amcs/internal/metadata/normalize.go

package metadata

import (
	"sort"
	"strings"

	"git.warky.dev/wdevs/amcs/internal/config"
	thoughttypes "git.warky.dev/wdevs/amcs/internal/types"
)

const (
	DefaultType          = "observation"
	DefaultTopicFallback = "uncategorized"
	maxTopics            = 10
)

var allowedTypes = map[string]struct{}{
	"observation": {},
	"task":        {},
	"idea":        {},
	"reference":   {},
	"person_note": {},
}

func Fallback(capture config.CaptureConfig) thoughttypes.ThoughtMetadata {
	topicFallback := strings.TrimSpace(capture.MetadataDefaults.TopicFallback)
	if topicFallback == "" {
		topicFallback = DefaultTopicFallback
	}

	return thoughttypes.ThoughtMetadata{
		People:         []string{},
		ActionItems:    []string{},
		DatesMentioned: []string{},
		Topics:         []string{topicFallback},
		Type:           normalizeType(capture.MetadataDefaults.Type),
		Source:         normalizeSource(capture.Source),
		Attachments:    []thoughttypes.ThoughtAttachment{},
	}
}

func Normalize(in thoughttypes.ThoughtMetadata, capture config.CaptureConfig) thoughttypes.ThoughtMetadata {
	out := thoughttypes.ThoughtMetadata{
		People:         normalizeList(in.People, 0),
		ActionItems:    normalizeList(in.ActionItems, 0),
		DatesMentioned: normalizeList(in.DatesMentioned, 0),
		Topics:         normalizeList(in.Topics, maxTopics),
		Type:           normalizeType(in.Type),
		Source:         normalizeSource(in.Source),
		Attachments:    normalizeAttachments(in.Attachments),
	}

	if len(out.Topics) == 0 {
		out.Topics = Fallback(capture).Topics
	}
	if out.Type == "" {
		out.Type = Fallback(capture).Type
	}
	if out.Source == "" {
		out.Source = Fallback(capture).Source
	}

	return out
}

func normalizeList(values []string, limit int) []string {
	seen := make(map[string]struct{}, len(values))
	result := make([]string, 0, len(values))

	for _, value := range values {
		trimmed := strings.Join(strings.Fields(strings.TrimSpace(value)), " ")
		if trimmed == "" {
			continue
		}

		key := strings.ToLower(trimmed)
		if _, ok := seen[key]; ok {
			continue
		}

		seen[key] = struct{}{}
		result = append(result, trimmed)

		if limit > 0 && len(result) >= limit {
			break
		}
	}

	return result
}

func normalizeType(value string) string {
	normalized := strings.ToLower(strings.TrimSpace(value))
	if normalized == "" {
		return DefaultType
	}
	if _, ok := allowedTypes[normalized]; ok {
		return normalized
	}
	return DefaultType
}

func normalizeSource(value string) string {
	normalized := strings.TrimSpace(value)
	if normalized == "" {
		return config.DefaultSource
	}
	return normalized
}

func Merge(base, patch thoughttypes.ThoughtMetadata, capture config.CaptureConfig) thoughttypes.ThoughtMetadata {
	merged := base

	if len(patch.People) > 0 {
		merged.People = append(append([]string{}, merged.People...), patch.People...)
	}
	if len(patch.ActionItems) > 0 {
		merged.ActionItems = append(append([]string{}, merged.ActionItems...), patch.ActionItems...)
	}
	if len(patch.DatesMentioned) > 0 {
		merged.DatesMentioned = append(append([]string{}, merged.DatesMentioned...), patch.DatesMentioned...)
	}
	if len(patch.Topics) > 0 {
		merged.Topics = append(append([]string{}, merged.Topics...), patch.Topics...)
	}
	if strings.TrimSpace(patch.Type) != "" {
		merged.Type = patch.Type
	}
	if strings.TrimSpace(patch.Source) != "" {
		merged.Source = patch.Source
	}
	if len(patch.Attachments) > 0 {
		merged.Attachments = append(append([]thoughttypes.ThoughtAttachment{}, merged.Attachments...), patch.Attachments...)
	}

	return Normalize(merged, capture)
}

func normalizeAttachments(values []thoughttypes.ThoughtAttachment) []thoughttypes.ThoughtAttachment {
	seen := make(map[string]struct{}, len(values))
	result := make([]thoughttypes.ThoughtAttachment, 0, len(values))

	for _, value := range values {
		if value.FileID.String() == "" || value.FileID.String() == "00000000-0000-0000-0000-000000000000" {
			continue
		}

		key := value.FileID.String()
		if _, ok := seen[key]; ok {
			continue
		}

		value.Name = strings.TrimSpace(value.Name)
		value.MediaType = strings.TrimSpace(value.MediaType)
		value.Kind = strings.TrimSpace(value.Kind)
		if value.SizeBytes < 0 {
			value.SizeBytes = 0
		}
		value.SHA256 = strings.TrimSpace(value.SHA256)

		seen[key] = struct{}{}
		result = append(result, value)
	}

	return result
}

func SortedTopCounts(in map[string]int, limit int) []thoughttypes.KeyCount {
	out := make([]thoughttypes.KeyCount, 0, len(in))
	for key, count := range in {
		if strings.TrimSpace(key) == "" {
			continue
		}
		out = append(out, thoughttypes.KeyCount{Key: key, Count: count})
	}

	sort.Slice(out, func(i, j int) bool {
		if out[i].Count == out[j].Count {
			return out[i].Key < out[j].Key
		}
		return out[i].Count > out[j].Count
	})

	if limit > 0 && len(out) > limit {
		return out[:limit]
	}

	return out
}