README

Milky2018/svgdiff/engine does not have a README file

#
AlignmentEvidence

Reviewable selection evidence for one Subject Alignment.

confidence remains null while confidence_status is not_calibrated; deterministic selection does not imply a calibrated probability.

#
AtomicDifference

Smallest independently reportable visual-semantic change.

Domain, evidence layers, computed relation, magnitude, and ordering remain independent so representation and measured visual impact are not conflated.

#
CauseEnvelope

Conservative Changed Fact candidates for one Difference Region.

A sound_overapproximation contains every actual changed cause within complete analysis coverage, but may contain false positives.

#
ChangedFact

One changed supported authored fact and the subjects it may affect.

#
ComparisonProfile

Rendering and normalization conditions recorded for one comparison.

The root schema 2.0 API accepts caller-provided viewport dimensions, Perceptual Background, FLIP Viewing Conditions, and FLIP error threshold, and canonicalizes the remaining fields to v1_default.

#
ComputedRelation

Relationship between the supported before and after computed facts.

status is equivalent, different, indeterminate, or not_applicable; reason_code and Diagnostic references explain why.

#
DeclaredVisualFact

One supported authored visual declaration with exact source provenance.

#
Diagnostic

Structured coverage limitation or analysis failure.

#
DiagnosticSourceLocation

One source-qualified location establishing a Diagnostic.

#
DifferenceMagnitude

Domain-appropriate magnitude vector for one Atomic Difference.

Optional components that were not computed serialize as null rather than zero.

#
DifferenceRegion

Pixel-space and CSS-space localization of one visual outcome component.

#
DomainOrdering

Versioned lexicographic components for ordering differences in one domain.

#
FeatureCoverage

Coverage of one encountered feature at each canonical evidence layer.

Layer status is covered, limited, not_applicable, or failed. Diagnostics explain every limited or failed row.

#
FlipErrorThreshold

Explicit optional reporting threshold for LDR-FLIP error values.

#
FlipViewingConditions

Explicit LDR-FLIP viewing conditions with a bounded filter domain.

#
ImpactAssessment

Additive versioned main-event assessment that preserves partial ordering.

#
ImpactDominationWitness

One deterministic explanation for excluding a dominated event.

#
ImpactFrontierGroup

One singleton or exact measured-tie group on the main-event frontier.

#
ImpactMeasurement

Common whole-canvas event measurements used by one Impact Assessment.

#
IntrinsicRasterMagnitude

Intrinsic decoded-raster comparison, independent of final SVG compositing.

#
PaintedBoundaryDisplacementMagnitude

Symmetric nearest-boundary statistics from the pinned renderer's alpha support.

#
PaintedCoverageDifferenceMagnitude

Symmetric alpha-only isolated painted-coverage comparison.

#
PerceptualBackground

One explicit opaque sRGB8 background for future perceptual compositing.

#
PerceptualColorEvidence

Availability and optional event-local perceptual-color magnitude.

#
PerceptualColorMagnitude

Displayed-color magnitude for one selected rendered outcome after explicit background compositing.

#
PerceptualFlipAreaAboveThreshold

Explicit-threshold spatial extent for one event-local LDR-FLIP response.

#
PerceptualFlipEvidence

Availability and optional event-local LDR-FLIP map.

#
PerceptualFlipMap

Deterministic quantized LDR-FLIP map for one selected rendered outcome.

#
PerceptualFlipStatistics

Separate pooled domains over one unquantized LDR-FLIP map.

#
PresenceMagnitude

Numeric footprint of inserted or deleted visual subjects.

#
RenderedEvidence

Availability and optional magnitude of canonical rendered evidence.

A computed zero magnitude is distinct from not_computed.

#
RenderedMagnitude

Continuous canonical raster measurements for one rendered outcome.

#
RendererCapabilityGap

One renderer capability gap encountered by the current comparison.

support_status is guarded or unavailable. Diagnostic references remain the authority for affected subjects and evidence layers.

#
ReportSourceSpan

Half-open UTF-16 offsets into the original SVG source.

#
ResolvedVisualFact

One supported visual fact after declaration precedence and inheritance.

#
SourceAuditDiagnostic

A bounded parse failure from one side of a source-only audit.

#
SourceAuditDifference

One changed, inserted, or deleted nonvisual source fact.

#
SourceAuditFact

One exact authored fact in the optional nonvisual source audit.

#
SourceAuditReport

An independently versioned source-only nonvisual metadata audit.

#
SourceResolutionPair

Before and after resolved facts for one subject property.

#
StructuredReport

Complete schema 2.0 result of one SVG comparison.

Read analysis_status and diagnostics before interpreting an empty atomic_differences array as profile-scoped equality.

#
SubjectAlignment

Set-to-set relationship between before and after visual subjects.

#
SubjectInstanceContext

Placement identity for a subject rendered through one or more use hosts.

The corresponding authored definition keeps its own identity and Source Spans; instance_id identifies only this rendered outcome.

#
SubjectReference

Report-local reference to one visual subject in an input SVG.

#
TransformEffectMagnitude

Tagged raw magnitude for one canonical transform-effect domain.

#
TransformResidualMagnitude

Exact singular linear transforms retained without a coefficient score.

#
TransformRotationMagnitude

Canonical paired rotation values and shortest signed angular change.

#
TransformScaleMagnitude

Canonical signed axis scales and their largest absolute paired delta.

#
TransformSkewMagnitude

Canonical paired X-skew values and shortest signed angular change.

#
TransformTranslationMagnitude

Viewport-space translation change and its Euclidean CSS-pixel magnitude.

#
VisualEvent

Agent-facing grouping of Atomic Differences around one subject outcome.

#
ComparisonInterrupted

pub(all) suberror ComparisonInterrupted {
Cancelled
CheckpointBudgetExceeded(max_checkpoints~ : Int)
} derive(Eq,
Debug
)

#
ComparisonControl

pub(all) struct ComparisonControl {
should_cancel : () -> Bool
max_checkpoints : Int?
}

#
ComparisonControl::unlimited

#
ResourceBundle

pub(all) struct ResourceBundle {
entries : Array[ResourceBundleEntry]
}

#
ResourceBundle::empty

#
ResourceBundleEntry

pub(all) struct ResourceBundleEntry {
locator : String
media_type : String
bytes : Bytes
}

#
audit_nonvisual_metadata

fn audit_nonvisual_metadata(before_svg : String, after_svg : String) ->
SourceAuditReport

#
compare_with_control

#
compare_with_control_and_resources

fn compare_with_control_and_resources(before_svg : String, after_svg : String, profile :
ComparisonProfile
, before_resources : ResourceBundle, after_resources : ResourceBundle, control : ComparisonControl) ->
StructuredReport
raise ComparisonInterrupted

#
compare_with_resources

fn compare_with_resources(before_svg : String, after_svg : String, profile :
ComparisonProfile
, before_resources : ResourceBundle, after_resources : ResourceBundle) ->
StructuredReport

Powered by MoonBit

Site sourceReport issuePackagesBuild queueSkillsStatistics

© 2026 mooncakes.io