Source:

• src/ometeotl_core/io/llm_export.py
• src/ometeotl_core/model/base.py

Local role: Provide a language-model-oriented representation that explicitly separates ontological reality from epistemic/perception-aware views.

Big-picture role: This export path implements the dedicated LLM/SLM view requirement (F-5) while preserving ownership boundaries:

• model owns canonical state (to_dict and domain classes)
• io owns formatting/projection for external consumers

Public API

world_to_llm_view(world, include_provenance=False)

Exports a world-focused LLM payload with:

• base object shape (id, type, optional sections)
• reality block
• member summaries and deterministic member lists

actor_to_llm_view(actor, perception=None, include_provenance=False)

Exports an actor with:

• ontological actor slice
• optional attached perception view
• explicit epistemic envelope

perception_to_llm_view(perception, include_provenance=False)

Exports a perception object with:

• perception payload (perceived_spaces, memberships, relations, component links)
• grouped epistemic statuses (certain, believed, hypothesis, projected, error)

ModelObject.to_llm_view()

Shared model-level entrypoint dispatching to LLMViewBuilder by object type. Unsupported object types use a generic fallback containing reality, perception, and epistemic sections.

Determinism and epistemic guarantees

• Perception payload collections are sorted with canonical keys.
• Epistemic status groups are produced deterministically.
• World member IDs (actors, spaces, resources) are sorted.
• Output explicitly exposes distinctions: reality, perception, belief, hypothesis, projection.

Validation and scope

• This exporter does not run validation.
• It assumes input objects are already structurally valid.
• It is a projection layer, not a business-logic layer.

Example:

from ometeotl_core.io.llm_export import (
    world_to_llm_view,
    actor_to_llm_view,
    perception_to_llm_view,
)

# World view: reality block + sorted member id lists
world_view = world_to_llm_view(world)
print(world_view["reality"]["actors"])

# Actor view with attached perception
actor_view = actor_to_llm_view(actor, perception=perception)
print(actor_view["epistemic"])

# Perception view: grouped by epistemic status
perc_view = perception_to_llm_view(perception)
print(perc_view["certain"])
print(perc_view["believed"])

# Via model object shortcut
llm_view = actor.to_llm_view()