Add item registry extraction and validation

This commit is contained in:
2026-08-05 19:27:49 +00:00
parent ce04387dbc
commit ad9ee076b5
19 changed files with 1204 additions and 0 deletions

View File

@@ -0,0 +1,88 @@
// Package sourcerelatedness warns when cited source text does not mention an item.
package sourcerelatedness
import (
"context"
"fmt"
"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared/diagnostics"
itemshape "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/validate/itemregistry/shape"
)
const (
Key = "extract/dnd/item-registry/source_relatedness"
WarningReasonCode = "item_not_near_source"
OmittedReasonCode = "item_relatedness_warnings_omitted"
policy = "dnd.item_registry.validator.source_relatedness.v1"
)
type Options struct{}
type Validator struct{}
var _ contracts.TypedValidator[dnd.ItemRegistry] = (*Validator)(nil)
var _ pipeline.CheckpointFingerprintProvider = (*Validator)(nil)
func New(Options) *Validator { return &Validator{} }
func (v *Validator) Name() string { return Key }
func (v *Validator) ExecutionClass() contracts.ExecutionClass {
return contracts.ExecutionClassDeterministic
}
func (v *Validator) CheckpointFingerprints() []pipeline.CheckpointFingerprint {
return []pipeline.CheckpointFingerprint{{Name: "policy", Value: policy}}
}
func (v *Validator) Validate(_ context.Context, req contracts.TypedValidationRequest[dnd.ItemRegistry]) (contracts.ValidationResult, error) {
if err := itemshape.Validate(req.Value); err != nil {
return contracts.ValidationResult{Approved: true}, nil
}
resolver, err := shared.NewCitationResolver(req.Source)
if err != nil {
return contracts.ValidationResult{Approved: true}, nil
}
citedTexts := make([]string, len(req.Value.Items))
for itemIndex, item := range req.Value.Items {
citedText, err := resolver.CitedText(item.SourceRefs)
if err != nil {
return contracts.ValidationResult{Approved: true}, nil
}
citedTexts[itemIndex] = citedText
}
warnings := make([]contracts.Warning, 0)
for itemIndex, item := range req.Value.Items {
if shared.ContainsTokenSequence(citedTexts[itemIndex], item.Name) {
continue
}
warnings = append(warnings, contracts.Warning{
Scope: fmt.Sprintf("items[%d]", itemIndex), ReasonCode: WarningReasonCode,
Message: fmt.Sprintf("Item %s was not found in cited source text", diagnostics.Quote(item.Name)),
})
}
return contracts.ValidationResult{Approved: true, Warnings: diagnostics.LimitWarnings(warnings, "items", OmittedReasonCode)}, nil
}
func Spec() pipeline.ValidatorSpec {
return pipeline.ValidatorSpec{Key: Key, ExecutionClass: contracts.ExecutionClassDeterministic}
}
func Register(registry *pipeline.ValidatorRegistry) error {
return pipeline.RegisterTypedValidatorBuilder(registry, dnd.ItemRegistryKind, Spec(), validateOptions, func(request pipeline.BuildRequest) (contracts.TypedValidator[dnd.ItemRegistry], error) {
options, err := DecodeOptions(request.Options)
if err != nil {
return nil, err
}
return New(options), nil
})
}
func DecodeOptions(options map[string]any) (Options, error) {
if err := pipeline.RejectUnknownOptions(options); err != nil {
return Options{}, err
}
return Options{}, nil
}
func validateOptions(options map[string]any) error { _, err := DecodeOptions(options); return err }

View File

@@ -0,0 +1,46 @@
package sourcerelatedness
import (
"context"
"reflect"
"testing"
"gitea.maximumdirect.net/eric/notarius/internal/core/source"
"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared/diagnostics"
)
func TestValidatorUsesOnlyCitedTranscriptText(t *testing.T) {
value := dnd.ItemRegistry{Items: []dnd.Item{{ID: "item", Name: "Star Compass", SourceRefs: []source.SourceRef{{SourceID: "session", StartUnitID: 1, EndUnitID: 1}}}}}
doc := &source.SourceDocument{ID: "session", Units: []source.SourceUnit{{ID: 1, Text: "They recover the star compass."}, {ID: 2, Text: "Unrelated text."}}}
result, err := New(Options{}).Validate(context.Background(), contracts.TypedValidationRequest[dnd.ItemRegistry]{Source: doc, Value: value})
if err != nil || !result.Approved || len(result.Warnings) != 0 {
t.Fatalf("cited match = %#v, %v; want approval without warnings", result, err)
}
value.Items[0].Name = "Glossary Relic"
before := value
result, err = New(Options{}).Validate(context.Background(), contracts.TypedValidationRequest[dnd.ItemRegistry]{Source: doc, References: contracts.ReferenceSet{Slots: map[string]contracts.ResolvedReferenceSlot{"glossary": {Items: []contracts.ReferenceItem{{Content: []byte("Glossary Relic")}}}}}, Value: value})
if err != nil || !result.Approved || len(result.Warnings) != 1 || result.Warnings[0].ReasonCode != WarningReasonCode || !reflect.DeepEqual(value, before) {
t.Fatalf("reference-only match = %#v, %v; want advisory warning", result, err)
}
}
func TestValidatorBoundsWarningsAndRegisters(t *testing.T) {
items := make([]dnd.Item, diagnostics.MaxWarnings+1)
for index := range items {
items[index] = dnd.Item{ID: "item", Name: "Missing", SourceRefs: []source.SourceRef{{SourceID: "session", StartUnitID: 1, EndUnitID: 1}}}
}
result, err := New(Options{}).Validate(context.Background(), contracts.TypedValidationRequest[dnd.ItemRegistry]{Source: &source.SourceDocument{ID: "session", Units: []source.SourceUnit{{ID: 1, Text: "Nothing here."}}}, Value: dnd.ItemRegistry{Items: items}})
if err != nil || !result.Approved || len(result.Warnings) != diagnostics.MaxWarnings || result.Warnings[len(result.Warnings)-1].ReasonCode != OmittedReasonCode {
t.Fatalf("bounded warnings = %#v, %v", result, err)
}
registry := pipeline.NewValidatorRegistry()
if err := Register(registry); err != nil {
t.Fatalf("Register() error = %v", err)
}
if got, ok := registry.Spec(Key); !ok || got != Spec() || got.ExecutionClass != contracts.ExecutionClassDeterministic {
t.Fatalf("registered spec = %#v, ok = %t", got, ok)
}
}