ultisuite-backend/internal/search/service.go

package search

import (
	"context"
	"fmt"
	"regexp"
	"sort"
	"strings"
	"time"
	"unicode"

	"github.com/jackc/pgx/v5/pgxpool"

	"github.com/ultisuite/ulti-backend/internal/api/query"
	"github.com/ultisuite/ulti-backend/internal/nextcloud"
)

type Service struct {
	db       *pgxpool.Pool
	nc       *nextcloud.Client
	engine   string
	external externalSearchClient
}

type ServiceOptions struct {
	Nextcloud           *nextcloud.Client
	Engine              string
	MeilisearchURL      string
	MeilisearchKey      string
	MeilisearchIndex    string
	TypesenseURL        string
	TypesenseKey        string
	TypesenseCollection string
}

func NewService(db *pgxpool.Pool, opts ServiceOptions) *Service {
	engine := normalizeEngine(opts.Engine)
	return &Service{
		db:       db,
		nc:       opts.Nextcloud,
		engine:   engine,
		external: newExternalSearchClient(engine, opts),
	}
}

type Result struct {
	Type      string  `json:"type"`
	ID        string  `json:"id"`
	Title     string  `json:"title"`
	Snippet   string  `json:"snippet"`
	Date      any     `json:"date,omitempty"`
	AccountID string  `json:"account_id,omitempty"`
	Score     float64 `json:"score,omitempty"`
}

type SearchResponse struct {
	Query      string               `json:"query"`
	Engine     string               `json:"engine,omitempty"`
	Results    []Result             `json:"results"`
	Count      int                  `json:"count"`
	Pagination query.PaginationMeta `json:"pagination,omitempty"`
}

type SearchFilters struct {
	AccountID string
}

func (s *Service) Search(ctx context.Context, externalID, q, typesRaw string, params query.ListParams, filters SearchFilters) (SearchResponse, error) {
	typeList := parseTypes(typesRaw)
	usedEngine := s.engine
	if s.external == nil && s.engine != "postgres" {
		usedEngine = "postgres"
	}

	var (
		allResults []Result
		err        error
	)
	if s.external != nil {
		allResults, err = s.external.Search(ctx, ExternalSearchParams{
			ExternalID: externalID,
			Query:      q,
			Types:      typeList,
			Params:     params,
			Filters:    filters,
		})
		if err != nil {
			usedEngine = "postgres"
		}
	}
	if s.external == nil || err != nil {
		allResults, err = s.searchFederated(ctx, externalID, q, typeList, params, filters)
		if err != nil {
			return SearchResponse{}, err
		}
	}

	finalizeExternalResults(allResults, q, params)
	allResults = discardFilteredResults(allResults)
	sortResultsByScore(allResults)

	total := int64(len(allResults))
	page, _ := paginateResults(allResults, params.Offset(), params.Limit())

	return SearchResponse{
		Query:      q,
		Engine:     usedEngine,
		Results:    page,
		Count:      len(page),
		Pagination: params.Meta(&total),
	}, nil
}

func (s *Service) searchFederated(ctx context.Context, externalID, q string, typeList []string, params query.ListParams, filters SearchFilters) ([]Result, error) {
	allResults := make([]Result, 0, params.Offset()+params.Limit())
	for _, t := range typeList {
		switch t {
		case "mail":
			mailResults, err := s.searchMail(ctx, externalID, q, params, filters)
			if err != nil {
				return nil, err
			}
			allResults = append(allResults, mailResults...)
		case "contacts":
			contactResults, err := s.searchContacts(ctx, externalID, q, filters)
			if err != nil {
				return nil, err
			}
			allResults = append(allResults, contactResults...)
		case "files":
			fileResults, err := s.searchFiles(ctx, externalID, q, params, filters)
			if err != nil {
				return nil, err
			}
			allResults = append(allResults, fileResults...)
		case "events":
			eventResults, err := s.searchEvents(ctx, externalID, q, params, filters)
			if err != nil {
				return nil, err
			}
			allResults = append(allResults, eventResults...)
		}
	}
	return allResults, nil
}

func sortResultsByScore(results []Result) {
	sort.SliceStable(results, func(i, j int) bool {
		if results[i].Score == results[j].Score {
			ti, okI := dateFromAny(results[i].Date)
			tj, okJ := dateFromAny(results[j].Date)
			if okI && okJ {
				return ti.After(tj)
			}
			return okI
		}
		return results[i].Score > results[j].Score
	})
}

func dateFromAny(raw any) (time.Time, bool) {
	switch t := raw.(type) {
	case time.Time:
		return t, true
	case *time.Time:
		if t == nil {
			return time.Time{}, false
		}
		return *t, true
	case string:
		return parseDateString(t)
	default:
		return time.Time{}, false
	}
}

func parseDateString(raw string) (time.Time, bool) {
	layouts := []string{
		time.RFC3339,
		"20060102T150405Z",
		"20060102T150405",
		"20060102",
	}
	for _, layout := range layouts {
		if ts, err := time.Parse(layout, raw); err == nil {
			return ts, true
		}
	}
	return time.Time{}, false
}

func normalizeEngine(raw string) string {
	switch strings.ToLower(strings.TrimSpace(raw)) {
	case "meilisearch":
		return "meilisearch"
	case "typesense":
		return "typesense"
	default:
		return "postgres"
	}
}

func defaultCalendarRange(params query.ListParams) (time.Time, time.Time) {
	from := time.Now().UTC().AddDate(0, -6, 0)
	to := time.Now().UTC().AddDate(1, 0, 0)
	if params.From != nil {
		from = params.From.UTC()
	}
	if params.To != nil {
		to = params.To.UTC()
	}
	if from.After(to) {
		from, to = to, from
	}
	return from, to
}

func parseTypes(typesRaw string) []string {
	if strings.TrimSpace(typesRaw) == "" {
		return []string{"mail", "contacts", "files", "events"}
	}
	parts := strings.Split(typesRaw, ",")
	out := make([]string, 0, len(parts))
	seen := make(map[string]struct{}, len(parts))
	for _, p := range parts {
		t := strings.TrimSpace(p)
		if t == "" {
			continue
		}
		if _, ok := seen[t]; ok {
			continue
		}
		seen[t] = struct{}{}
		out = append(out, t)
	}
	return out
}

func paginateResults(results []Result, offset, limit int) ([]Result, int64) {
	total := int64(len(results))
	if offset >= len(results) {
		return []Result{}, total
	}
	end := offset + limit
	if end > len(results) {
		end = len(results)
	}
	return results[offset:end], total
}

func (s *Service) searchMail(ctx context.Context, externalID, queryText string, params query.ListParams, filters SearchFilters) ([]Result, error) {
	tsQuery := toTSQuery(queryText)
	if tsQuery == "" {
		return []Result{}, nil
	}

	limit := params.Offset() + params.Limit()
	if limit < 20 {
		limit = 20
	}

	args := []any{tsQuery, externalID}
	argIdx := 3
	base := `
		FROM messages m
		JOIN mail_accounts ma ON m.account_id = ma.id
		WHERE ma.user_id = (SELECT id FROM users WHERE external_id = $2)
		  AND m.search_vector @@ to_tsquery('simple', $1)
	`
	if filters.AccountID != "" {
		base += fmt.Sprintf(" AND m.account_id = $%d", argIdx)
		args = append(args, filters.AccountID)
		argIdx++
	}
	if params.From != nil {
		base += fmt.Sprintf(" AND m.date >= $%d", argIdx)
		args = append(args, params.From.UTC())
		argIdx++
	}
	if params.To != nil {
		base += fmt.Sprintf(" AND m.date <= $%d", argIdx)
		args = append(args, params.To.UTC())
		argIdx++
	}

	querySQL := `
		SELECT m.id, m.account_id, m.subject, m.snippet, m.date,
		       ts_rank(m.search_vector, to_tsquery('simple', $1)) as rank,
		       ts_headline(
		         'simple',
		         COALESCE(m.subject, '') || ' ' || COALESCE(m.snippet, ''),
		         to_tsquery('simple', $1),
		         'StartSel=<mark>,StopSel=</mark>,MaxWords=24,MinWords=8,MaxFragments=2'
		       ) as highlighted
	` + base + fmt.Sprintf(`
		ORDER BY rank DESC, m.date DESC
		LIMIT $%d
	`, argIdx)
	args = append(args, limit)

	rows, err := s.db.Query(ctx, querySQL, args...)
	if err != nil {
		return nil, err
	}
	defer rows.Close()

	results := make([]Result, 0)
	for rows.Next() {
		var id, accountID, subject, snippet string
		var date time.Time
		var rank float64
		var highlighted string
		if err := rows.Scan(&id, &accountID, &subject, &snippet, &date, &rank, &highlighted); err != nil {
			return nil, err
		}
		finalSnippet := strings.TrimSpace(highlighted)
		if finalSnippet == "" {
			finalSnippet = contextualSnippet(snippet, queryText, 220)
		}
		results = append(results, Result{
			Type:      "mail",
			ID:        id,
			Title:     highlightTerms(subject, queryText),
			Snippet:   finalSnippet,
			Date:      date.UTC(),
			AccountID: accountID,
			Score:     rank + recencyBoost(date),
		})
	}
	if err := rows.Err(); err != nil {
		return nil, err
	}
	return results, nil
}

func (s *Service) searchContacts(ctx context.Context, userID, queryText string, filters SearchFilters) ([]Result, error) {
	if s.nc == nil {
		return []Result{}, nil
	}

	bookIDs := make([]string, 0, 4)
	if filters.AccountID != "" {
		bookIDs = append(bookIDs, filters.AccountID)
	} else {
		books, err := s.nc.ListAddressBooks(ctx, userID)
		if err != nil {
			return nil, err
		}
		for _, b := range books {
			if b.ID == "" {
				continue
			}
			bookIDs = append(bookIDs, b.ID)
		}
	}

	seen := make(map[string]struct{})
	results := make([]Result, 0, 16)
	for _, bookID := range bookIDs {
		contacts, err := s.nc.ListContacts(ctx, userID, cardBookPath(userID, bookID))
		if err != nil {
			return nil, err
		}
		for _, c := range contacts {
			content := strings.TrimSpace(strings.Join([]string{c.FullName, c.Email, c.Phone, c.Org}, " "))
			if !matchesQuery(content, queryText) {
				continue
			}
			key := strings.TrimSpace(bookID + ":" + c.UID + ":" + c.Email)
			if _, ok := seen[key]; ok {
				continue
			}
			seen[key] = struct{}{}

			title := strings.TrimSpace(c.FullName)
			if title == "" {
				title = strings.TrimSpace(c.Email)
			}
			snippetSrc := joinNonEmpty(" • ", c.Email, c.Phone, c.Org)
			results = append(results, Result{
				Type:      "contacts",
				ID:        key,
				Title:     highlightTerms(title, queryText),
				Snippet:   contextualSnippet(snippetSrc, queryText, 180),
				AccountID: bookID,
				Score:     textMatchScore(content, queryText, 1.2),
			})
		}
	}
	return results, nil
}

func (s *Service) searchEvents(ctx context.Context, userID, queryText string, params query.ListParams, filters SearchFilters) ([]Result, error) {
	if s.nc == nil {
		return []Result{}, nil
	}

	from, to := defaultCalendarRange(params)
	calIDs := make([]string, 0, 4)
	if filters.AccountID != "" {
		calIDs = append(calIDs, filters.AccountID)
	} else {
		cals, err := s.nc.ListCalendars(ctx, userID)
		if err != nil {
			return nil, err
		}
		for _, c := range cals {
			if c.ID == "" {
				continue
			}
			calIDs = append(calIDs, c.ID)
		}
	}

	seen := make(map[string]struct{})
	results := make([]Result, 0, 16)
	for _, calID := range calIDs {
		events, err := s.nc.ListEvents(ctx, userID, calPath(userID, calID), from, to)
		if err != nil {
			return nil, err
		}
		for _, ev := range events {
			content := strings.TrimSpace(strings.Join([]string{ev.Summary, ev.Description, ev.Location}, " "))
			if !matchesQuery(content, queryText) {
				continue
			}
			eventDate, _ := parseDateString(ev.Start)
			key := strings.TrimSpace(calID + ":" + ev.UID + ":" + ev.Start)
			if _, ok := seen[key]; ok {
				continue
			}
			seen[key] = struct{}{}

			snippetSrc := joinNonEmpty(" • ", ev.Description, ev.Location)
			result := Result{
				Type:      "events",
				ID:        key,
				Title:     highlightTerms(ev.Summary, queryText),
				Snippet:   contextualSnippet(snippetSrc, queryText, 200),
				AccountID: calID,
				Score:     textMatchScore(content, queryText, 1.1),
			}
			if !eventDate.IsZero() {
				result.Date = eventDate.UTC()
				result.Score += recencyBoost(eventDate)
			}
			results = append(results, result)
		}
	}
	return results, nil
}

func (s *Service) searchFiles(ctx context.Context, userID, queryText string, params query.ListParams, filters SearchFilters) ([]Result, error) {
	if s.nc == nil {
		return []Result{}, nil
	}

	basePath := "/"
	if strings.TrimSpace(filters.AccountID) != "" {
		basePath = strings.TrimSpace(filters.AccountID)
	}

	files, err := s.nc.ListFiles(ctx, userID, basePath)
	if err != nil {
		return nil, err
	}

	results := make([]Result, 0, len(files))
	for _, f := range files {
		content := strings.TrimSpace(strings.Join([]string{f.Name, f.Path, f.MimeType}, " "))
		if !matchesQuery(content, queryText) {
			continue
		}

		var modifiedTime time.Time
		if ts, err := time.Parse(time.RFC1123, strings.TrimSpace(f.LastModified)); err == nil {
			modifiedTime = ts
		}
		if !modifiedTime.IsZero() {
			if params.From != nil && modifiedTime.Before(*params.From) {
				continue
			}
			if params.To != nil && modifiedTime.After(*params.To) {
				continue
			}
		}

		result := Result{
			Type:      "files",
			ID:        f.Path,
			Title:     highlightTerms(f.Name, queryText),
			Snippet:   contextualSnippet(joinNonEmpty(" • ", f.Path, f.MimeType), queryText, 220),
			AccountID: basePath,
			Score:     textMatchScore(content, queryText, 1.0),
		}
		if !modifiedTime.IsZero() {
			result.Date = modifiedTime.UTC()
			result.Score += recencyBoost(modifiedTime)
		}
		results = append(results, result)
	}
	return results, nil
}

func recencyBoost(ts time.Time) float64 {
	if ts.IsZero() {
		return 0
	}
	days := time.Since(ts).Hours() / 24
	if days < 0 {
		days = -days / 2
	}
	switch {
	case days <= 1:
		return 1.5
	case days <= 7:
		return 1.0
	case days <= 30:
		return 0.6
	case days <= 180:
		return 0.3
	default:
		return 0
	}
}

func textMatchScore(content, q string, weight float64) float64 {
	content = strings.ToLower(strings.TrimSpace(content))
	q = strings.ToLower(strings.TrimSpace(q))
	if content == "" || q == "" {
		return 0
	}
	score := 0.0
	if strings.Contains(content, q) {
		score += 3.0 * weight
	}
	for _, term := range splitSearchTerms(q) {
		hits := strings.Count(content, term)
		if hits == 0 {
			continue
		}
		score += float64(hits) * weight
		if strings.HasPrefix(content, term) {
			score += 0.25 * weight
		}
	}
	return score
}

func toTSQuery(input string) string {
	terms := splitSearchTerms(input)
	if len(terms) == 0 {
		return ""
	}
	out := make([]string, 0, len(terms))
	for _, term := range terms {
		out = append(out, term+":*")
	}
	return strings.Join(out, " & ")
}

func splitSearchTerms(input string) []string {
	rawWords := strings.Fields(strings.ToLower(strings.TrimSpace(input)))
	terms := make([]string, 0, len(rawWords))
	seen := make(map[string]struct{}, len(rawWords))
	for _, raw := range rawWords {
		clean := strings.Map(func(r rune) rune {
			if unicode.IsLetter(r) || unicode.IsDigit(r) {
				return r
			}
			return ' '
		}, raw)
		for _, part := range strings.Fields(clean) {
			if len(part) == 0 {
				continue
			}
			if _, ok := seen[part]; ok {
				continue
			}
			seen[part] = struct{}{}
			terms = append(terms, part)
		}
	}
	return terms
}

func highlightTerms(text, q string) string {
	text = strings.TrimSpace(text)
	if text == "" {
		return ""
	}
	terms := splitSearchTerms(q)
	if len(terms) == 0 {
		return text
	}
	patterns := make([]string, 0, len(terms))
	for _, term := range terms {
		patterns = append(patterns, regexp.QuoteMeta(term))
	}
	re := regexp.MustCompile(`(?i)\b(` + strings.Join(patterns, "|") + `)\b`)
	return re.ReplaceAllStringFunc(text, func(match string) string {
		return "<mark>" + match + "</mark>"
	})
}

func contextualSnippet(text, q string, maxLen int) string {
	text = strings.Join(strings.Fields(strings.TrimSpace(text)), " ")
	if text == "" {
		return ""
	}
	if maxLen <= 0 {
		maxLen = 180
	}

	snippet := text
	if len(text) > maxLen {
		snippet = text[:maxLen]
	}

	terms := splitSearchTerms(q)
	lower := strings.ToLower(text)
	bestIdx := -1
	for _, term := range terms {
		idx := strings.Index(lower, term)
		if idx >= 0 {
			bestIdx = idx
			break
		}
	}

	if bestIdx >= 0 {
		start := bestIdx - (maxLen / 3)
		if start < 0 {
			start = 0
		}
		end := start + maxLen
		if end > len(text) {
			end = len(text)
		}
		snippet = text[start:end]
		if start > 0 {
			snippet = "..." + snippet
		}
		if end < len(text) {
			snippet += "..."
		}
	}

	return highlightTerms(snippet, q)
}

func matchesQuery(content, q string) bool {
	terms := splitSearchTerms(q)
	if len(terms) == 0 {
		return true
	}
	content = strings.ToLower(content)
	for _, term := range terms {
		if strings.Contains(content, term) {
			return true
		}
	}
	return false
}

func joinNonEmpty(sep string, parts ...string) string {
	filtered := make([]string, 0, len(parts))
	for _, part := range parts {
		part = strings.TrimSpace(part)
		if part != "" {
			filtered = append(filtered, part)
		}
	}
	return strings.Join(filtered, sep)
}

func discardFilteredResults(results []Result) []Result {
	out := make([]Result, 0, len(results))
	for _, r := range results {
		if r.Score < 0 {
			continue
		}
		out = append(out, r)
	}
	return out
}

func cardBookPath(userID, bookID string) string {
	return "/remote.php/dav/addressbooks/users/" + userID + "/" + bookID + "/"
}

func calPath(userID, calID string) string {
	return "/remote.php/dav/calendars/" + userID + "/" + calID + "/"
}