pipeworks_mud_mapper.services.ollama_state

State helpers for applying Ollama outputs to map data.

This module centralizes the domain logic for: - building generation metadata - applying generated descriptions to rooms - attaching or clearing validation metadata

Keeping these operations here allows callbacks to remain thin and makes unit testing easier.

Functions

build_generation_metadata(*, model, actual_seed, ...)

Build a metadata dictionary matching OllamaGenerationInfo fields.

apply_generation_to_room(*, zone_data, room_id, ...)

Return updated zone data with the generated description applied.

Module Contents

pipeworks_mud_mapper.services.ollama_state.build_generation_metadata(*, model, actual_seed, template_id, temperature, top_k, top_p, num_ctx, num_predict, target_words, system_prompt, user_prompt, generated_at=None)[source]

Build a metadata dictionary matching OllamaGenerationInfo fields.

Parameters:

  • model (str) – Ollama model identifier used for generation.

  • actual_seed (int) – Seed that was actually used (>= 0).

  • template_id (str) – Template ID used for generation.

  • temperature (float) – Temperature parameter used for generation.

  • top_k (int) – Top-K sampling parameter.

  • top_p (float) – Top-P sampling parameter.

  • num_ctx (int) – Context window size used for generation.

  • num_predict (int) – Maximum output tokens requested.

  • target_words (int) – Target word count used in the system prompt.

  • system_prompt (str | None) – Full system prompt (compiled or custom).

  • user_prompt (str) – User prompt entered by the author.

  • generated_at (str | None) – Optional ISO timestamp; if None, uses current UTC.

Returns:

Metadata dictionary for storage in map files.

Return type:

dict[str, Any]

pipeworks_mud_mapper.services.ollama_state.apply_generation_to_room(*, zone_data, room_id, description, generation_info, validation_info)[source]

Return updated zone data with the generated description applied.

Parameters:

  • zone_data (dict) – Current zone data object.

  • room_id (str) – Room ID to update.

  • description (str) – Generated description text to apply.

  • generation_info (dict[str, Any] | None) – Metadata about the LLM generation; when None, metadata is cleared.

  • validation_info (dict[str, Any] | None) – Validator metadata to attach; when None, validation info is cleared.

Returns:

Updated zone data with modified room data.

Return type:

dict

Raises:

KeyError – If room_id is not in the zone data.