descrybe/apps/api/internal/feeds/suggest.go

package feeds

import "strings"

// MappingSuggestion is a suggested source→target pair from schema extraction.
type MappingSuggestion struct {
	Source     string  `json:"source"`
	Target     string  `json:"target"`
	Confidence string  `json:"confidence"`
	Score      float64 `json:"score"`
}

var sourceAliases = map[string][]string{
	"gtin":             {"gtin"},
	"ean":              {"gtin"},
	"upc":              {"gtin"},
	"barcode":          {"gtin"},
	"title":            {"title"},
	"name":             {"title"},
	"productname":      {"title"},
	"brand":            {"brand"},
	"manufacturer":     {"brand"},
	"description":      {"description"},
	"desc":             {"description"},
	"price":            {"price"},
	"regularprice":     {"price"},
	"saleprice":        {"sale_price", "price"},
	"currency":         {"currency"},
	"image":            {"image_url", "main_image", "image"},
	"imageurl":         {"image_url", "main_image", "image"},
	"imagelink":        {"image_url", "main_image", "image"},
	"mainimage":        {"image_url", "main_image", "image"},
	"mainimageurl":     {"image_url", "main_image", "image"},
	"link":             {"product_url"},
	"url":              {"product_url"},
	"producturl":       {"product_url"},
	"sku":              {"sku"},
	"mpn":              {"mpn"},
	"category":         {"category"},
	"availability":     {"availability"},
	"stockstatus":      {"availability", "stock_status"},
	"stock":            {"stock"},
	"quantity":         {"stock"},
	"qty":              {"stock"},
	"color":            {"color"},
	"size":             {"size"},
	"material":         {"material"},
	"weight":           {"weight"},
	"netmass":          {"weight"},
	"purchaseprice":    {"purchase_price", "price"},
	"buyprice":         {"purchase_price"},
	"cost":             {"purchase_price"},
	"costprice":        {"purchase_price"},
	"officiallink":     {"official_link"},
	"warranty":         {"warranty"},
	"garancija":        {"warranty"},
	"service":          {"service"},
	"servis":           {"service"},
	"productmodel":     {"product_model"},
	"model":            {"product_model"},
	"modelnumber":      {"product_model"},
	"eprelid":          {"eprel_id"},
	"eprel":            {"eprel_id"},
	"specifications":   {"specs", "specifications"},
	"specs":            {"specs", "specifications"},
	"specification":    {"specs", "specifications"},
	"techspecs":        {"specs", "specifications"},
}

func normalizeSuggestKey(raw string) string {
	s := strings.ToLower(strings.TrimSpace(raw))
	var b strings.Builder
	for _, r := range s {
		if (r >= 'a' && r <= 'z') || (r >= '0' && r <= '9') {
			b.WriteRune(r)
		}
	}
	return b.String()
}

func leafSuggestKey(path string) string {
	leaf := path
	if i := strings.LastIndex(path, "/"); i >= 0 {
		leaf = path[i+1:]
	}
	if i := strings.LastIndex(leaf, ":"); i >= 0 {
		leaf = leaf[i+1:]
	}
	return normalizeSuggestKey(leaf)
}

// SuggestTarget returns the best canonical target key for a single source name/path,
// or "" when no alias matches. Used during schema extract to annotate fields.
func SuggestTarget(source string) string {
	keys := []string{normalizeSuggestKey(source), leafSuggestKey(source)}
	for _, key := range keys {
		if key == "" {
			continue
		}
		if aliases, ok := sourceAliases[key]; ok && len(aliases) > 0 {
			return aliases[0]
		}
		// Exact identity for known-looking keys already in alias map as targets.
		for _, aliases := range sourceAliases {
			for _, a := range aliases {
				if a == key {
					return a
				}
			}
		}
	}
	return ""
}

// SuggestMappings fuzzy-matches schema fields onto enabled target keys (1:1).
func SuggestMappings(schema []SchemaField, enabledTargets []string) []MappingSuggestion {
	enabled := make(map[string]struct{}, len(enabledTargets))
	for _, t := range enabledTargets {
		t = strings.TrimSpace(t)
		if t != "" {
			enabled[t] = struct{}{}
		}
	}
	if len(schema) == 0 || len(enabled) == 0 {
		return nil
	}
	used := map[string]struct{}{}
	type scored struct {
		MappingSuggestion
		order int
	}
	var candidates []scored
	for i, f := range schema {
		keys := []string{
			normalizeSuggestKey(f.FieldName),
			leafSuggestKey(f.Path),
			normalizeSuggestKey(f.Path),
		}
		var hit *MappingSuggestion
		for _, key := range keys {
			if key == "" {
				continue
			}
			if _, ok := enabled[key]; ok {
				if _, taken := used[key]; !taken {
					hit = &MappingSuggestion{Source: f.Path, Target: key, Confidence: "exact", Score: 1}
					break
				}
			}
			if aliases, ok := sourceAliases[key]; ok {
				for _, a := range aliases {
					if _, ok := enabled[a]; !ok {
						continue
					}
					if _, taken := used[a]; taken {
						continue
					}
					hit = &MappingSuggestion{Source: f.Path, Target: a, Confidence: "alias", Score: 0.95}
					break
				}
				if hit != nil {
					break
				}
			}
		}
		if hit != nil {
			candidates = append(candidates, scored{MappingSuggestion: *hit, order: i})
		}
	}
	// Prefer higher score, then earlier schema order.
	for i := 0; i < len(candidates); i++ {
		for j := i + 1; j < len(candidates); j++ {
			if candidates[j].Score > candidates[i].Score ||
				(candidates[j].Score == candidates[i].Score && candidates[j].order < candidates[i].order) {
				candidates[i], candidates[j] = candidates[j], candidates[i]
			}
		}
	}
	out := make([]MappingSuggestion, 0, len(candidates))
	for _, c := range candidates {
		if _, taken := used[c.Target]; taken {
			continue
		}
		used[c.Target] = struct{}{}
		out = append(out, c.MappingSuggestion)
	}
	return out
}