docs: pin metrics compute layer designation

[erinepshovel-code]

GPT generated; context, prompt Erin Spencer

Summary

Resolves the edcm_metrics.py / metrics compute layer-designation question as A — Behavioral/orchestration, not Operator L0.

This PR updates:

• backend/src/edcmbone/metrics/compute.py
• tests/test_metrics_layer_designation.py

Decision

metrics/compute.py is not the bones-only Operator layer. It is an upper-layer / orchestration module that currently lives in edcmbone for compatibility and is marked for future migration to edcm.

The L0 Operator entry point remains separate:

core/operator/operator_extractor.py

Why A

Existing canon and migration docs already say:

• edcmbone is canonical L0 only
• metrics/compute.py orchestrates upper-layer computation and should migrate upstream
• the 9 Behavioral metrics are not the L0 Operator vector
• Operator output comes from core/operator/operator_extractor.py

So this is documentation / layer-contract debt, not a rewrite-to-bones bug.

New layer constants

Added to edcmbone.metrics.compute:

LAYER_DECISION = "A_BEHAVIORAL_ORCHESTRATION"
LAYER_DESIGNATION = "L1_L2_L3_ORCHESTRATOR_PENDING_EDCM_MIGRATION"
LAYER_INPUT_SUBSTRATE = "round_text_tokens_marker_stats_prior_round_context"
OPERATOR_LAYER_SUBSTRATE = "bones_only"
OPERATES_ON_OPERATOR_BONES = False
CONSUMES_BONE_COUNT_FOR_AUDIT = True
MIGRATION_TARGET = "edcm"
OPERATOR_ENTRYPOINT = "core.operator.operator_extractor"

Test enforcement

Adds tests that fail loudly if future edits blur the layer again:

• metrics compute is Behavioral/orchestration, not Operator L0
• Operator entry point remains separate
• RoundMetrics remains an upper-layer vector shape while carrying bone_count only for audit continuity

hmmm

[Copilot]

Pull request overview

This PR clarifies the architectural/layer contract for backend/src/edcmbone/metrics/compute.py, explicitly designating it as Behavioral/orchestration (not Operator L0), and adds a regression test to prevent future drift of that designation.

Changes:

• Added an explicit “Layer designation” section to the edcmbone.metrics.compute module docstring.
• Introduced module-level “layer contract” constants (decision, substrate, migration target, operator entrypoint).
• Added tests to enforce the layer boundary and to assert RoundMetrics exposes an upper-layer vector shape while carrying bone_count only as audit metadata.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
backend/src/edcmbone/metrics/compute.py	Adds layer designation documentation and a set of constants encoding the layer contract; minor signature formatting.
tests/test_metrics_layer_designation.py	Adds tests that lock the compute module’s layer decision/entrypoint separation and the RoundMetrics vector shape.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.