1 oracle.pipelines.surface_reference_panel

oracle.pipelines.surface_reference_panel

Measure a fixed area-sampled mesh reference on the frozen 96-pose panel.

Each state is reconstructed once from its original full candidate-table prefix. The resulting cropped mesh and prepared root, current and successor clouds are retained before either sample count or seed is evaluated. This is a diagnostic new metric version; historical labels and rollout stores are never rewritten.

1.1 Functions

Name	Description
compare_prepared_triangle_replay	Prove persisted second-pass tensors reproduce first-pass scorer components.
freeze_assignment	Bind the completed stride cohort, source, rasters and surface policy.
measure_prepared_state	Score all frozen surface treatments from one persisted cloud payload.
summarize_surface_state	Compare decisions only within each named evaluator’s score units.
surface_evaluator_identity	Return the closed identity of the corrected surface metric.

1.1.1 compare_prepared_triangle_replay

oracle.pipelines.surface_reference_panel.compare_prepared_triangle_replay(
    baseline,
    scored,
)

Prove persisted second-pass tensors reproduce first-pass scorer components.

1.1.2 freeze_assignment

oracle.pipelines.surface_reference_panel.freeze_assignment(
    repo,
    output,
    stride_assignment_path,
    stride_run,
)

Bind the completed stride cohort, source, rasters and surface policy.

1.1.3 measure_prepared_state

oracle.pipelines.surface_reference_panel.measure_prepared_state(
    input_path,
    *,
    expected_hash,
    counts=COUNTS,
    anchor_count=None,
)

Score all frozen surface treatments from one persisted cloud payload.

1.1.4 summarize_surface_state

oracle.pipelines.surface_reference_panel.summarize_surface_state(
    baseline,
    surface,
)

Compare decisions only within each named evaluator’s score units.

1.1.5 surface_evaluator_identity

oracle.pipelines.surface_reference_panel.surface_evaluator_identity()

Return the closed identity of the corrected surface metric.

# oracle.pipelines.surface_reference_panel { #aria_nbv.oracle.pipelines.surface_reference_panel } `oracle.pipelines.surface_reference_panel` Measure a fixed area-sampled mesh reference on the frozen 96-pose panel. Each state is reconstructed once from its original full candidate-table prefix. The resulting cropped mesh and prepared root, current and successor clouds are retained before either sample count or seed is evaluated. This is a diagnostic new metric version; historical labels and rollout stores are never rewritten. ## Functions | Name | Description | | --- | --- | | [compare_prepared_triangle_replay](#aria_nbv.oracle.pipelines.surface_reference_panel.compare_prepared_triangle_replay) | Prove persisted second-pass tensors reproduce first-pass scorer components. | | [freeze_assignment](#aria_nbv.oracle.pipelines.surface_reference_panel.freeze_assignment) | Bind the completed stride cohort, source, rasters and surface policy. | | [measure_prepared_state](#aria_nbv.oracle.pipelines.surface_reference_panel.measure_prepared_state) | Score all frozen surface treatments from one persisted cloud payload. | | [summarize_surface_state](#aria_nbv.oracle.pipelines.surface_reference_panel.summarize_surface_state) | Compare decisions only within each named evaluator's score units. | | [surface_evaluator_identity](#aria_nbv.oracle.pipelines.surface_reference_panel.surface_evaluator_identity) | Return the closed identity of the corrected surface metric. | ### compare_prepared_triangle_replay { #aria_nbv.oracle.pipelines.surface_reference_panel.compare_prepared_triangle_replay } ```python oracle.pipelines.surface_reference_panel.compare_prepared_triangle_replay( baseline, scored, ) ``` Prove persisted second-pass tensors reproduce first-pass scorer components. ### freeze_assignment { #aria_nbv.oracle.pipelines.surface_reference_panel.freeze_assignment } ```python oracle.pipelines.surface_reference_panel.freeze_assignment( repo, output, stride_assignment_path, stride_run, ) ``` Bind the completed stride cohort, source, rasters and surface policy. ### measure_prepared_state { #aria_nbv.oracle.pipelines.surface_reference_panel.measure_prepared_state } ```python oracle.pipelines.surface_reference_panel.measure_prepared_state( input_path, *, expected_hash, counts=COUNTS, anchor_count=None, ) ``` Score all frozen surface treatments from one persisted cloud payload. ### summarize_surface_state { #aria_nbv.oracle.pipelines.surface_reference_panel.summarize_surface_state } ```python oracle.pipelines.surface_reference_panel.summarize_surface_state( baseline, surface, ) ``` Compare decisions only within each named evaluator's score units. ### surface_evaluator_identity { #aria_nbv.oracle.pipelines.surface_reference_panel.surface_evaluator_identity } ```python oracle.pipelines.surface_reference_panel.surface_evaluator_identity() ``` Return the closed identity of the corrected surface metric.