Skip to content

Fixture 002: Linear Transport Baseline

Purpose

Fixture 002 is the next sub-canonical, xZeno-canonical characterization fixture for the repo. It keeps the hardened Fixture 001 runtime and reporting pipeline but replaces the 3x5 source grid with a single centered source row so transport behavior is easier to read spatially.

Fixture 002 is not intended to be a visually dramatic departure from Fixture 001. Its purpose is to preserve the same radial field family while simplifying the source topology from a dense grid to a single centered row, improving attribution and DOE interpretability without introducing additional field-position or geometry confounders.

The intent is not to supplant Fixture 001. Fixture 001 remains the richer radial GRIN baseline. Fixture 002 exists to provide a simpler control surface for future DOE work where we want cleaner separation between source hits, rendered no-hit area, and unrendered area.

Why It Exists Relative To Fixture 001

  • Fixture 001 is the verified baseline for the current full radial-grid harness.
  • Fixture 002 keeps the same capture, verification, ledger, and reporting architecture, but reduces geometric complexity.
  • The single-row source layout should make lateral transport changes easier to interpret without introducing new runtime or reporting machinery.

Current Harness

  • Harness script: scripts/run_fixture_002.sh
  • Reporter: tools/fixture_002_report.py
  • Shared report core: tools/fixture_characterization_report.py
  • Scene: res://test-grin-basic-visual-linear-minimal.tscn
  • Fixture identity expected by launch audit: grin_basic_visual_linear_minimal
  • Fixture id: fixture_002
  • Launcher token configured by the harness: run_fixture_002

Baseline Assumptions

  • Launch only through scripts/run_fixture_002.sh.
  • Rebuild against the hardened Windows runtime mirror before each run.
  • Treat Fixture 002 as non-destructive and independent of Fixture 001 history.
  • Use the same verified artifact contract already established for Fixture 001.
  • Prefer diagnostic_flat for visual interpretation and categorical_final when explicit final-hit / rendered-no-hit / unrendered separation is needed.

Canonical Analysis Basis

  • Primary analysis basis: image-center
  • Companion analysis basis: field-relative
  • Reason: Fixture 002 keeps the transport field centered, so image center and effective field center coincide for interpretation purposes. The original radial and left/right image-center artifacts therefore remain the primary canonical basis.
  • Field-relative artifacts may still be emitted for consistency across the shared fixture reporting flow and to keep direct comparison structure aligned with later asymmetric fixtures.

Control-Validation Contract

Treat a Fixture 002 run as usable for control-surface interpretation only when all of these checks are true:

  • runtime_fingerprint_present = true
  • assembly_timestamp_present = true
  • effective_step_matches_requested = true
  • row_diagnostics_present = true
  • scheduler_clean = true

The output contract is the same as Fixture 001:

  • summary.json
  • metrics.json
  • summary.txt
  • ledger append through tools/characterization_ledger/ledger_writer.py
  • verification fields in run artifacts
  • row coverage artifacts
  • radial profile artifacts when categorical_final is used
  • radial sector profile artifacts when categorical_final is used
  • field-relative radial artifacts when categorical_final is used
  • field-relative radial sector artifacts when categorical_final is used

Planned Operating-Point Characterization Workflow

  1. Start from the default Fixture 002 harness with the linear-row geometry.
  2. Hold scene topology fixed and vary transport controls through harness env overrides such as FIXTURE_002_STEP_LENGTH, FIXTURE_002_MIN_STEP_LENGTH, FIXTURE_002_TURN_THRESHOLD, and FIXTURE_002_ERROR_TOLERANCE.
  3. Keep interpretation restricted to fully verified runs.
  4. Use diagnostic_flat for quick visual sanity checks.
  5. Use categorical_final when capture classification and radial/structured summaries are needed for comparison.

Expected Capture Modes

  • Visual mode: diagnostic_flat
  • Analysis capture mode: categorical_final
  • Alternate analysis capture mode: resolved_film

Expected Artifacts

  • capture.png if later exported manually from a selected canonical run
  • capture.png
  • analysis_capture.png
  • debug_capture.png
  • row_coverage.txt
  • row_coverage.json
  • radial_profile.txt when applicable
  • radial_profile.json when applicable
  • radial_sector_profile.txt when applicable
  • radial_sector_profile.json when applicable
  • field_radial_profile.txt when applicable
  • field_radial_profile.json when applicable
  • field_radial_sector_profile.txt when applicable
  • field_radial_sector_profile.json when applicable

For Fixture 002, radial analysis is still meaningful when we use categorical_final, because the underlying transport field remains centered and radially organized even though the source geometry is simpler.

The harness uses capture.png as the primary analysis artifact and mirrors that same image to analysis_capture.png so Fixture 002 can participate in the same comparison/reporting rhythm while still exposing the explicit artifact name requested for fixture-note work.

Best Verified Operating Point

The March 26, 2026 rerun is accepted as the refreshed post-field-fix comparison baseline for Fixture 002.

  • Timestamp: 2026-03-26T21-15-16
  • Output path: output/fixture_runs/fixture_002/2026-03-26T21-15-16
  • Status: ok
  • guard_progress exits: 0
  • forcedAdvance events: 0
  • processed_rows: 164
  • traced_pixels: 54880
  • source_hits: 2149
  • miss_hits: 52731
  • radial_overall_hit_fraction: 0.027299
  • radial_sector_left_hit_fraction: 0.030615
  • radial_sector_right_hit_fraction: 0.023984
  • turnThreshold: 4

This run is accepted even though run_verified = false against the older pre-fix comparison target, because the mismatch is attributable to the corrected r_outer / r_inner field cutoff behavior rather than to scheduler instability.

Lower-Variance Fallback

Placeholder. To be filled once repeatable verified runs establish the first stable fallback branch.

Accepted Post-Fix Comparison Baseline

  • Timestamp: 2026-03-26T21-15-16
  • Scene: res://test-grin-basic-visual-linear-minimal.tscn
  • Fixture: grin_basic_visual_linear_minimal
  • Transport: GRIN_Optical
  • Status: ok
  • Capture succeeded: true
  • guard_progress exits: 0
  • forcedAdvance events: 0
  • processed_rows: 164
  • traced_pixels: 54880
  • source_hits: 2149
  • miss_hits: 52731

This accepted comparison baseline supersedes the older pre-fix capture for future post-r_outer / r_inner validation work.

Topology Interpretation

Placeholder. Expected initial interpretation focus:

  • whether hits remain concentrated near the center row or spread asymmetrically
  • whether rendered no-hit area expands without corresponding hit conversion
  • whether unrendered area stays clearly separated from rendered outcomes