BasePydanticAiHtmlDistiller

Base class for the built-in HTML distillers.

A distiller reduces raw HTML to a compact representation that an LLM can read cheaply. Subclasses implement distill. The base stores the prompt notes and returns them from get_prompt_notes. Override get_prompt_notes when the notes depend on several constructor arguments.

The public interface is the PydanticAiHtmlDistiller protocol. The concrete distillers are PydanticAiCleanHtmlDistiller and PydanticAiSkeletonDistiller.

Hierarchy

• BasePydanticAiHtmlDistiller
  • PydanticAiCleanHtmlDistiller

Index

Methods

• __init__
• distill
• get_prompt_notes

Methods

__init__

• __init__(*, prompt_notes): None

• Initialize a new instance.

  ---

  Parameters

  • optionalkeyword-onlyprompt_notes: str | None = None

    Short description of the final representation. Appended to the LLM task instructions by extractors. None means no notes are appended.

  Returns None

distill

• distill(html): str

• Convert raw HTML to a compact representation suitable for an LLM.

  ---

  Parameters

  • html: str

  Returns str

get_prompt_notes

• get_prompt_notes(): str | None

• Return the configured prompt notes, or None when not set.

  ---

  Returns str | None