PraisonAI 3.0.0__py3-none-any.whl

praisonai/capabilities/embeddings.py

@@ -0,0 +1,158 @@
+"""
+Embeddings Capabilities Module
+
+Provides text embedding functionality via LiteLLM.
+"""
+
+from dataclasses import dataclass, field
+from typing import Optional, Union, List, Any, Dict
+
+
+@dataclass
+class EmbeddingResult:
+    """Result from embedding generation."""
+    embeddings: List[List[float]]
+    model: Optional[str] = None
+    usage: Optional[Dict[str, int]] = None
+    metadata: Dict[str, Any] = field(default_factory=dict)
+
+
+def embed(
+    input: Union[str, List[str]],
+    model: str = "text-embedding-3-small",
+    dimensions: Optional[int] = None,
+    encoding_format: str = "float",
+    timeout: float = 600.0,
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> EmbeddingResult:
+    """
+    Generate embeddings for text using LiteLLM.
+    
+    Args:
+        input: Text or list of texts to embed
+        model: Model name (e.g., "text-embedding-3-small", "text-embedding-3-large")
+        dimensions: Optional output dimensions (for models that support it)
+        encoding_format: "float" or "base64"
+        timeout: Request timeout in seconds
+        api_key: Optional API key override
+        api_base: Optional API base URL override
+        metadata: Optional metadata for tracing
+        
+    Returns:
+        EmbeddingResult with embeddings list
+        
+    Example:
+        >>> result = embed("Hello, world!")
+        >>> print(len(result.embeddings[0]))
+        
+        >>> result = embed(["Hello", "World"], model="text-embedding-3-large")
+    """
+    import litellm
+    
+    call_kwargs = {
+        'model': model,
+        'input': input,
+        'encoding_format': encoding_format,
+        'timeout': timeout,
+    }
+    
+    if dimensions:
+        call_kwargs['dimensions'] = dimensions
+    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.embedding(**call_kwargs)
+    
+    embeddings = []
+    if hasattr(response, 'data'):
+        for item in response.data:
+            if isinstance(item, dict):
+                embeddings.append(item.get('embedding', []))
+            else:
+                embeddings.append(getattr(item, 'embedding', []))
+    
+    usage = None
+    if hasattr(response, 'usage'):
+        usage = {
+            'prompt_tokens': getattr(response.usage, 'prompt_tokens', 0),
+            'total_tokens': getattr(response.usage, 'total_tokens', 0),
+        }
+    
+    return EmbeddingResult(
+        embeddings=embeddings,
+        model=model,
+        usage=usage,
+        metadata=metadata or {},
+    )
+
+
+async def aembed(
+    input: Union[str, List[str]],
+    model: str = "text-embedding-3-small",
+    dimensions: Optional[int] = None,
+    encoding_format: str = "float",
+    timeout: float = 600.0,
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> EmbeddingResult:
+    """
+    Async: Generate embeddings for text using LiteLLM.
+    
+    See embed() for full documentation.
+    """
+    import litellm
+    
+    call_kwargs = {
+        'model': model,
+        'input': input,
+        'encoding_format': encoding_format,
+        'timeout': timeout,
+    }
+    
+    if dimensions:
+        call_kwargs['dimensions'] = dimensions
+    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.aembedding(**call_kwargs)
+    
+    embeddings = []
+    if hasattr(response, 'data'):
+        for item in response.data:
+            if isinstance(item, dict):
+                embeddings.append(item.get('embedding', []))
+            else:
+                embeddings.append(getattr(item, 'embedding', []))
+    
+    usage = None
+    if hasattr(response, 'usage'):
+        usage = {
+            'prompt_tokens': getattr(response.usage, 'prompt_tokens', 0),
+            'total_tokens': getattr(response.usage, 'total_tokens', 0),
+        }
+    
+    return EmbeddingResult(
+        embeddings=embeddings,
+        model=model,
+        usage=usage,
+        metadata=metadata or {},
+    )

praisonai/capabilities/files.py

@@ -0,0 +1,467 @@
+"""
+Files Capabilities Module
+
+Provides file upload and management functionality via LiteLLM.
+"""
+
+from dataclasses import dataclass, field
+from typing import Optional, Union, List, Any, Dict, Literal, BinaryIO
+
+
+@dataclass
+class FileResult:
+    """Result from file operations."""
+    id: str
+    object: str = "file"
+    bytes: Optional[int] = None
+    created_at: Optional[int] = None
+    filename: Optional[str] = None
+    purpose: Optional[str] = None
+    status: Optional[str] = None
+    metadata: Dict[str, Any] = field(default_factory=dict)
+
+
+def file_create(
+    file: Union[str, bytes, BinaryIO],
+    purpose: Literal["assistants", "batch", "fine-tune"] = "assistants",
+    custom_llm_provider: str = "openai",
+    timeout: float = 600.0,
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> FileResult:
+    """
+    Upload a file for use with assistants, batch, or fine-tuning.
+    
+    Args:
+        file: File path, bytes, or file-like object
+        purpose: Purpose of the file ("assistants", "batch", "fine-tune")
+        custom_llm_provider: Provider ("openai", "azure", "vertex_ai", "bedrock")
+        timeout: Request timeout in seconds
+        api_key: Optional API key override
+        api_base: Optional API base URL override
+        metadata: Optional metadata for tracing
+        
+    Returns:
+        FileResult with file ID and metadata
+        
+    Example:
+        >>> result = file_create("./data.jsonl", purpose="batch")
+        >>> print(result.id)
+    """
+    import litellm
+    
+    file_obj = file
+    if isinstance(file, str):
+        file_obj = open(file, 'rb')
+    
+    try:
+        call_kwargs = {
+            'file': file_obj,
+            'purpose': purpose,
+            'custom_llm_provider': custom_llm_provider,
+            'timeout': timeout,
+        }
+        
+        if api_key:
+            call_kwargs['api_key'] = api_key
+        if api_base:
+            call_kwargs['api_base'] = api_base
+        
+        call_kwargs.update(kwargs)
+        
+        response = litellm.create_file(**call_kwargs)
+        
+        return FileResult(
+            id=getattr(response, 'id', ''),
+            object=getattr(response, 'object', 'file'),
+            bytes=getattr(response, 'bytes', None),
+            created_at=getattr(response, 'created_at', None),
+            filename=getattr(response, 'filename', None),
+            purpose=getattr(response, 'purpose', purpose),
+            status=getattr(response, 'status', None),
+            metadata=metadata or {},
+        )
+    finally:
+        if isinstance(file, str) and hasattr(file_obj, 'close'):
+            file_obj.close()
+
+
+async def afile_create(
+    file: Union[str, bytes, BinaryIO],
+    purpose: Literal["assistants", "batch", "fine-tune"] = "assistants",
+    custom_llm_provider: str = "openai",
+    timeout: float = 600.0,
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    metadata: Optional[Dict[str, Any]] = None,
+    **kwargs
+) -> FileResult:
+    """
+    Async: Upload a file for use with assistants, batch, or fine-tuning.
+    
+    See file_create() for full documentation.
+    """
+    import litellm
+    
+    file_obj = file
+    if isinstance(file, str):
+        file_obj = open(file, 'rb')
+    
+    try:
+        call_kwargs = {
+            'file': file_obj,
+            'purpose': purpose,
+            'custom_llm_provider': custom_llm_provider,
+            'timeout': timeout,
+        }
+        
+        if api_key:
+            call_kwargs['api_key'] = api_key
+        if api_base:
+            call_kwargs['api_base'] = api_base
+        
+        call_kwargs.update(kwargs)
+        
+        response = await litellm.acreate_file(**call_kwargs)
+        
+        return FileResult(
+            id=getattr(response, 'id', ''),
+            object=getattr(response, 'object', 'file'),
+            bytes=getattr(response, 'bytes', None),
+            created_at=getattr(response, 'created_at', None),
+            filename=getattr(response, 'filename', None),
+            purpose=getattr(response, 'purpose', purpose),
+            status=getattr(response, 'status', None),
+            metadata=metadata or {},
+        )
+    finally:
+        if isinstance(file, str) and hasattr(file_obj, 'close'):
+            file_obj.close()
+
+
+def file_list(
+    purpose: Optional[str] = None,
+    custom_llm_provider: str = "openai",
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    **kwargs
+) -> List[FileResult]:
+    """
+    List uploaded files.
+    
+    Args:
+        purpose: Filter by purpose
+        custom_llm_provider: Provider
+        api_key: Optional API key override
+        api_base: Optional API base URL override
+        
+    Returns:
+        List of FileResult objects
+    """
+    import litellm
+    
+    call_kwargs = {
+        'custom_llm_provider': custom_llm_provider,
+    }
+    
+    if purpose:
+        call_kwargs['purpose'] = purpose
+    if api_key:
+        call_kwargs['api_key'] = api_key
+    if api_base:
+        call_kwargs['api_base'] = api_base
+    
+    call_kwargs.update(kwargs)
+    
+    response = litellm.file_list(**call_kwargs)
+    
+    results = []
+    data = getattr(response, 'data', response) if hasattr(response, 'data') else response
+    if isinstance(data, list):
+        for item in data:
+            results.append(FileResult(
+                id=getattr(item, 'id', ''),
+                object=getattr(item, 'object', 'file'),
+                bytes=getattr(item, 'bytes', None),
+                created_at=getattr(item, 'created_at', None),
+                filename=getattr(item, 'filename', None),
+                purpose=getattr(item, 'purpose', None),
+                status=getattr(item, 'status', None),
+            ))
+    
+    return results
+
+
+async def afile_list(
+    purpose: Optional[str] = None,
+    custom_llm_provider: str = "openai",
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    **kwargs
+) -> List[FileResult]:
+    """
+    Async: List uploaded files.
+    
+    See file_list() for full documentation.
+    """
+    import litellm
+    
+    call_kwargs = {
+        'custom_llm_provider': custom_llm_provider,
+    }
+    
+    if purpose:
+        call_kwargs['purpose'] = purpose
+    if api_key:
+        call_kwargs['api_key'] = api_key
+    if api_base:
+        call_kwargs['api_base'] = api_base
+    
+    call_kwargs.update(kwargs)
+    
+    response = await litellm.afile_list(**call_kwargs)
+    
+    results = []
+    data = getattr(response, 'data', response) if hasattr(response, 'data') else response
+    if isinstance(data, list):
+        for item in data:
+            results.append(FileResult(
+                id=getattr(item, 'id', ''),
+                object=getattr(item, 'object', 'file'),
+                bytes=getattr(item, 'bytes', None),
+                created_at=getattr(item, 'created_at', None),
+                filename=getattr(item, 'filename', None),
+                purpose=getattr(item, 'purpose', None),
+                status=getattr(item, 'status', None),
+            ))
+    
+    return results
+
+
+def file_retrieve(
+    file_id: str,
+    custom_llm_provider: str = "openai",
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    **kwargs
+) -> FileResult:
+    """
+    Retrieve a file by ID.
+    
+    Args:
+        file_id: The file ID
+        custom_llm_provider: Provider
+        api_key: Optional API key override
+        api_base: Optional API base URL override
+        
+    Returns:
+        FileResult with file metadata
+    """
+    import litellm
+    
+    call_kwargs = {
+        'file_id': file_id,
+        'custom_llm_provider': custom_llm_provider,
+    }
+    
+    if api_key:
+        call_kwargs['api_key'] = api_key
+    if api_base:
+        call_kwargs['api_base'] = api_base
+    
+    call_kwargs.update(kwargs)
+    
+    response = litellm.file_retrieve(**call_kwargs)
+    
+    return FileResult(
+        id=getattr(response, 'id', file_id),
+        object=getattr(response, 'object', 'file'),
+        bytes=getattr(response, 'bytes', None),
+        created_at=getattr(response, 'created_at', None),
+        filename=getattr(response, 'filename', None),
+        purpose=getattr(response, 'purpose', None),
+        status=getattr(response, 'status', None),
+    )
+
+
+async def afile_retrieve(
+    file_id: str,
+    custom_llm_provider: str = "openai",
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    **kwargs
+) -> FileResult:
+    """
+    Async: Retrieve a file by ID.
+    
+    See file_retrieve() for full documentation.
+    """
+    import litellm
+    
+    call_kwargs = {
+        'file_id': file_id,
+        'custom_llm_provider': custom_llm_provider,
+    }
+    
+    if api_key:
+        call_kwargs['api_key'] = api_key
+    if api_base:
+        call_kwargs['api_base'] = api_base
+    
+    call_kwargs.update(kwargs)
+    
+    response = await litellm.afile_retrieve(**call_kwargs)
+    
+    return FileResult(
+        id=getattr(response, 'id', file_id),
+        object=getattr(response, 'object', 'file'),
+        bytes=getattr(response, 'bytes', None),
+        created_at=getattr(response, 'created_at', None),
+        filename=getattr(response, 'filename', None),
+        purpose=getattr(response, 'purpose', None),
+        status=getattr(response, 'status', None),
+    )
+
+
+def file_delete(
+    file_id: str,
+    custom_llm_provider: str = "openai",
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    **kwargs
+) -> bool:
+    """
+    Delete a file by ID.
+    
+    Args:
+        file_id: The file ID
+        custom_llm_provider: Provider
+        api_key: Optional API key override
+        api_base: Optional API base URL override
+        
+    Returns:
+        True if deleted successfully
+    """
+    import litellm
+    
+    call_kwargs = {
+        'file_id': file_id,
+        'custom_llm_provider': custom_llm_provider,
+    }
+    
+    if api_key:
+        call_kwargs['api_key'] = api_key
+    if api_base:
+        call_kwargs['api_base'] = api_base
+    
+    call_kwargs.update(kwargs)
+    
+    response = litellm.file_delete(**call_kwargs)
+    
+    return getattr(response, 'deleted', True)
+
+
+async def afile_delete(
+    file_id: str,
+    custom_llm_provider: str = "openai",
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    **kwargs
+) -> bool:
+    """
+    Async: Delete a file by ID.
+    
+    See file_delete() for full documentation.
+    """
+    import litellm
+    
+    call_kwargs = {
+        'file_id': file_id,
+        'custom_llm_provider': custom_llm_provider,
+    }
+    
+    if api_key:
+        call_kwargs['api_key'] = api_key
+    if api_base:
+        call_kwargs['api_base'] = api_base
+    
+    call_kwargs.update(kwargs)
+    
+    response = await litellm.afile_delete(**call_kwargs)
+    
+    return getattr(response, 'deleted', True)
+
+
+def file_content(
+    file_id: str,
+    custom_llm_provider: str = "openai",
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    **kwargs
+) -> bytes:
+    """
+    Get file content by ID.
+    
+    Args:
+        file_id: The file ID
+        custom_llm_provider: Provider
+        api_key: Optional API key override
+        api_base: Optional API base URL override
+        
+    Returns:
+        File content as bytes
+    """
+    import litellm
+    
+    call_kwargs = {
+        'file_id': file_id,
+        'custom_llm_provider': custom_llm_provider,
+    }
+    
+    if api_key:
+        call_kwargs['api_key'] = api_key
+    if api_base:
+        call_kwargs['api_base'] = api_base
+    
+    call_kwargs.update(kwargs)
+    
+    response = litellm.file_content(**call_kwargs)
+    
+    if hasattr(response, 'content'):
+        return response.content
+    return bytes(response) if response else b''
+
+
+async def afile_content(
+    file_id: str,
+    custom_llm_provider: str = "openai",
+    api_key: Optional[str] = None,
+    api_base: Optional[str] = None,
+    **kwargs
+) -> bytes:
+    """
+    Async: Get file content by ID.
+    
+    See file_content() for full documentation.
+    """
+    import litellm
+    
+    call_kwargs = {
+        'file_id': file_id,
+        'custom_llm_provider': custom_llm_provider,
+    }
+    
+    if api_key:
+        call_kwargs['api_key'] = api_key
+    if api_base:
+        call_kwargs['api_base'] = api_base
+    
+    call_kwargs.update(kwargs)
+    
+    response = await litellm.afile_content(**call_kwargs)
+    
+    if hasattr(response, 'content'):
+        return response.content
+    return bytes(response) if response else b''