eric/notarius
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Use contextual descriptors for entity reconciliation

Browse Source
This commit is contained in:
Eric Rakestraw
2026-08-08 15:05:35 +00:00
parent 51d62de1f3
commit fc449863f2
18 changed files with 435 additions and 100 deletions
Download Patch File Download Diff File Expand all files Collapse all files

116
internal/modules/dnd/shared/entityreconcile/context.go
View File

@@ -22,13 +22,34 @@ type Candidate struct {
SourceRefs []source.SourceRef
}
// Selector identifies one candidate through its canonical name and source-free
// evidence ranges. It is the complete model-facing candidate descriptor.
type Selector struct {
Name string `json:"name"`
SourceRefs []SourceRange `json:"source_refs"`
}
// Clone returns an owned copy of the selector.
func (s Selector) Clone() Selector {
s.SourceRefs = cloneSourceRanges(s.SourceRefs)
return s
}
// SourceRange is a source-free evidence coordinate used in a selector.
type SourceRange struct {
StartUnitID int `json:"start_unit_id"`
EndUnitID int `json:"end_unit_id"`
}
// Materials contains owned prompt inputs and opaque candidate-key mappings.
type Materials struct {
Candidates contracts.LLMInputMaterial
Transcript contracts.LLMInputMaterial
candidateKeys []string
eligible map[string]struct{}
candidateKeys []string
eligible map[string]struct{}
keyBySelector map[string]string
collidedSelectors map[string]struct{}
}
// CandidateKeys returns all deterministic keys in candidate input order.
@@ -49,18 +70,7 @@ func (m Materials) EligibleCandidateKeys() []string {
}
type candidateInput struct {
Candidates []candidateView `json:"candidates"`
}
type candidateView struct {
Key string `json:"key"`
Name string `json:"name"`
SourceRefs []candidateSourceRef `json:"source_refs"`
}
type candidateSourceRef struct {
StartUnitID int `json:"start_unit_id"`
EndUnitID int `json:"end_unit_id"`
Candidates []Selector `json:"candidates"`
}
type transcriptInput struct {
@@ -91,8 +101,10 @@ func BuildContext(doc *source.SourceDocument, candidates []Candidate, radius int
return Materials{}, false, fmt.Errorf("build entity reconciliation context: radius must not be negative")
}
materials := Materials{
candidateKeys: make([]string, len(candidates)),
eligible: make(map[string]struct{}),
candidateKeys: make([]string, len(candidates)),
eligible: make(map[string]struct{}),
keyBySelector: make(map[string]string),
collidedSelectors: make(map[string]struct{}),
}
for index := range candidates {
key := fmt.Sprintf(candidateKeyFormat, index+1)
@@ -103,7 +115,15 @@ func BuildContext(doc *source.SourceDocument, candidates []Candidate, radius int
}
index := source.NewDocumentIndex(doc)
views := make([]candidateView, 0, len(candidates))
type preparedCandidate struct {
key string
selector Selector
intervals []sourceInterval
lookupKey string
}
prepared := make([]preparedCandidate, 0, len(candidates))
selectorCounts := make(map[string]int, len(candidates))
views := make([]Selector, 0, len(candidates))
intervals := make([]sourceInterval, 0)
cited := make([]bool, len(doc.Units))
for candidateIndex, candidate := range candidates {
@@ -112,9 +132,23 @@ func BuildContext(doc *source.SourceDocument, candidates []Candidate, radius int
continue
}
key := materials.candidateKeys[candidateIndex]
materials.eligible[key] = struct{}{}
views = append(views, candidateView{Key: key, Name: candidate.Name, SourceRefs: references})
for _, interval := range candidateIntervals {
selector := Selector{Name: candidate.Name, SourceRefs: references}
lookupKey, err := selectorLookupKey(selector)
if err != nil {
return Materials{}, false, fmt.Errorf("build entity reconciliation context: invalid candidate material")
}
prepared = append(prepared, preparedCandidate{key: key, selector: selector, intervals: candidateIntervals, lookupKey: lookupKey})
selectorCounts[lookupKey]++
}
for _, candidate := range prepared {
if selectorCounts[candidate.lookupKey] != 1 {
materials.collidedSelectors[candidate.lookupKey] = struct{}{}
continue
}
materials.eligible[candidate.key] = struct{}{}
materials.keyBySelector[candidate.lookupKey] = candidate.key
views = append(views, candidate.selector.Clone())
for _, interval := range candidate.intervals {
for position := interval.start; position <= interval.end; position++ {
cited[position] = true
}
@@ -145,24 +179,56 @@ func BuildContext(doc *source.SourceDocument, candidates []Candidate, radius int
return materials, true, nil
}
func candidateReferences(index source.DocumentIndex, refs []source.SourceRef) ([]candidateSourceRef, []sourceInterval, bool) {
func candidateReferences(index source.DocumentIndex, refs []source.SourceRef) ([]SourceRange, []sourceInterval, bool) {
if len(refs) == 0 {
return nil, nil, false
}
references := make([]candidateSourceRef, 0, len(refs))
intervals := make([]sourceInterval, 0, len(refs))
type referencedInterval struct {
reference SourceRange
interval sourceInterval
}
prepared := make([]referencedInterval, 0, len(refs))
for _, ref := range refs {
if err := index.ValidateRef(ref); err != nil {
return nil, nil, false
}
start, _ := index.Position(ref.StartUnitID)
end, _ := index.Position(ref.EndUnitID)
references = append(references, candidateSourceRef{StartUnitID: ref.StartUnitID, EndUnitID: ref.EndUnitID})
intervals = append(intervals, sourceInterval{start: start, end: end})
prepared = append(prepared, referencedInterval{reference: SourceRange{StartUnitID: ref.StartUnitID, EndUnitID: ref.EndUnitID}, interval: sourceInterval{start: start, end: end}})
}
sort.Slice(prepared, func(left, right int) bool {
if prepared[left].interval.start != prepared[right].interval.start {
return prepared[left].interval.start < prepared[right].interval.start
}
return prepared[left].interval.end < prepared[right].interval.end
})
references := make([]SourceRange, 0, len(prepared))
intervals := make([]sourceInterval, 0, len(prepared))
for _, item := range prepared {
if len(references) > 0 && references[len(references)-1] == item.reference {
continue
}
references = append(references, item.reference)
intervals = append(intervals, item.interval)
}
return references, intervals, true
}
func selectorLookupKey(selector Selector) (string, error) {
content, err := json.Marshal(selector.Clone())
if err != nil {
return "", err
}
return string(content), nil
}
func cloneSourceRanges(values []SourceRange) []SourceRange {
if len(values) == 0 {
return []SourceRange{}
}
return append([]SourceRange(nil), values...)
}
func coalesceIntervals(intervals []sourceInterval) []sourceInterval {
if len(intervals) == 0 {
return nil