​
Chat Integration

​
Processing Chat Conversations

import epistemic_me
from epistemic_me.models import DialecticType

def process_chat_conversation(self_model_id: str, conversation: list[dict]):
    """
    Process a chat conversation to build a belief system.
    
    Args:
        self_model_id: ID of the self model
        conversation: List of message dictionaries with 'role' and 'content' keys
    """
    # Create initial dialectic
    dialectic = epistemic_me.Dialectic.create(
        self_model_id=self_model_id,
        dialectic_type=DialecticType.DEFAULT
    )

    # Track created beliefs during the conversation
    created_beliefs = []
    last_dialectic_belief_system = None

    # Process messages and create question/answer pairs
    for message in conversation:
        if message["role"] == "assistant":
            # Handle assistant messages as questions
            dialectic.update(question=message["content"])
        else:
            # Handle user messages as answers
            update_resp = dialectic.update(answer=message["content"])
            
            # Track beliefs from the response
            if update_resp?.dialectic?.beliefSystem:
                for belief in update_resp.dialectic.beliefSystem.beliefs:
                    if belief?.id:
                        created_beliefs.append(belief.id)
                last_dialectic_belief_system = update_resp.dialectic.beliefSystem

    return created_beliefs, last_dialectic_belief_system

​
Example Usage

# Example conversation
conversation = [
    {"role": "assistant", "content": "How would you describe your sleep habits?"},
    {"role": "user", "content": "I usually get about 8 hours of sleep per night"},
    {"role": "assistant", "content": "What time do you typically go to bed?"},
    {"role": "user", "content": "Around 10:30 PM most nights"}
]

# Process the conversation
beliefs, belief_system = process_chat_conversation(
    self_model_id="user123",
    conversation=conversation
)

print(f"Created beliefs: {beliefs}")
print(f"Final belief system: {belief_system}")