Schemas Reference

Calibrex ships JSON Schema definitions for every config, result, and report sidecar artifact. Schemas live in schemas/ at the repository root and are regenerated with:

calibrex schema all --output-dir schemas

The same command writes schema_ledger.json, a machine-readable ledger of every artifact kind: its current and accepted schema_version values, schema file and SHA-256, whether it is digest-bound, and the versions that were retired (with the remedy) or are upgraded by a migration. calibrex schema ledger prints it. A released schema_version remains readable: calibrex validate either reads it directly, upgrades it through a lossless migration, or explains which release retired it and what to do instead.

Schema	Description
assessment.schema.json	AssessmentArtifact — machine-readable falsification assessment (assessment.json).
comparison.schema.json	ResultComparison — machine-readable comparison between two Calibrex results (comparison.json).
benchmark_definition.schema.json	BenchmarkDefinition — shared method × split trial matrix before aggregation.
benchmark.schema.json	BenchmarkArtifact — N-way failure rates, distributions, rankings, paired bootstrap comparisons, and provenance.
calibration_ci.schema.json	CalibrationCIArtifact — digest-bound CI inputs, falsification and compatibility gates, artifact paths, and execution provenance.
calibration_ci_lifecycle.schema.json	CalibrationCILifecycleArtifact — what one Calibration CI decision did to a lifecycle registry edge: action, digest-bound CI decision and candidate snapshot, appended events, incumbent before/after, and registry verification.
calibration_lifecycle.schema.json	CalibrationLifecycleArtifact — incumbent/candidate adoption timeline with evidence windows, weak-observability rejection, and rollback provenance.
config.schema.json	CalibrationConfig — schema for calibrex calibrate configuration files.
dataset_manifest.schema.json	DatasetManifest — portable dataset manifest independent of ROS message classes.
remote_archive_selection.schema.json	RemoteArchiveSelectionArtifact — HTTP archive identity, endpoint-inclusive frame selection, ZIP member CRC/size, materialized file hashes, dataset license, and downloader provenance.
doctor.schema.json	DoctorArtifact — environment, dataset readiness, workflow suggestions, and generation provenance from calibrex doctor.
depth_provider.schema.json	DepthProviderArtifact — model/environment identity and digest-pinned depth-map observations.
continuous_time_camera_lidar_problem.schema.json	ContinuousTimeCameraLidarProblemArtifact — dataset license, per-point firing times, exposure/readout times, probabilistic correspondences, initial and optional reference extrinsic/clock state, and either supplied body twists or a digest-pinned piecewise SE(3) trajectory.
continuous_time_camera_lidar_result.schema.json	ContinuousTimeCameraLidarResultArtifact — joint extrinsic/clock estimate, optional reference errors, selected trajectory model, complete candidate trace, observability, disjoint holdout evidence, and source digest.
continuous_time_trajectory.schema.json	ContinuousTimeTrajectoryContract — typed T_world_body knots with the interpolation model, knot-domain validity, clock/capture-time semantics, and provenance; converts to the piecewise adapter.
continuous_time_trajectory_measurements.schema.json	ContinuousTimeTrajectoryMeasurements — pose, point, point-to-plane, IMU pre-integration, lever-arm, IMU clock-offset, accelerometer-bias, and gravity fields constraining a trajectory fit, with frame semantics and unique IDs.
continuous_time_trajectory_fit.schema.json	ContinuousTimeTrajectoryFitResultArtifact — sparse manifold fit result with fitted knots, initial values, convergence evidence, per-type RMSE, optional gyro bias, lever arm, IMU clock offset, accelerometer bias, gravity, and input digests.
continuous_time_lidar_point_to_plane.schema.json	ContinuousTimeLidarPointToPlaneArtifact — synthetic knot-path point-to-plane recovery with disjoint temporal holdout and a signed known-bad control.
continuous_time_imu_preintegration.schema.json	ContinuousTimeImuPreintegrationArtifact — synthetic knot-path gyro pre-integration recovery with shared bias, holdout, and a signed known-bad control.
continuous_time_imu_lever_arm.schema.json	ContinuousTimeImuLeverArmArtifact — synthetic knot-path IMU lever-arm recovery with holdout and a signed known-bad control.
continuous_time_imu_clock_offset.schema.json	ContinuousTimeImuClockOffsetArtifact — synthetic knot-path IMU clock-offset recovery with holdout and a signed known-bad control.
continuous_time_imu_accel_bias.schema.json	ContinuousTimeImuAccelBiasArtifact — synthetic knot-path accelerometer-bias and gravity recovery with holdout and a signed known-bad control.
continuous_time_imu_intrinsics.schema.json	ContinuousTimeImuIntrinsicsArtifact — synthetic knot-path diagonal gyro and accelerometer scale recovery with holdout and a signed known-bad control.
continuous_time_lidar_imu_joint.schema.json	ContinuousTimeLidarImuJointArtifact — synthetic joint LiDAR-IMU recovery (IMU mounting rotation, gyro bias, lever arm, clock offset, accelerometer bias, gravity) from LiDAR point-to-plane knots, with interleaved holdout, signed per-axis controls and block separability.
lidar_imu_bag_calibration.schema.json	LidarImuBagCalibrationArtifact — joint B-spline LiDAR-IMU calibration of one bag segment (calibrex imu-lidar joint): T_lidar_imu, clock offset, biases and gravity with formal std, map-round convergence, the angular-rate starting estimate, held-out IMU controls, separability and a PASS/WARN/FAIL/INCONCLUSIVE policy.
continuous_time_sliding_window.schema.json	ContinuousTimeSlidingWindowArtifact — synthetic sliding-window Schur marginalization with batch consistency, holdout, and a skip-marginalization control.
empirical_se3_uncertainty.schema.json	EmpiricalSe3UncertaintyArtifact — deterministic temporal-block-resampled SE(3) tangent-space intervals, observed coverage against injected truth, an overconfident interval control, and the PASS/WARN/FAIL/INCONCLUSIVE policy decision.
probabilistic_correspondence.schema.json	ProbabilisticCorrespondenceArtifact — dataset family/sequence, separate dataset/provider license lineage, plus 2D–3D means, covariance, outlier probability, and reliability.
probabilistic_pnp_result.schema.json	ProbabilisticPnpResultArtifact — robust pose output, uncertainty-aware diagnostics, tool identity, and immutable input provenance.
probabilistic_refinement_result.schema.json	ProbabilisticRefinementResultArtifact — v0.2 separates the optimized candidate from the selected output, records fit-only acceptance/initializer rollback, keeps holdout out of selection, and retains reference errors, ablations, trace, and input digests.
camera_lidar_problem.schema.json	CameraLidarCalibrationProblem — fixed depth/LiDAR observations, transforms, bounds, splits, and digests.
camera_lidar_correspondence_export.schema.json	CameraLidarCorrespondenceExportManifest — KITTI/KITTI-360 provider NPZ exports, sensor intrinsics, dataset/provider licenses, and frame-level digests before conversion to probabilistic correspondences.
camera_lidar_confidence_calibration.schema.json	CameraLidarConfidenceCalibrationArtifact — development-only confidence sweep, geometric train/holdout support gates, immutable runtime options when locked, and explicit rejection otherwise.
camera_lidar_provider_support_comparison.schema.json	CameraLidarProviderSupportComparisonArtifact — like-for-like development provider ranking with exact protocol/source digests, Pareto diagnostics, and a hard prohibition on selecting rejected calibrations.
camera_lidar_pose_initializer_protocol.schema.json	CameraLidarPoseInitializerBenchmarkProtocol — frozen external-provider identity, camera stream, frame IDs, left-camera-frame SE(3) perturbations, strict hit rule, isolation declaration, metrics, and provenance.
camera_lidar_pose_initializer_failure_analysis.schema.json	CameraLidarPoseInitializerFailureAnalysis — digest-bound required/predicted correction vectors, per-frame errors, signed parallel response, orthogonal leakage, non-causal findings, and complete source provenance.
camera_lidar_initializer_calibration.schema.json	CameraLidarInitializerCalibrationArtifact — development-only aggregate PnP confidence/RANSAC recovery grid and immutable initializer lock.
camera_lidar_correspondence_quality.schema.json	CameraLidarCorrespondenceQualityArtifact — result-bound per-frame confidence/projection gates, image coverage, residual distributions, scale-normalized six-DoF information, weak axes, support grades, and provenance.
camera_lidar_failure_analysis.schema.json	CameraLidarFailureAnalysisArtifact — digest-bound D2D/refinement cases, selected and rejected-candidate errors, initializer rollback, p90 errors, holdout degradation, and explicitly non-causal findings.
camera_lidar_sota_audit_protocol.schema.json	CameraLidarSotaAuditProtocol — immutable category, numerical gates, coverage requirements, artifact locators, and digests.
camera_lidar_sota_audit_result.schema.json	CameraLidarSotaAuditResult — requirement-by-requirement achieved/contradicted/missing evidence and supported/refuted/incomplete claim verdict.
gnss_lidar_lever_arm.schema.json	GnssLidarLeverArmArtifact — GNSS antenna lever arm in the LiDAR frame and clock offset, with analytic and window-jackknife std, held-out known-bad controls, and reference differences.
imu_lidar_rotation.schema.json	ImuLidarRotationArtifact — rotation of T_lidar_imu, clock offset, and gyro bias from angular rates, with analytic and window-jackknife std, held-out known-bad controls, and design-reference differences.
imu_lidar_translation.schema.json	ImuLidarTranslationArtifact — IMU lever arm (translation of T_lidar_imu) from the accelerometer, built on a digest-pinned rotation artifact, with analytic, window-jackknife, and segment-sensitivity std, held-out known-bad controls, and design-reference differences.
ins_lidar_hand_eye.schema.json	InsLidarHandEyeArtifact — joint INS/GNSS-LiDAR trajectory hand-eye estimate with per-DoF analytic and block-jackknife std, estimated/prior/unobservable status, held-out known-bad controls, and reference-calibration differences.
lidar_lidar_extrinsic.schema.json	LidarLidarExtrinsicArtifact — T_reference_target from map registration, with analytic and block-jackknife std, held-out known-bad controls, the held-out cost of a reference transform, and per-DoF reference differences.
vehicle_frame_rotation.schema.json	VehicleFrameRotationArtifact — sensor-to-vehicle rotation from non-holonomic motion, with analytic and block-jackknife std, held-out known-bad controls, and reference differences with their held-out cost.
gnss_imu_lever_arm.schema.json	GnssImuLeverArmArtifact — GNSS antenna in the IMU frame composed from digest-pinned GNSS-LiDAR and IMU-LiDAR artifacts, with propagated std and inherited observability.
lidar_wheel_odometry.schema.json	LidarWheelOdometryArtifact — LiDAR-to-vehicle rotation, wheel-speed scale, clock offset, and lever against wheel odometry, with analytic and block-jackknife std, held-out known-bad controls, and references.
camera_focal_scale.schema.json	CameraFocalScaleArtifact — target-free camera focal-length check from camera/gyro rate ratios about each camera axis, with held-out ratios, jackknife std, and the optical-axis control.
camera_lidar_edge.schema.json	CameraLidarEdgeArtifact — targetless camera-LiDAR T_camera_lidar by image-edge / LiDAR depth-edge alignment, with per-DoF jackknife std, held-out known-bad controls, the objective at the start and the estimate, and the synchronized frames used.
calibration_check.schema.json	CalibrationCheckArtifact — audit of the calibration deployed on a robot: candidate sources, topic roles, frame tree, and one record per sensor pair (Phase A: plan only).
check_frames.schema.json	CheckFramesFile — hand-written static frame tree accepted by calibrex check --tf.
bag_estimate.schema.json	BagEstimateArtifact — calibration estimated from a bag with no calibration (calibrex estimate): per pair the estimated transform, each axis with its std and whether the data observed it, the frames exported to a check_frames tree and those omitted, estimator settings, evidence, bag digest and provenance. calibrex estimate --html renders it as a self-contained report.
calibration_drift.schema.json	CalibrationDriftArtifact — whether a rig's calibration changed between recordings (calibrex drift): the per-bag estimate references and digests, per pair and axis the estimates with std, pairwise differences, chi-square homogeneity test, minimum detectable change, the verdict (stable / drift / inconclusive) with the deviating bag and the size of the change, thresholds and provenance.
sota_audit_protocol.schema.json	SotaAuditProtocol — pair-agnostic claim scope (modalities, estimated quantities, category), numerical gates, coverage requirements, and evidence digests.
sota_audit_result.schema.json	SotaAuditResult — requirement-by-requirement evidence status and the supported/refuted/incomplete verdict for any sensor pair.
sota_leaderboard.schema.json	SotaLeaderboard — per-pair standings copied from digest-pinned audit results, including target pairs without a claim.
camera_lidar_benchmark_protocol.schema.json	CameraLidarBenchmarkProtocol — immutable frame sampling, perturbations, hit rule, isolation, and metrics.
calibration_candidate_trace.schema.json	CalibrationCandidateTrace — every objective evaluation, accepted update, stopping reason, output, and independent hit decision.
bullseye_plot.schema.json	BullseyePlotArtifact — digest-pinned three-plane rotation recovery visualization and all plotted trial points.
kitti_benchmark_input.schema.json	KITTIBenchmarkInputManifest — digest-locked official KITTI raw frame, timestamp, and calibration inputs.
evidence_bundle.schema.json	EvidenceBundleManifest — digest manifest for report and evidence artifacts (bundle.json).
evidence_bundle_verification.schema.json	EvidenceBundleVerification — machine-readable verification result for an evidence bundle (verification.json).
external_run.schema.json	ExternalCalibrationRunArtifact — digest-bound external execution/import identity, conventions, isolation, status, parsed outputs, and provenance (external-run.json).
camera_imu_motion_recording.schema.json	CameraImuMotionRecording — camera orientations and IMU gyro samples on their own clocks, used only to evaluate an imported Camera--IMU candidate.
external_camera_imu_evidence.schema.json	ExternalCameraImuEvidenceArtifact — PASS/WARN/FAIL/INCONCLUSIVE/BLOCKED holdout angular-rate evidence for an imported T_cam_imu/time shift with signed rotation/time controls and digest-bound provenance.
koide_runner_config.schema.json	KoideRunnerConfig — typed argv stages, commercial/research policy, container isolation, frame bindings, and reproducibility inputs for the Koide adapter.
koide_execution_lock.schema.json	KoideExecutionLock — self-digesting commercial handoff pinning the upstream source commit, published image digest, manual initialization stages, and unresolved base-image rebuild requirement.
koide_readiness.schema.json	KoideReadinessArtifact — digest-bound camera/LiDAR input checks, synchronization, intrinsics, frame binding, texture, geometry, motion, dynamic-scene risk, and profile guidance before an external Koide run.
koide_pilot.schema.json	KoidePilotArtifact — frozen external candidate acceptance with digest-bound input/readiness/config provenance, independent KITTI holdout projection/edge/depth metrics, twelve signed 6-DoF controls, and an explicit PASS/WARN/FAIL/INCONCLUSIVE/BLOCKED adoption decision.
koide_real_pilot_request.schema.json	KoideRealPilotRequest — non-executing official-data handoff with source/image pins, archive metadata state, fitting/holdout IDs/digests, exact argv/environment/mount policy, native output contract, known-bad controls, and KPI thresholds.
koide_real_pilot_verification.schema.json	KoideRealPilotVerification — fail-closed evidence verification for archive, raw inventory, readiness, native calib.json, external-run identity, execution logs/environment, isolation, and test-vs-official label.
koide_real_pilot_finalization.schema.json	KoideRealPilotFinalization — terminal BLOCKED/TEST_ONLY/PASS/FAIL decision requiring independent holdout KPIs and all twelve known-bad controls; only official executed PASS/ADOPT may claim an official result.
kitti_falsification.schema.json	KITTIFalsificationBenchmarkArtifact — pinned KITTI frames, source metadata, input/output digests, frozen protocol, reference/known-bad trial decisions, and provenance (benchmark.json).
online_timeline.schema.json	OnlineCalibrationTimelineArtifact — per-batch evidence timeline for an online/streaming calibration session (timeline.json).
trajectory.schema.json	TrajectoryArtifact — schema-validated odometry trajectory and ground-truth-free trajectory quality evidence.
trajectory_window_drift.schema.json	TrajectoryWindowDriftArtifact — per-capture-window odometry/map consistency evidence and local-vs-long-span interpretation.
capture_readiness.schema.json	CaptureReadinessArtifact — motion excitation, plane-normal diversity, estimated observable DoF, and actionable recapture guidance before calibration.
capture_manifest.schema.json	CaptureManifest — digest-bound capture/session/vehicle/sensor-kit identity, source and stream inventory, cross-stream frame/clock bindings, and ready/warn/blocked intake evidence.
lifecycle_registry.schema.json	RegistryManifest — filesystem-first v0.2 lifecycle registry identity and provenance.
lifecycle_event.schema.json	LifecycleEvent — append-only register/install/capture/evaluate/hold/promote/reject/rollback/remove event with a SHA-256 hash chain.
lifecycle_evaluation.schema.json	LifecycleEvaluationArtifact — fail-closed candidate admission with exact capture/result/evidence/assessment/promotion/smoke digests.
lifecycle_registry_state.schema.json	LifecycleRegistryState — replayed vehicle, sensor-kit, physical-sensor, edge, and capture projection.
lifecycle_head.schema.json	RegistryHead — event/state digests and checkpoint sequence.
lifecycle_verification.schema.json	RegistryVerificationArtifact — typed chain/source verification report.
lifecycle_status.schema.json	LifecycleRegistryStatus — verified registry counts and active-edge/sensor summary.
capture_manifest_verification.schema.json	CaptureManifestVerification — fail-closed self-digest and source/config path revalidation report emitted by calibrex capture verify.
mcap_integrity.schema.json	McapIntegrityEvidence — bounded official MCAP framing, link, Summary/Index, and directly verifiable CRC evidence emitted during calibrex capture inspect --type mcap.
continuous_time_lidar_pair_result.schema.json	ContinuousTimeLidarPairArtifact — extrinsic re-optimization while profiling a scalar LiDAR clock offset on a fixed piecewise-SE(3) trajectory.
continuous_time_lidar_train_diagnostics.schema.json	ContinuousTimeLidarTrainDiagnostics — train-only residual distributions, range bins, MAD rejection, clock-profile trace, and provenance.
continuous_time_lidar_ablation.schema.json	ContinuousTimeLidarAblationManifest — digest-bound uniform-vs-adaptive voxel and outlier-policy comparison under shared capture, holdout, and solver-budget controls.
solid_state_cross_dataset_benchmark_config.schema.json	SolidStateCrossDatasetBenchmarkSpec v0.2 — public-dataset declarations, split/seed replicate protocol, reference/identity-control labels, and bootstrap settings.
solid_state_cross_dataset_benchmark.schema.json	SolidStateCrossDatasetBenchmarkManifest v0.2 — digest-bound replicate table, failure categories, adaptive/uniform holdout comparison, confidence intervals, and provenance.
solid_state_failure_analysis.schema.json	SolidStateFailureAnalysisManifest — digest-bound per-replicate diagnostics and explicitly caveated counterexample findings.
solid_state_synthetic_benchmark.schema.json	SolidStateSyntheticBenchmarkArtifact — absolute extrinsic/clock error gates, deterministic known-bad control, and generator provenance.
solid_state_metrology_evaluation.schema.json	SolidStateMetrologyEvaluationArtifact — per-session independent physical reference, remount-session, downstream holdout, uncertainty, absolute-error, and evidence-integrity gates.
policy.schema.json	PolicyArtifact — machine-readable assessment policy declarations for one run (policy.json).
protocol.schema.json	ProtocolArtifact — machine-readable evidence protocol declarations for one run (protocol.json).
report_comparison.schema.json	ReportComparison — machine-readable N-way comparison across labeled Calibrex results (report_comparison.json).
report_degeneracy.schema.json	ReportDegeneracyArtifact — schema for degeneracy.json report sidecars.
report_evidence.schema.json	ReportEvidenceArtifact — schema for evidence.json report sidecars.
report_metrics.schema.json	ReportMetricsArtifact — schema for metrics.json report sidecars.
report_observability.schema.json	ReportObservabilityArtifact — schema for observability.json report sidecars.
report_summary.schema.json	ReportSummaryArtifact — schema for summary.json report sidecars.
result.schema.json	CalibrationResult — schema for result.yaml, the primary calibration output.
multi_lidar_service_plan.schema.json	MultiLidarServicePlan — digest-bound vehicle/sensor identities, baseline and candidate edge artifacts, exact mutable incident edge IDs, and read-only downstream references.
multi_lidar_service_evaluation.schema.json	MultiLidarServiceEvaluation — fail-closed READY/HOLD decision with replay gates, edge drift, graph connectedness, cycle closure, and provenance.
camera_imu_service_plan.schema.json	CameraImuServicePlan — digest-bound old/new camera and IMU identities, explicit evidence roles, metric observations, budgets, and mutable-component allowlist.
camera_imu_service_evaluation.schema.json	CameraImuServiceEvaluation — fail-closed READY/HOLD replay, evidence, binding, metric-budget, and component-scope decision with provenance.
radar_service_plan.schema.json	RadarServicePlan — digest-bound old/new radar identities, captured RadarScan and cross-modal evidence roles, explicit Doppler/frame conventions, budgets, and mutable-component allowlist.
radar_service_evaluation.schema.json	RadarServiceEvaluation — fail-closed READY/HOLD replay, evidence, convention, metric-budget, binding, and component-scope decision with provenance.
autoware_export.schema.json	AutowareExportArtifact — Autoware sensor-kit calibration mapping, exact frame/inverse records, static-TF commands, and digest-bound provenance.
autoware_promotion.schema.json	AutowarePromotionArtifact — package-aware PASS/WARN/FAIL/BLOCKED admission, baseline/candidate digests, deterministic patch inventory, and atomic rollback evidence.
autoware_smoke.schema.json	AutowareSmokeArtifact — digest-bound downstream xacro/build/TF/launch/perception evidence with argv, output digests/tails, isolation policy, and provenance.
solid_state_context.schema.json	SolidStateLidarCalibrationContext — acquisition semantics, capture windows, and provenance for solid-state LiDAR.
livox_time_ablation.schema.json	LivoxTimeAblationManifest — digest-bound deskew/point-time validation variants and their result evidence.
transforms.schema.json	TransformArtifact — machine-readable transform estimates for one calibration run (transforms.json).