spsays-bff/internal/repository/sitemap_repo.go

package repository

import (
	"context"
	"fmt"
	"sort"
	"strings"

	"go.mongodb.org/mongo-driver/v2/bson"
	"go.mongodb.org/mongo-driver/v2/mongo/options"
)

// SitemapRepo enumerates every public content leaf path for the sitemap.
// All queries are projection-only (no content hydration), so a single sweep
// over the ~tens-of-thousands of content docs stays cheap. URL construction
// mirrors cmd/indexer (the canonical URLs already indexed into Typesense), so
// sitemap entries match the real, crawlable pages exactly.
type SitemapRepo struct{ db *MongoDB }

func NewSitemapRepo(db *MongoDB) *SitemapRepo { return &SitemapRepo{db: db} }

// ContentURLs returns the sorted, de-duplicated set of content leaf paths
// (relative — the caller prefixes the origin).
func (r *SitemapRepo) ContentURLs(ctx context.Context) ([]string, error) {
	set := make(map[string]struct{}, 40000)
	add := func(u string) {
		if u != "" {
			set[u] = struct{}{}
		}
	}

	// Verses (bg/sb/cc/sb1962/iso/noi/bs/mm/nbs/…).
	if err := r.each(ctx, "verses",
		bson.M{"book": 1, "division": 1, "chapter": 1, "verse": 1},
		func(m bson.M) {
			add(verseURL(asString(m["book"]), asString(m["division"]),
				asInt(m["chapter"]), asInt(m["verse"])))
		}); err != nil {
		return nil, err
	}

	// Prose chapters (KB volumes, CC chapter summaries, NoD/BBD/… front matter).
	if err := r.each(ctx, "prose_chapters",
		bson.M{"book": 1, "volume": 1, "slug": 1, "type": 1},
		func(m bson.M) {
			add(proseURL(asString(m["book"]), asInt(m["volume"]),
				asString(m["slug"]), asString(m["type"])))
		}); err != nil {
		return nil, err
	}

	// Slug-keyed collections.
	for _, c := range []struct{ coll, prefix string }{
		{"letters", "/letters/"},
		{"transcriptions", "/transcriptions/"},
		{"essays", "/essays/"},
	} {
		if err := r.each(ctx, c.coll, bson.M{"slug": 1}, func(m bson.M) {
			if s := asString(m["slug"]); s != "" {
				add(c.prefix + s)
			}
		}); err != nil {
			return nil, err
		}
	}

	// Book-scoped chapter collections (book code is the URL slug, per indexer).
	for _, c := range []struct{ coll, prefix string }{
		{"journal_chapters", "/journals/"},
		{"magazine_issues", "/magazines/"},
		{"other_document_chapters", "/other-documents/"},
	} {
		if err := r.each(ctx, c.coll, bson.M{"book": 1, "slug": 1}, func(m bson.M) {
			b, s := asString(m["book"]), asString(m["slug"])
			if b != "" && s != "" {
				add(fmt.Sprintf("%s%s/%s", c.prefix, b, s))
			}
		}); err != nil {
			return nil, err
		}
	}

	// Glossary letter pages — one per distinct first letter that has entries.
	letters := make(map[string]struct{})
	if err := r.each(ctx, "glossary", bson.M{"term": 1}, func(m bson.M) {
		if l := firstAlphaLower(asString(m["term"])); l != "" {
			letters[l] = struct{}{}
		}
	}); err != nil {
		return nil, err
	}
	for l := range letters {
		add("/glossary/" + l)
	}

	out := make([]string, 0, len(set))
	for u := range set {
		out = append(out, u)
	}
	sort.Strings(out)
	return out, nil
}

// each streams a collection with the given projection, invoking fn per doc.
func (r *SitemapRepo) each(ctx context.Context, coll string, projection bson.M, fn func(bson.M)) error {
	cur, err := r.db.Collection(coll).Find(ctx, bson.M{}, options.Find().SetProjection(projection))
	if err != nil {
		return fmt.Errorf("sitemap query %s: %w", coll, err)
	}
	defer cur.Close(ctx)
	for cur.Next(ctx) {
		var m bson.M
		if err := cur.Decode(&m); err != nil {
			continue
		}
		fn(m)
	}
	return cur.Err()
}

// --- URL builders (mirror cmd/indexer/main.go) ---

func verseURL(book, division string, chapter, verse int) string {
	switch book {
	case "bg":
		return fmt.Sprintf("/books/bg/%d/%d", chapter, verse)
	case "sb":
		if division != "" {
			return fmt.Sprintf("/books/sb/%s/%d/%d", division, chapter, verse)
		}
		return fmt.Sprintf("/books/sb/%d/%d", chapter, verse)
	case "cc":
		if division != "" {
			return fmt.Sprintf("/books/cc/%s/%d/%d", division, chapter, verse)
		}
		return fmt.Sprintf("/books/cc/%d/%d", chapter, verse)
	case "sb1962":
		return fmt.Sprintf("/books/sb1962/%d/%d", chapter, verse)
	case "iso":
		return fmt.Sprintf("/books/iso/%d", chapter)
	case "noi":
		return fmt.Sprintf("/books/noi/%d", chapter)
	case "bs":
		return fmt.Sprintf("/books/bs/5/%d", verse)
	case "mm":
		return fmt.Sprintf("/books/mm/%d/%d", chapter, verse)
	case "nbs":
		return fmt.Sprintf("/books/nbs/%d/%d", chapter, verse)
	case "":
		return ""
	default:
		return fmt.Sprintf("/books/%s/%d/%d", book, chapter, verse)
	}
}

func proseURL(book string, volume int, slug, typ string) string {
	if slug == "" {
		return ""
	}
	switch {
	case book == "kb":
		return fmt.Sprintf("/books/kb/%d/%s", volume, slug)
	case typ == "chapter-summary":
		if parts := strings.SplitN(slug, "-", 2); len(parts) == 2 {
			return fmt.Sprintf("/books/%s/%s/%s", book, parts[0], parts[1])
		}
		return fmt.Sprintf("/books/%s/%s", book, slug)
	default:
		return fmt.Sprintf("/books/%s/%s", book, slug)
	}
}

// firstAlphaLower returns the lowercased first ASCII letter of a term, or ""
// when it doesn't start with a-z (matches the a–z glossary letter routes).
func firstAlphaLower(term string) string {
	term = strings.TrimSpace(term)
	if term == "" {
		return ""
	}
	c := term[0]
	if c >= 'A' && c <= 'Z' {
		c += 'a' - 'A'
	}
	if c >= 'a' && c <= 'z' {
		return string(c)
	}
	return ""
}

func asString(v any) string {
	if s, ok := v.(string); ok {
		return s
	}
	return ""
}

func asInt(v any) int {
	switch n := v.(type) {
	case int32:
		return int(n)
	case int64:
		return int(n)
	case int:
		return n
	case float64:
		return int(n)
	default:
		return 0
	}
}