NaviWatcher/internal/musicbrainz/api.go

package musicbrainz

import (
	"context"
	"fmt"
	"net/url"
	"regexp"
	"strings"
	"unicode"

	"naviwatcher/internal/database"
)

// excludedStatuses contains release-group statuses that should be filtered out.
var excludedStatuses = map[string]bool{
	"Bootleg":        true,
	"Promotion":      true,
	"Pseudo-Release": true,
}

// includedTypes contains release-group types that should be included.
var includedTypes = map[string]bool{
	"Album":       true,
	"Single":      true,
	"EP":          true,
	"Compilation": true,
}

// Precompiled regexes for NormalizeString — compiled once at package init.
var (
	bracketRe = regexp.MustCompile(`\[[^\]]*\]`)
	parenRe   = regexp.MustCompile(`\([^)]*\)`)
	yearRe    = regexp.MustCompile(`\b(1[0-9]{3}|2[0-9]{3})\b`)
	spaceRe   = regexp.MustCompile(`\s+`)
)

// GetArtistReleaseGroups fetches all release groups for a given artist from MusicBrainz.
// It queries the artist's release groups via the MusicBrainz Web Service API,
// parses the XML response, and applies status and type filtering.
//
// The method handles pagination automatically by following offset parameters
// until all release groups are fetched.
func (c *MusicBrainzClient) GetArtistReleaseGroups(ctx context.Context, artistMBID string) ([]ReleaseGroup, error) {
	var allGroups []ReleaseGroup
	offset := 0
	limit := 100 // MusicBrainz max limit per request

	for {
		params := url.Values{}
		params.Set("artist", artistMBID)
		params.Set("limit", fmt.Sprintf("%d", limit))
		params.Set("offset", fmt.Sprintf("%d", offset))
		path := buildPath("/release-group", params)

		body, err := c.doGet(ctx, path)
		if err != nil {
			return nil, fmt.Errorf("fetch release groups for artist %s: %w", artistMBID, err)
		}

		parsed, err := ParseReleaseGroups(body)
		if err != nil {
			return nil, fmt.Errorf("parse release groups for artist %s: %w", artistMBID, err)
		}

		allGroups = append(allGroups, parsed.ReleaseGroups...)

		// If we've fetched all results, we've reached the end.
		// Also break on empty page to prevent infinite loop if API
		// returns fewer items than advertised by count.
		if len(parsed.ReleaseGroups) == 0 || offset+len(parsed.ReleaseGroups) >= parsed.Count {
			break
		}

		// Check context cancellation between pages for responsive shutdown.
		if err := ctx.Err(); err != nil {
			return nil, fmt.Errorf("fetch release groups for artist %s: %w", artistMBID, err)
		}

		offset += limit
	}

	return allGroups, nil
}

// FilterReleaseGroups applies status and type filtering to a list of release groups.
// It excludes Bootleg, Promotion, and Pseudo-Release statuses.
// It includes only Album, Single, EP, and Compilation types.
func FilterReleaseGroups(groups []ReleaseGroup) []ReleaseGroup {
	var filtered []ReleaseGroup
	for _, rg := range groups {
		if IsStatusExcluded(rg.Status) {
			continue
		}
		if !IsTypeIncluded(rg.Type) {
			continue
		}
		filtered = append(filtered, rg)
	}
	return filtered
}

// IsStatusExcluded returns true if the given status should be excluded.
func IsStatusExcluded(status string) bool {
	return excludedStatuses[status]
}

// IsTypeIncluded returns true if the given type is in the base included set.
func IsTypeIncluded(releaseType string) bool {
	return includedTypes[releaseType]
}

// NormalizeString normalizes a string for fuzzy matching by:
//   - Converting to lowercase
//   - Removing special characters (keeping only letters, digits, and spaces)
//   - Removing years (4-digit numbers that look like years)
//   - Removing bracketed keywords (e.g., [Deluxe], [Remastered])
//   - Collapsing multiple spaces into one
//   - Trimming leading/trailing whitespace
func NormalizeString(s string) string {
	// Convert to lowercase
	s = strings.ToLower(s)

	// Remove bracketed content first (e.g., [Deluxe Edition], [Remastered 2020])
	s = bracketRe.ReplaceAllString(s, "")

	// Remove parenthesized content (e.g., (Deluxe), (Remastered))
	s = parenRe.ReplaceAllString(s, "")

	// Remove years (4-digit numbers between 1000-2999)
	s = yearRe.ReplaceAllString(s, "")

	// Replace common separators with spaces before stripping other special chars
	s = strings.ReplaceAll(s, "-", " ")
	s = strings.ReplaceAll(s, "_", " ")

	// Keep only letters, digits, and spaces
	var b strings.Builder
	for _, r := range s {
		if unicode.IsLetter(r) || unicode.IsDigit(r) || unicode.IsSpace(r) {
			b.WriteRune(r)
		}
	}
	s = b.String()

	// Collapse multiple spaces
	s = spaceRe.ReplaceAllString(s, " ")

	// Trim
	s = strings.TrimSpace(s)

	return s
}

// NormalizeArtistName normalizes an artist name for comparison.
// It applies NormalizeString and additionally handles common prefixes.
func NormalizeArtistName(name string) string {
	name = NormalizeString(name)

	// Remove common leading articles for better matching
	prefixes := []string{"the ", "a ", "an "}
	for _, prefix := range prefixes {
		if strings.HasPrefix(name, prefix) {
			name = strings.TrimPrefix(name, prefix)
			break
		}
	}

	return strings.TrimSpace(name)
}

// ToExternalRelease converts a ReleaseGroup to an ExternalRelease
// for database persistence.
func (rg *ReleaseGroup) ToExternalRelease() *database.ExternalRelease {
	return &database.ExternalRelease{
		RGID:        rg.ID,
		ArtistID:    rg.ArtistID,
		Title:       rg.Title,
		Type:        rg.Type,
		ReleaseDate: rg.ReleaseDate,
	}
}