Documentation
¶
Index ¶
- func ApplySemanticEvidence(baseSimilarity float64, signals1, signals2 SemanticSignals) float64
- func CompareCFGFeatures(f1, f2 *CFGFeatures) float64
- func CompareDFAFeatures(f1, f2 *dfa.DFAFeatures) float64
- func ComputeSimilarity(cfg1, cfg2 *cfg.CFG, dfa1, dfa2 *dfa.DFAInfo, config Config) float64
- type CFGFeatures
- type Config
- type SemanticSignals
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func ApplySemanticEvidence ¶
func ApplySemanticEvidence(baseSimilarity float64, signals1, signals2 SemanticSignals) float64
ApplySemanticEvidence adjusts a base (CFG/DFA-derived) similarity with AST-level counter-evidence: fully disjoint string-literal vocabularies, absence of any shared strong signal, or incompatible return categories each discount the score. Missing evidence on either side is given the benefit of the doubt.
func CompareCFGFeatures ¶
func CompareCFGFeatures(f1, f2 *CFGFeatures) float64
CompareCFGFeatures computes similarity between two CFG feature sets. Returns a value between 0.0 (completely different) and 1.0 (identical).
func CompareDFAFeatures ¶
func CompareDFAFeatures(f1, f2 *dfa.DFAFeatures) float64
CompareDFAFeatures computes similarity between two DFA feature sets. Returns a value between 0.0 (completely different) and 1.0 (identical).
func ComputeSimilarity ¶
ComputeSimilarity computes the combined CFG+DFA semantic similarity. Returns 0.0 when either CFG is nil (e.g. parse/build failures) to avoid false-positive matches on absent semantic data.
Types ¶
type CFGFeatures ¶
type CFGFeatures struct {
BlockCount int
EdgeCount int
EdgeTypeCounts map[cfg.EdgeType]int
CyclomaticNumber int
BranchingFactor float64
LoopEdgeCount int
ConditionalCount int
}
CFGFeatures holds structural features extracted from a CFG.
func ExtractCFGFeatures ¶
func ExtractCFGFeatures(c *cfg.CFG) *CFGFeatures
ExtractCFGFeatures extracts structural features from a CFG.
type Config ¶
Config holds configuration for semantic similarity computation.
func DefaultConfig ¶
func DefaultConfig() Config
DefaultConfig returns the default semantic similarity configuration.
type SemanticSignals ¶
type SemanticSignals struct {
StrongSignals map[string]struct{}
ReturnCategories map[string]struct{}
StringLiterals map[string]struct{}
}
SemanticSignals holds the language-independent semantic evidence extracted from a fragment's AST. Language adapters populate the sets: StrongSignals with kind-prefixed identifiers (e.g. "call:open", "attr:write"), ReturnCategories with return-shape categories (e.g. "none", "literal", "collection"), and StringLiterals with meaningful string constants.
func NewSemanticSignals ¶
func NewSemanticSignals() SemanticSignals
NewSemanticSignals returns an empty signal set ready to be populated.
Source Files