pt", "af_bella+af_sarah.pt"]), \
-         patch("os.makedirs"):
-        voices = await voice_manager.list_voices()
-        assert len(voices) == 3
-        assert "af_bella" in voices
-        assert "af_sarah" in voices
-        assert "af_bella+af_sarah" in voices
-
-
-@pytest.mark.asyncio
-async def test_load_combined_voice(voice_manager, mock_voice_tensor):
-    """Test loading a combined voice"""
-    with patch("api.src.core.paths.load_voice_tensor", new_callable=AsyncMock) as mock_load:
-        mock_load.return_value = mock_voice_tensor
-        with patch("os.path.exists", return_value=True):
-            voice = await voice_manager.load_voice("af_bella+af_sarah", "cpu")
-            assert torch.equal(voice, mock_voice_tensor)
-
-
-def test_cache_management(mock_voice_tensor):
-    """Test voice cache management"""
-    # Create voice manager with small cache size
-    config = VoiceConfig(cache_size=2)
-    voice_manager = VoiceManager(config)
-    
-    # Add items to cache
-    voice_manager._voice_cache = {
-        "voice1_cpu": torch.randn(5, 5),
-        "voice2_cpu": torch.randn(5, 5),
-        "voice3_cpu": torch.randn(5, 5),  # Add one more than cache size
-    }
-    
-    # Try managing cache
-    voice_manager._manage_cache()
-    
-    # Check cache size maintained
-    assert len(voice_manager._voice_cache) <= 2
-
-
-@pytest.mark.asyncio
-async def test_voice_loading_with_cache(voice_manager, mock_voice_tensor):
-    """Test voice loading with cache enabled"""
-    with patch("api.src.core.paths.load_voice_tensor", new_callable=AsyncMock) as mock_load, \
-         patch("os.path.exists", return_value=True):
-        
-        mock_load.return_value = mock_voice_tensor
-        
-        # First load should hit disk
-        voice1 = await voice_manager.load_voice("af_bella", "cpu")
-        assert mock_load.call_count == 1
-        
-        # Second load should hit cache
-        voice2 = await voice_manager.load_voice("af_bella", "cpu")
-        assert mock_load.call_count == 1  # Still 1
-        
-        assert torch.equal(voice1, voice2)

docker/cpu/Dockerfile

@@ -1,12 +1,14 @@
-FROM --platform=$BUILDPLATFORM python:3.10-slim
+FROM --platform=$BUILDPLATFORM python:3.10-slim 
 
-# Install dependencies
+# Install dependencies and check espeak location
-RUN apt-get update && apt-get install -y --no-install-recommends \
+RUN apt-get update && apt-get install -y \
     espeak-ng \
     git \
     libsndfile1 \
     curl \
     ffmpeg \
+    && dpkg -L espeak-ng \
+    && find / -name "espeak-ng-data" \
     && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
 
@@ -15,12 +17,12 @@ RUN curl -LsSf https://astral.sh/uv/install.sh | sh && \
     mv /root/.local/bin/uv /usr/local/bin/ && \
     mv /root/.local/bin/uvx /usr/local/bin/
 
-# Create non-root user
+# Create non-root user and set up directories and permissions
-RUN useradd -m -u 1000 appuser
+RUN useradd -m -u 1000 appuser && \
+    mkdir -p /app/api/src/models/v1_0 && \
+    chown -R appuser:appuser /app && \
+    chown -R appuser:appuser /lib/x86_64-linux-gnu/espeak-ng-data
 
-# Create directories and set ownership
-RUN mkdir -p /app/api/src/models/v1_0 && \
-    chown -R appuser:appuser /app
 
 USER appuser
 WORKDIR /app
@@ -49,7 +51,6 @@ ENV PYTHONUNBUFFERED=1 \
     UV_LINK_MODE=copy \
     USE_GPU=false
 
-# Core settings that differ from config.py defaults
 ENV DOWNLOAD_MODEL=true
 # Download model if enabled
 RUN if [ "$DOWNLOAD_MODEL" = "true" ]; then \
@@ -57,4 +58,4 @@ RUN if [ "$DOWNLOAD_MODEL" = "true" ]; then \
     fi
 
 # Run FastAPI server
-CMD ["uv", "run", "python", "-m", "uvicorn", "api.src.main:app", "--host", "0.0.0.0", "--port", "8880", "--log-level", "debug"]
+CMD ["uv", "run", "python", "-m", "uvicorn", "api.src.main:app", "--host", "0.0.0.0", "--port", "8880", "--log-level", "debug"]

docker/gpu/Dockerfile

@@ -1,9 +1,10 @@
-FROM --platform=$BUILDPLATFORM nvidia/cuda:12.1.0-runtime-ubuntu22.04
+FROM --platform=$BUILDPLATFORM nvidia/cuda:12.1.0-cudnn8-runtime-ubuntu22.04
+
 # Set non-interactive frontend
 ENV DEBIAN_FRONTEND=noninteractive
 
 # Install Python and other dependencies
-RUN apt-get update && apt-get install -y --no-install-recommends \
+RUN apt-get update && apt-get install -y \
     python3.10 \
     python3.10-venv \
     espeak-ng \
@@ -11,19 +12,26 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
     libsndfile1 \
     curl \
     ffmpeg \
- && apt-get clean && rm -rf /var/lib/apt/lists/*
+ && ls -la /usr/lib/x86_64-linux-gnu/espeak-ng-data \
+ && apt-get clean \
+ && rm -rf /var/lib/apt/lists/*
 
-# Install UV using the installer script
+# Create user and set up permissions 
+RUN useradd -m -u 1000 appuser && \
+    mkdir -p /app/api/src/models/v1_0 && \
+    chown -R appuser:appuser /app && \
+    chown -R appuser:appuser /usr/lib/x86_64-linux-gnu/espeak-ng-data
+
+
+# Rest of your Dockerfile...
+
+# Install UV in a separate step
 RUN curl -LsSf https://astral.sh/uv/install.sh | sh && \
     mv /root/.local/bin/uv /usr/local/bin/ && \
-    mv /root/.local/bin/uvx /usr/local/bin/ && \
+    mv /root/.local/bin/uvx /usr/local/bin/
-    useradd -m -u 1000 appuser && \
-    mkdir -p /app/api/src/models/v1_0 && \
-    chown -R appuser:appuser /app
 
 USER appuser
 WORKDIR /app
-
 # Copy dependency files
 COPY --chown=appuser:appuser pyproject.toml ./pyproject.toml