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 +1,97 @@
// Package shape validates the required extracted item fields.
package shape
import (
"context"
"fmt"
"strings"
"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"
)
const (
Key = "extract/dnd/item-registry/shape"
ReasonCode = "invalid_item_shape"
policy = "dnd.item_registry.validator.shape.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) {
issues := issuesFor(req.Value)
if len(issues) > 0 {
return rejection(diagnostics.Aggregate("invalid item shape", issues)), nil
}
return contracts.ValidationResult{Approved: true}, nil
}
func Validate(value dnd.ItemRegistry) error {
issues := issuesFor(value)
if len(issues) == 0 {
return nil
}
return fmt.Errorf("%s", diagnostics.Aggregate("invalid item shape", issues))
}
func issuesFor(value dnd.ItemRegistry) []string {
if value.Items == nil {
return []string{"items must be present"}
}
issues := make([]string, 0)
for index, item := range value.Items {
prefix := fmt.Sprintf("items[%d]", index)
if strings.TrimSpace(item.ID) == "" {
issues = append(issues, prefix+".id must not be empty")
}
if strings.TrimSpace(item.Name) == "" {
issues = append(issues, prefix+".name must not be empty")
}
if len(item.SourceRefs) == 0 {
issues = append(issues, prefix+".source_refs must not be empty")
}
}
return issues
}
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 }
func rejection(message string) contracts.ValidationResult {
return contracts.ValidationResult{Approved: false, ReasonCode: ReasonCode, Message: message}
}

View File

@@ -0,0 +1,44 @@
package shape
import (
"context"
"reflect"
"strings"
"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"
)
func TestValidatorRejectsMalformedItemsWithoutMutation(t *testing.T) {
value := dnd.ItemRegistry{Items: []dnd.Item{{Name: "", SourceRefs: []source.SourceRef{}}}}
before := value
result, err := New(Options{}).Validate(context.Background(), contracts.TypedValidationRequest[dnd.ItemRegistry]{Value: value})
if err != nil || result.Approved || result.ReasonCode != ReasonCode || !strings.Contains(result.Message, "id must not be empty") || !strings.Contains(result.Message, "source_refs must not be empty") || !reflect.DeepEqual(value, before) {
t.Fatalf("Validate() = %#v, %v; want non-mutating shape rejection", result, err)
}
result, err = New(Options{}).Validate(context.Background(), contracts.TypedValidationRequest[dnd.ItemRegistry]{Value: dnd.ItemRegistry{}})
if err != nil || result.Approved || !strings.Contains(result.Message, "items must be present") {
t.Fatalf("missing items = %#v, %v; want rejection", result, err)
}
}
func TestValidatorApprovesAndRegisters(t *testing.T) {
value := dnd.ItemRegistry{Items: []dnd.Item{{ID: "item", Name: "Rope", SourceRefs: []source.SourceRef{{SourceID: "session", StartUnitID: 1, EndUnitID: 1}}}}}
result, err := New(Options{}).Validate(context.Background(), contracts.TypedValidationRequest[dnd.ItemRegistry]{Value: value})
if err != nil || !result.Approved {
t.Fatalf("Validate() = %#v, %v; want approval", 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)
}
if _, err := DecodeOptions(map[string]any{"unexpected": true}); err == nil {
t.Fatal("DecodeOptions() accepted unknown options")
}
}

View File

@@ -0,0 +1,101 @@
// Package sourcerefs validates item citations against the current source.
package sourcerefs
import (
"context"
"fmt"
"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"
itemshape "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/validate/itemregistry/shape"
)
const (
Key = "extract/dnd/item-registry/source_refs"
ReasonCode = "invalid_item_source_refs"
policy = "dnd.item_registry.validator.source_refs.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 req.Stage == string(pipeline.StageExtract) && req.Chunk == nil {
return contracts.ValidationResult{}, fmt.Errorf("item source-reference validator requires the current extraction chunk")
}
if err := itemshape.Validate(req.Value); err != nil {
return contracts.ValidationResult{Approved: true}, nil
}
index := source.NewDocumentIndex(req.Source)
issues := make([]string, 0)
for itemIndex, item := range req.Value.Items {
for refIndex, ref := range item.SourceRefs {
if err := index.ValidateRef(ref); err != nil {
issues = append(issues, fmt.Sprintf("items[%d].source_refs[%d]: %s", itemIndex, refIndex, diagnostics.Truncate(err.Error())))
continue
}
if req.Stage == string(pipeline.StageExtract) && !chunkContainsRef(req.Chunk, ref) {
issues = append(issues, fmt.Sprintf("items[%d].source_refs[%d]: source reference is outside the current extraction chunk", itemIndex, refIndex))
}
}
}
if len(issues) == 0 {
return contracts.ValidationResult{Approved: true}, nil
}
return rejection(diagnostics.Aggregate("invalid item source references", issues)), nil
}
func chunkContainsRef(chunk *source.Chunk, ref source.SourceRef) bool {
if chunk == nil || ref.SourceID != chunk.SourceID {
return false
}
startFound := false
endFound := false
for _, unit := range chunk.Units {
startFound = startFound || unit.ID == ref.StartUnitID
endFound = endFound || unit.ID == ref.EndUnitID
}
return startFound && endFound
}
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 }
func rejection(message string) contracts.ValidationResult {
return contracts.ValidationResult{Approved: false, ReasonCode: ReasonCode, Message: message}
}

View File

@@ -0,0 +1,60 @@
package sourcerefs
import (
"context"
"reflect"
"strings"
"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"
)
func TestValidatorRejectsInvalidAndOutOfChunkReferencesWithoutMutation(t *testing.T) {
value := validItemRegistry()
value.Items[0].SourceRefs = []source.SourceRef{{SourceID: "foreign", StartUnitID: 1, EndUnitID: 1}}
before := value
result, err := New(Options{}).Validate(context.Background(), request(validDocument(), value))
if err != nil || result.Approved || result.ReasonCode != ReasonCode || !strings.Contains(result.Message, "items[0].source_refs[0]") || !reflect.DeepEqual(value, before) {
t.Fatalf("Validate() = %#v, %v; want source-reference rejection", result, err)
}
doc := &source.SourceDocument{ID: "session", Units: []source.SourceUnit{{ID: 1}, {ID: 2}}}
value = validItemRegistry()
value.Items[0].SourceRefs[0].EndUnitID = 2
result, err = New(Options{}).Validate(context.Background(), contracts.TypedValidationRequest[dnd.ItemRegistry]{Stage: string(pipeline.StageExtract), Source: doc, Chunk: &source.Chunk{SourceID: "session", Units: []source.SourceUnit{{ID: 1}}}, Value: value})
if err != nil || result.Approved || !strings.Contains(result.Message, "outside the current extraction chunk") {
t.Fatalf("out-of-chunk evidence = %#v, %v", result, err)
}
}
func TestValidatorDefersMalformedShapeAndRegisters(t *testing.T) {
result, err := New(Options{}).Validate(context.Background(), request(nil, dnd.ItemRegistry{Items: []dnd.Item{{Name: "Missing"}}}))
if err != nil || !result.Approved {
t.Fatalf("malformed shape = %#v, %v; want deferral", result, err)
}
_, err = New(Options{}).Validate(context.Background(), contracts.TypedValidationRequest[dnd.ItemRegistry]{Stage: string(pipeline.StageExtract), Source: validDocument(), Value: validItemRegistry()})
if err == nil || !strings.Contains(err.Error(), "requires the current extraction chunk") {
t.Fatalf("missing chunk error = %v", 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)
}
}
func request(doc *source.SourceDocument, value dnd.ItemRegistry) contracts.TypedValidationRequest[dnd.ItemRegistry] {
return contracts.TypedValidationRequest[dnd.ItemRegistry]{Source: doc, Value: value}
}
func validDocument() *source.SourceDocument {
return &source.SourceDocument{ID: "session", Units: []source.SourceUnit{{ID: 1, Text: "The party finds rope."}}}
}
func validItemRegistry() dnd.ItemRegistry {
return dnd.ItemRegistry{Items: []dnd.Item{{ID: "item", Name: "Rope", SourceRefs: []source.SourceRef{{SourceID: "session", StartUnitID: 1, EndUnitID: 1}}}}}
}

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)
}
}