PraisonAI 3.0.0__py3-none-any.whl

praisonai/capabilities/images.py

@@ -0,0 +1,330 @@
+"""
+Images Capabilities Module
+
+Provides image generation and editing functionality via LiteLLM.
+"""
+
+from dataclasses import dataclass, field
+from typing import Optional, Union, List, Any, Dict
+
+
+@dataclass
+class ImageResult:
+    """Result from image generation or editing."""
+    url: Optional[str] = None
+    b64_json: Optional[str] = None
+    revised_prompt: Optional[str] = None
+    model: Optional[str] = None
+    metadata: Dict[str, Any] = field(default_factory=dict)
+    
+    def save(self, path: str) -> str:
+        """Save image to file (from base64 or download from URL)."""
+        if self.b64_json:
+            import base64
+            with open(path, 'wb') as f:
+                f.write(base64.b64decode(self.b64_json))
+        elif self.url:
+            import urllib.request
+            urllib.request.urlretrieve(self.url, path)
+        else:
+            raise ValueError("No image data available")
+        return path
+
+
+def image_generate(
+    prompt: str,
+    model: str = "dall-e-3",
+    n: int = 1,
+    size: str = "1024x1024",
+    quality: str = "standard",
+    style: Optional[str] = None,
+    response_format: str = "url",
+    timeout: float = 600.0,
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> List[ImageResult]:
+    """
+    Generate images from a text prompt using LiteLLM.
+    
+    Args:
+        prompt: Text description of the image to generate
+        model: Model name (e.g., "dall-e-3", "dall-e-2", "stability/...")
+        n: Number of images to generate
+        size: Image size (e.g., "1024x1024", "1792x1024", "1024x1792")
+        quality: Image quality ("standard" or "hd")
+        style: Image style ("vivid" or "natural")
+        response_format: "url" or "b64_json"
+        timeout: Request timeout in seconds
+        api_key: Optional API key override
+        api_base: Optional API base URL override
+        metadata: Optional metadata for tracing
+        
+    Returns:
+        List of ImageResult objects
+        
+    Example:
+        >>> results = image_generate("A sunset over mountains")
+        >>> results[0].save("sunset.png")
+    """
+    import litellm
+    
+    call_kwargs = {
+        'model': model,
+        'prompt': prompt,
+        'n': n,
+        'size': size,
+        'response_format': response_format,
+        'timeout': timeout,
+    }
+    
+    # Only add quality/style for dall-e-3 (not supported by dall-e-2)
+    if 'dall-e-3' in model.lower():
+        call_kwargs['quality'] = quality
+        if style:
+            call_kwargs['style'] = style
+    if api_key:
+        call_kwargs['api_key'] = api_key
+    if api_base:
+        call_kwargs['api_base'] = api_base
+    
+    call_kwargs.update(kwargs)
+    
+    if metadata:
+        call_kwargs['metadata'] = metadata
+    
+    response = litellm.image_generation(**call_kwargs)
+    
+    results = []
+    if hasattr(response, 'data'):
+        for item in response.data:
+            results.append(ImageResult(
+                url=getattr(item, 'url', None),
+                b64_json=getattr(item, 'b64_json', None),
+                revised_prompt=getattr(item, 'revised_prompt', None),
+                model=model,
+                metadata=metadata or {},
+            ))
+    
+    return results
+
+
+async def aimage_generate(
+    prompt: str,
+    model: str = "dall-e-3",
+    n: int = 1,
+    size: str = "1024x1024",
+    quality: str = "standard",
+    style: Optional[str] = None,
+    response_format: str = "url",
+    timeout: float = 600.0,
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> List[ImageResult]:
+    """
+    Async: Generate images from a text prompt using LiteLLM.
+    
+    See image_generate() for full documentation.
+    """
+    import litellm
+    
+    call_kwargs = {
+        'model': model,
+        'prompt': prompt,
+        'n': n,
+        'size': size,
+        'response_format': response_format,
+        'timeout': timeout,
+    }
+    
+    # Only add quality/style for dall-e-3 (not supported by dall-e-2)
+    if 'dall-e-3' in model.lower():
+        call_kwargs['quality'] = quality
+        if style:
+            call_kwargs['style'] = style
+    if api_key:
+        call_kwargs['api_key'] = api_key
+    if api_base:
+        call_kwargs['api_base'] = api_base
+    
+    call_kwargs.update(kwargs)
+    
+    if metadata:
+        call_kwargs['metadata'] = metadata
+    
+    response = await litellm.aimage_generation(**call_kwargs)
+    
+    results = []
+    if hasattr(response, 'data'):
+        for item in response.data:
+            results.append(ImageResult(
+                url=getattr(item, 'url', None),
+                b64_json=getattr(item, 'b64_json', None),
+                revised_prompt=getattr(item, 'revised_prompt', None),
+                model=model,
+                metadata=metadata or {},
+            ))
+    
+    return results
+
+
+def image_edit(
+    image: Union[str, bytes],
+    prompt: str,
+    model: str = "dall-e-2",
+    mask: Optional[Union[str, bytes]] = None,
+    n: int = 1,
+    size: str = "1024x1024",
+    response_format: str = "url",
+    timeout: float = 600.0,
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> List[ImageResult]:
+    """
+    Edit an image using a text prompt.
+    
+    Args:
+        image: Path to image file or image bytes
+        prompt: Text description of the edit
+        model: Model name
+        mask: Optional mask image (transparent areas will be edited)
+        n: Number of images to generate
+        size: Output image size
+        response_format: "url" or "b64_json"
+        timeout: Request timeout in seconds
+        api_key: Optional API key override
+        api_base: Optional API base URL override
+        metadata: Optional metadata for tracing
+        
+    Returns:
+        List of ImageResult objects
+    """
+    import litellm
+    
+    # Handle file paths
+    image_file = image
+    if isinstance(image, str):
+        image_file = open(image, 'rb')
+    
+    mask_file = mask
+    if mask and isinstance(mask, str):
+        mask_file = open(mask, 'rb')
+    
+    try:
+        call_kwargs = {
+            'model': model,
+            'image': image_file,
+            'prompt': prompt,
+            'n': n,
+            'size': size,
+            'response_format': response_format,
+            'timeout': timeout,
+        }
+        
+        if mask_file:
+            call_kwargs['mask'] = mask_file
+        if api_key:
+            call_kwargs['api_key'] = api_key
+        if api_base:
+            call_kwargs['api_base'] = api_base
+        
+        call_kwargs.update(kwargs)
+        
+        if metadata:
+            call_kwargs['metadata'] = metadata
+        
+        response = litellm.image_edit(**call_kwargs)
+        
+        results = []
+        if hasattr(response, 'data'):
+            for item in response.data:
+                results.append(ImageResult(
+                    url=getattr(item, 'url', None),
+                    b64_json=getattr(item, 'b64_json', None),
+                    model=model,
+                    metadata=metadata or {},
+                ))
+        
+        return results
+    finally:
+        if isinstance(image, str) and hasattr(image_file, 'close'):
+            image_file.close()
+        if mask and isinstance(mask, str) and hasattr(mask_file, 'close'):
+            mask_file.close()
+
+
+async def aimage_edit(
+    image: Union[str, bytes],
+    prompt: str,
+    model: str = "dall-e-2",
+    mask: Optional[Union[str, bytes]] = None,
+    n: int = 1,
+    size: str = "1024x1024",
+    response_format: str = "url",
+    timeout: float = 600.0,
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> List[ImageResult]:
+    """
+    Async: Edit an image using a text prompt.
+    
+    See image_edit() for full documentation.
+    """
+    import litellm
+    
+    image_file = image
+    if isinstance(image, str):
+        image_file = open(image, 'rb')
+    
+    mask_file = mask
+    if mask and isinstance(mask, str):
+        mask_file = open(mask, 'rb')
+    
+    try:
+        call_kwargs = {
+            'model': model,
+            'image': image_file,
+            'prompt': prompt,
+            'n': n,
+            'size': size,
+            'response_format': response_format,
+            'timeout': timeout,
+        }
+        
+        if mask_file:
+            call_kwargs['mask'] = mask_file
+        if api_key:
+            call_kwargs['api_key'] = api_key
+        if api_base:
+            call_kwargs['api_base'] = api_base
+        
+        call_kwargs.update(kwargs)
+        
+        if metadata:
+            call_kwargs['metadata'] = metadata
+        
+        response = await litellm.aimage_edit(**call_kwargs)
+        
+        results = []
+        if hasattr(response, 'data'):
+            for item in response.data:
+                results.append(ImageResult(
+                    url=getattr(item, 'url', None),
+                    b64_json=getattr(item, 'b64_json', None),
+                    model=model,
+                    metadata=metadata or {},
+                ))
+        
+        return results
+    finally:
+        if isinstance(image, str) and hasattr(image_file, 'close'):
+            image_file.close()
+        if mask and isinstance(mask, str) and hasattr(mask_file, 'close'):
+            mask_file.close()

praisonai/capabilities/mcp.py

@@ -0,0 +1,190 @@
+"""
+MCP (Model Context Protocol) Capabilities Module
+
+Provides MCP server interaction functionality.
+"""
+
+from dataclasses import dataclass, field
+from typing import Optional, Any, Dict, List
+
+
+@dataclass
+class MCPResult:
+    """Result from MCP operations."""
+    tools: Optional[List[Dict[str, Any]]] = None
+    resources: Optional[List[Dict[str, Any]]] = None
+    prompts: Optional[List[Dict[str, Any]]] = None
+    server_name: Optional[str] = None
+    metadata: Dict[str, Any] = field(default_factory=dict)
+
+
+@dataclass
+class MCPToolCallResult:
+    """Result from MCP tool call."""
+    result: Any
+    tool_name: str
+    server_name: Optional[str] = None
+    metadata: Dict[str, Any] = field(default_factory=dict)
+
+
+def mcp_list_tools(
+    command: str,
+    args: Optional[List[str]] = None,
+    env: Optional[Dict[str, str]] = None,
+    timeout: float = 30.0,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> MCPResult:
+    """
+    List tools from an MCP server.
+    
+    Args:
+        command: MCP server command (e.g., "npx")
+        args: Command arguments
+        env: Environment variables
+        timeout: Request timeout in seconds
+        metadata: Optional metadata for tracing
+        
+    Returns:
+        MCPResult with available tools
+        
+    Example:
+        >>> result = mcp_list_tools("npx", ["-y", "@modelcontextprotocol/server-filesystem", "."])
+        >>> for tool in result.tools:
+        ...     print(tool['name'])
+    """
+    try:
+        from praisonaiagents.mcp import MCP
+        
+        mcp = MCP(command=command, args=args or [], env=env)
+        tools = mcp.get_tools()
+        
+        tool_list = []
+        for tool in tools:
+            tool_list.append({
+                'name': getattr(tool, 'name', str(tool)),
+                'description': getattr(tool, 'description', ''),
+            })
+        
+        return MCPResult(
+            tools=tool_list,
+            server_name=command,
+            metadata=metadata or {},
+        )
+    except ImportError:
+        return MCPResult(
+            tools=[],
+            server_name=command,
+            metadata={"error": "MCP not available", **(metadata or {})},
+        )
+
+
+async def amcp_list_tools(
+    command: str,
+    args: Optional[List[str]] = None,
+    env: Optional[Dict[str, str]] = None,
+    timeout: float = 30.0,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> MCPResult:
+    """
+    Async: List tools from an MCP server.
+    
+    See mcp_list_tools() for full documentation.
+    """
+    return mcp_list_tools(
+        command=command,
+        args=args,
+        env=env,
+        timeout=timeout,
+        metadata=metadata,
+        **kwargs
+    )
+
+
+def mcp_call_tool(
+    command: str,
+    tool_name: str,
+    tool_args: Optional[Dict[str, Any]] = None,
+    args: Optional[List[str]] = None,
+    env: Optional[Dict[str, str]] = None,
+    timeout: float = 60.0,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> MCPToolCallResult:
+    """
+    Call a tool on an MCP server.
+    
+    Args:
+        command: MCP server command
+        tool_name: Name of the tool to call
+        tool_args: Arguments for the tool
+        args: Command arguments
+        env: Environment variables
+        timeout: Request timeout in seconds
+        metadata: Optional metadata for tracing
+        
+    Returns:
+        MCPToolCallResult with tool result
+    """
+    try:
+        from praisonaiagents.mcp import MCP
+        
+        mcp = MCP(command=command, args=args or [], env=env)
+        
+        # Get the tool and call it
+        tools = mcp.get_tools()
+        for tool in tools:
+            if getattr(tool, 'name', str(tool)) == tool_name:
+                if callable(tool):
+                    result = tool(**(tool_args or {}))
+                else:
+                    result = None
+                
+                return MCPToolCallResult(
+                    result=result,
+                    tool_name=tool_name,
+                    server_name=command,
+                    metadata=metadata or {},
+                )
+        
+        return MCPToolCallResult(
+            result=None,
+            tool_name=tool_name,
+            server_name=command,
+            metadata={"error": f"Tool '{tool_name}' not found", **(metadata or {})},
+        )
+    except ImportError:
+        return MCPToolCallResult(
+            result=None,
+            tool_name=tool_name,
+            server_name=command,
+            metadata={"error": "MCP not available", **(metadata or {})},
+        )
+
+
+async def amcp_call_tool(
+    command: str,
+    tool_name: str,
+    tool_args: Optional[Dict[str, Any]] = None,
+    args: Optional[List[str]] = None,
+    env: Optional[Dict[str, str]] = None,
+    timeout: float = 60.0,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> MCPToolCallResult:
+    """
+    Async: Call a tool on an MCP server.
+    
+    See mcp_call_tool() for full documentation.
+    """
+    return mcp_call_tool(
+        command=command,
+        tool_name=tool_name,
+        tool_args=tool_args,
+        args=args,
+        env=env,
+        timeout=timeout,
+        metadata=metadata,
+        **kwargs
+    )