notarius/internal/modules/dnd/normalize/itemoccurrences/normalizer.go

// Package itemoccurrences normalizes merged D&D item-occurrence candidates.
package itemoccurrences

import (
	"context"
	"fmt"
	"sort"

	"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"
	itemoccurrencemodel "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/itemoccurrences"
	itemregistry "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/items/registry"
	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared/diagnostics"
)

const (
	Key                 = "dnd/item-occurrences"
	normalizationPolicy = "dnd.item_occurrences.normalize.v1"
	NormalizationPolicy = normalizationPolicy

	ReasonCodeNameCanonicalized    = "item_occurrence_name_canonicalized"
	ReasonCodeUnknownItemID        = "item_occurrence_unknown_item_id"
	ReasonCodeSourceRefsNormalized = "source_references_normalized"
	ReasonCodeOccurrencesReordered = "item_occurrences_reordered"
	ReasonCodeDuplicateCollapsed   = "duplicate_item_occurrence_collapsed"
)

const (
	ItemRegistryReferenceSlot = itemregistry.ReferenceSlot
	ItemRegistryMaxBytes      = itemregistry.MaxBytes
)

var requiredCapabilities = []string{"merged"}
var providedCapabilities = []string{"normalized"}

var _ contracts.Normalizer[dnd.ItemOccurrenceList] = (*Normalizer)(nil)
var _ contracts.ManifestMetadataProvider = (*Normalizer)(nil)
var _ pipeline.CheckpointFingerprintProvider = (*Normalizer)(nil)

type Options struct{}

type Normalizer struct{ itemResolver *itemregistry.Resolver }

func New(_ Options, references ...contracts.ReferenceSet) (*Normalizer, error) {
	if len(references) > 1 {
		return nil, normalizerErrorf("at most one reference set may be supplied")
	}
	var referenceSet contracts.ReferenceSet
	if len(references) == 1 {
		referenceSet = references[0]
	}
	resolver, err := itemregistry.NewResolver(referenceSet)
	if err != nil {
		return nil, normalizerErrorf("prepare item registry: %w", err)
	}
	return &Normalizer{itemResolver: resolver}, nil
}

func (n *Normalizer) Key() string { return Key }

func (n *Normalizer) ReferenceSlots() []contracts.ReferenceSlot { return referenceSlots() }

func (n *Normalizer) ManifestMetadata() map[string]any {
	if n == nil || n.itemResolver == nil {
		return nil
	}
	metadata := map[string]any{"normalization_policy": normalizationPolicy, "comparison_policy": shared.TextComparisonPolicy}
	seeded := n.itemResolver.Seeded()
	if seeded.Bound() {
		metadata["item_registry_digest"] = seeded.Digest()
		metadata["item_count"] = seeded.Count()
	}
	return metadata
}

func (n *Normalizer) CheckpointFingerprints() []pipeline.CheckpointFingerprint {
	if n == nil || n.itemResolver == nil {
		return nil
	}
	return []pipeline.CheckpointFingerprint{
		{Name: "normalization_policy", Value: normalizationPolicy},
		{Name: "comparison_policy", Value: shared.TextComparisonPolicy},
		{Name: "item_registry", Value: n.itemResolver.Seeded().ProjectionDigest()},
	}
}

func (n *Normalizer) Normalize(ctx context.Context, req contracts.TypedNormalizeRequest[dnd.ItemOccurrenceList]) (contracts.TypedNormalizeResult[dnd.ItemOccurrenceList], error) {
	if n == nil || n.itemResolver == nil {
		return contracts.TypedNormalizeResult[dnd.ItemOccurrenceList]{}, normalizerErrorf("normalizer must not be nil")
	}
	if ctx == nil {
		return contracts.TypedNormalizeResult[dnd.ItemOccurrenceList]{}, normalizerErrorf("context must not be nil")
	}
	if err := ctx.Err(); err != nil {
		return contracts.TypedNormalizeResult[dnd.ItemOccurrenceList]{}, normalizerErrorf("context error before normalize: %w", err)
	}

	registry, err := n.itemResolver.Resolve(req.References)
	if err != nil {
		return contracts.TypedNormalizeResult[dnd.ItemOccurrenceList]{}, normalizerErrorf("resolve item registry: %w", err)
	}
	if !registry.Bound() {
		return contracts.TypedNormalizeResult[dnd.ItemOccurrenceList]{}, normalizerErrorf("item registry reference is required")
	}
	index := source.NewDocumentIndex(req.Source)
	value, findings := normalizeList(req.MergeOutput.Value, index, shared.NewSourceRefOrderFromIndex(index), registry)
	diagnosticGroups, err := diagnostics.NormalizationDiagnostics(findings, ReasonCodeUnknownItemID)
	if err != nil {
		return contracts.TypedNormalizeResult[dnd.ItemOccurrenceList]{}, normalizerErrorf("collect diagnostics: %w", err)
	}
	return contracts.TypedNormalizeResult[dnd.ItemOccurrenceList]{Value: value, Diagnostics: diagnosticGroups}, nil
}

type normalizedRecord struct {
	occurrence dnd.ItemOccurrence
	identity   string
	inputIndex int
}

func normalizeList(input dnd.ItemOccurrenceList, index source.DocumentIndex, order shared.SourceRefOrder, registry *itemregistry.Registry) (dnd.ItemOccurrenceList, []diagnostics.Finding) {
	if input.Occurrences == nil {
		return dnd.ItemOccurrenceList{}, nil
	}

	records := make([]normalizedRecord, len(input.Occurrences))
	findings := make([]diagnostics.Finding, 0)
	for index, inputOccurrence := range input.Occurrences {
		occurrence, changedFields, found, refsChanged := normalizeOccurrence(inputOccurrence, order, registry)
		records[index] = normalizedRecord{occurrence: occurrence, identity: itemoccurrencemodel.CanonicalExactIdentity(occurrence), inputIndex: index}
		if len(changedFields) != 0 {
			findings = append(findings, diagnostics.Finding{
				Scope:      occurrenceScope(index),
				ReasonCode: ReasonCodeNameCanonicalized,
				Message: fmt.Sprintf("input index %d: normalized display whitespace in %s", index,
					diagnostics.Aggregate("fields", changedFields)),
			})
		}
		if found && inputOccurrence.Name != occurrence.Name {
			findings = append(findings, diagnostics.Finding{Scope: occurrenceScope(index), ReasonCode: ReasonCodeNameCanonicalized,
				Message: fmt.Sprintf("input index %d: item name canonicalized from %s to %s", index, diagnostics.Quote(inputOccurrence.Name), diagnostics.Quote(occurrence.Name))})
		}
		if !found {
			findings = append(findings, diagnostics.Finding{Scope: occurrenceScope(index), ReasonCode: ReasonCodeUnknownItemID,
				Message: fmt.Sprintf("input index %d: item ID %s is not in the supplied registry", index, diagnostics.Quote(inputOccurrence.ItemID))})
		}
		if refsChanged {
			findings = append(findings, diagnostics.Finding{
				Scope:      occurrenceScope(index),
				ReasonCode: ReasonCodeSourceRefsNormalized,
				Message: fmt.Sprintf("input index %d: source references normalized (original count %d, final count %d)",
					index, len(inputOccurrence.SourceRefs), len(occurrence.SourceRefs)),
			})
		}
	}

	sort.SliceStable(records, func(left, right int) bool {
		return itemoccurrencemodel.CanonicalLess(order, records[left].occurrence, records[right].occurrence)
	})
	for position, record := range records {
		if position == record.inputIndex {
			continue
		}
		findings = append(findings, diagnostics.Finding{
			Scope:      occurrenceScope(record.inputIndex),
			ReasonCode: ReasonCodeOccurrencesReordered,
			Message:    fmt.Sprintf("input index %d moved to normalized position %d", record.inputIndex, position),
		})
	}

	output, duplicateFindings := collapseDuplicates(records, index)
	findings = append(findings, duplicateFindings...)
	return dnd.ItemOccurrenceList{Occurrences: output}, findings
}

func normalizeOccurrence(input dnd.ItemOccurrence, order shared.SourceRefOrder, registry *itemregistry.Registry) (dnd.ItemOccurrence, []string, bool, bool) {
	output := cloneOccurrence(input)
	canonical, found := registry.LookupID(input.ItemID)
	if found {
		output.Name = canonical.Name
	}
	changedFields := make([]string, 0, 2)
	for _, field := range []struct {
		name  string
		value *string
	}{
		{name: "from", value: &output.From},
		{name: "to", value: &output.To},
	} {
		trimmed := itemoccurrencemodel.DisplayValue(*field.value)
		if *field.value != trimmed {
			*field.value = trimmed
			changedFields = append(changedFields, field.name)
		}
	}
	output.SourceRefs = order.Canonicalize(input.SourceRefs)
	return output, changedFields, found, !itemoccurrencemodel.SourceRefsEqual(input.SourceRefs, output.SourceRefs)
}

func cloneOccurrence(input dnd.ItemOccurrence) dnd.ItemOccurrence {
	output := input
	if input.Quantity != nil {
		quantity := *input.Quantity
		output.Quantity = &quantity
	}
	if input.SourceRefs != nil {
		output.SourceRefs = append([]source.SourceRef(nil), input.SourceRefs...)
	}
	return output
}

type duplicateGroup struct {
	retainedIndex  int
	retainedRecord int
	removed        []int
}

func collapseDuplicates(records []normalizedRecord, index source.DocumentIndex) ([]dnd.ItemOccurrence, []diagnostics.Finding) {
	if len(records) == 0 {
		return make([]dnd.ItemOccurrence, 0), nil
	}

	keep := make([]bool, len(records))
	groups := make([]duplicateGroup, 0)
	groupByKey := make(map[string][]int)
	for recordIndex, record := range records {
		if !itemoccurrencemodel.ValidSourceRefs(index, record.occurrence.SourceRefs) {
			keep[recordIndex] = true
			continue
		}
		groupIndex := -1
		for _, candidate := range groupByKey[record.identity] {
			if itemoccurrencemodel.CanonicalExactEqual(records[groups[candidate].retainedRecord].occurrence, record.occurrence) {
				groupIndex = candidate
				break
			}
		}
		if groupIndex < 0 {
			groupIndex = len(groups)
			groupByKey[record.identity] = append(groupByKey[record.identity], groupIndex)
			groups = append(groups, duplicateGroup{retainedIndex: record.inputIndex, retainedRecord: recordIndex})
			keep[recordIndex] = true
			continue
		}
		groups[groupIndex].removed = append(groups[groupIndex].removed, record.inputIndex)
	}

	output := make([]dnd.ItemOccurrence, 0, len(records))
	for recordIndex, record := range records {
		if keep[recordIndex] {
			output = append(output, cloneOccurrence(record.occurrence))
		}
	}
	findings := make([]diagnostics.Finding, 0)
	for _, group := range groups {
		if len(group.removed) != 0 {
			findings = append(findings, duplicateFinding(group.retainedIndex, group.removed))
		}
	}
	return output, findings
}

func duplicateFinding(retainedIndex int, removed []int) diagnostics.Finding {
	issues := make([]string, len(removed))
	for index, removedIndex := range removed {
		issues[index] = fmt.Sprintf("removed input index %d", removedIndex)
	}
	return diagnostics.Finding{
		Scope:      occurrenceScope(retainedIndex),
		ReasonCode: ReasonCodeDuplicateCollapsed,
		Message: diagnostics.Aggregate(
			fmt.Sprintf("duplicate item occurrence collapsed; retained input index %d", retainedIndex), issues),
	}
}

func occurrenceScope(index int) string { return fmt.Sprintf("occurrences[%d]", index) }

func ModuleSpec() pipeline.ModuleSpec {
	return pipeline.ModuleSpec{
		Key:            Key,
		Stage:          pipeline.StageNormalize,
		ExecutionClass: contracts.ExecutionClassDeterministic,
		Requires:       append([]string(nil), requiredCapabilities...),
		Provides:       append([]string(nil), providedCapabilities...),
		ArtifactKind:   dnd.ItemOccurrenceListKind,
		ReferenceSlots: referenceSlots(),
	}
}

func Register(registry *pipeline.NormalizerRegistry) error {
	return pipeline.RegisterNormalizerBuilder(registry, ModuleSpec(), validateOptions, func(request pipeline.BuildRequest) (contracts.Normalizer[dnd.ItemOccurrenceList], error) {
		options, err := DecodeOptions(request.Options)
		if err != nil {
			return nil, err
		}
		return New(options, request.References)
	})
}

func referenceSlots() []contracts.ReferenceSlot {
	return []contracts.ReferenceSlot{{
		Name:                  ItemRegistryReferenceSlot,
		Description:           "Required normalized item registry used only for item identity grounding.",
		Required:              true,
		AcceptedMediaTypes:    []string{"application/json"},
		AcceptedArtifactKinds: []contracts.ArtifactKind{dnd.ItemRegistryKind},
		MaxBytes:              ItemRegistryMaxBytes,
	}}
}

func DecodeOptions(options map[string]any) (Options, error) {
	if err := pipeline.RejectUnknownOptions(options); err != nil {
		return Options{}, normalizerErrorf("%w", err)
	}
	return Options{}, nil
}

func validateOptions(options map[string]any) error { _, err := DecodeOptions(options); return err }

func normalizerErrorf(format string, args ...any) error {
	return fmt.Errorf("dnd item occurrences normalizer: "+format, args...)
}