Types - Data Structures

Overview

The Types component defines the core data structures and models used throughout the TKM AI Agency Platform. It provides standardized types for message handling, agent responses, and state management using Pydantic models and enums.

Directory Structure

Backend/Core_Components/
└── types.py           # Core type definitions

Main Components

Enums

Core enumeration types:

• BlockingState
• MessageType
• Response states
• Message formats

Data Models

Pydantic models for:

• Agent responses
• Folder structures
• Request formats
• Response formats

Type Definitions

1. State Management

   • Blocking states
   • Process states
   • Error states
   • Response states

2. Message Handling

   • Message types
   • Content formats
   • Metadata structures
   • Response formats

Data Structures

BlockingState Enum

class BlockingState(str, Enum):
    ACTIVE = "ACTIVE"
    WAITING_CONFIRMATION = "WAITING_CONFIRMATION"
    WAITING_INPUT = "WAITING_INPUT"
    DONE = "DONE"
    FAILED = "FAILED"
    ERROR = "ERROR"
    WAITING = "WAITING"

MessageType Enum

class MessageType(str, Enum):
    AUDIO = "audio"
    TEXT = "text"
    IMAGE = "image"
    DOCUMENT = "document"

AgentResponse Model

class AgentResponse(BaseModel):
    success: bool
    message: str
    source_agent: str
    conversation_id: str
    metadata: Dict[str, Any] = {}
    tokens: Optional[int] = None
    folder_structure: Optional[Dict[str, Any]] = None

FolderStructure Model

class FolderStructure(BaseModel):
    folder_name: str
    file_name: str
    category: str
    subcategory: Optional[str] = None
    state: BlockingState = BlockingState.ACTIVE
    confidence: float

Key Features

1. Type Safety

   • Pydantic validation
   • Type checking
   • Default values
   • Optional fields

2. State Management

   • Process states
   • Blocking states
   • Error states
   • Response states

3. Message Formats

   • Content types
   • Response formats
   • Metadata structures
   • Validation rules

Usage Examples

Agent Response

response = AgentResponse(
    success=True,
    message="Operation completed",
    source_agent="prometheus",
    conversation_id="conv_123",
    metadata={"process_time": "0.5s"}
)

Folder Structure

structure = FolderStructure(
    folder_name="documents",
    file_name="report.pdf",
    category="reports",
    subcategory="financial",
    state=BlockingState.ACTIVE,
    confidence=0.95
)

Integration

Model Usage

• Request validation
• Response formatting
• State tracking
• Error handling

Type Applications

• Data validation
• State management
• Message processing
• Response handling

Error Handling

1. Validation Errors

   • Type mismatches
   • Missing fields
   • Invalid states
   • Format errors

2. State Errors

   • Invalid transitions
   • Undefined states
   • State conflicts
   • Recovery handling

Best Practices

1. Type Usage

   • Always use type hints
   • Validate inputs
   • Handle optionals
   • Check states

2. Model Validation

   • Use Pydantic models
   • Define constraints
   • Handle defaults
   • Document fields