Types API

pydantic_ai_summarization.types

Type definitions for summarization-pydantic-ai.

TokenCounter = Callable[[Sequence[ModelMessage]], int] module-attribute

Function type that counts tokens in a sequence of messages.

Example

Python

def my_token_counter(messages: Sequence[ModelMessage]) -> int:
    return sum(len(str(msg)) for msg in messages) // 4

processor = SummarizationProcessor(
    model="openai:gpt-4.1",
    token_counter=my_token_counter,
)

ContextSize = ContextFraction | ContextTokens | ContextMessages module-attribute

Union type for all context size specifications.

Can be: - ("fraction", float) - fraction of max_input_tokens (requires max_input_tokens) - ("tokens", int) - absolute token count - ("messages", int) - message count

Examples:

Python

# Trigger at 80% of context window
trigger: ContextSize = ("fraction", 0.8)

# Trigger at 100k tokens
trigger: ContextSize = ("tokens", 100000)

# Trigger at 50 messages
trigger: ContextSize = ("messages", 50)

ContextFraction = tuple[Literal['fraction'], float] module-attribute

Context size specified as a fraction of max_input_tokens.

Example: ("fraction", 0.8) means 80% of max_input_tokens.

ContextTokens = tuple[Literal['tokens'], int] module-attribute

Context size specified as an absolute token count.

Example: ("tokens", 100000) means 100,000 tokens.

ContextMessages = tuple[Literal['messages'], int] module-attribute

Context size specified as a message count.

Example: ("messages", 50) means 50 messages.