streamrip-go/cmd/rip/lastfm.go

package main

import (
	"context"
	"encoding/json"
	"fmt"
	"html"
	"io"
	"net/http"
	"net/url"
	"strconv"
	"strings"
	"time"

	"streamrip-go/internal/app"
	"streamrip-go/internal/jsonutil"
	"streamrip-go/internal/netutil"
	"streamrip-go/internal/provider"
)

func parseLastFMArgs(args []string, defaultSource, defaultFallback string) (lastFMOptions, error) {
	opts := lastFMOptions{Source: strings.ToLower(strings.TrimSpace(defaultSource)), FallbackSource: strings.ToLower(strings.TrimSpace(defaultFallback))}
	for i := 0; i < len(args); i++ {
		switch args[i] {
		case "-s", "--source":
			if i+1 >= len(args) {
				return lastFMOptions{}, fmt.Errorf("--source requires a value")
			}
			opts.Source = strings.ToLower(strings.TrimSpace(args[i+1]))
			i++
		case "-fs", "--fallback-source":
			if i+1 >= len(args) {
				return lastFMOptions{}, fmt.Errorf("--fallback-source requires a value")
			}
			opts.FallbackSource = strings.ToLower(strings.TrimSpace(args[i+1]))
			i++
		default:
			if strings.HasPrefix(args[i], "-") {
				return lastFMOptions{}, fmt.Errorf("unknown option %q", args[i])
			}
			if opts.PlaylistURL != "" {
				return lastFMOptions{}, fmt.Errorf("unexpected extra argument %q", args[i])
			}
			opts.PlaylistURL = strings.TrimSpace(args[i])
		}
	}
	if opts.Source == "" {
		opts.Source = "qobuz"
	}
	if opts.PlaylistURL == "" {
		return lastFMOptions{}, fmt.Errorf("missing playlist url")
	}
	if !isValidLastFMPlaylistURL(opts.PlaylistURL) {
		return lastFMOptions{}, fmt.Errorf("playlist url must be a last.fm url")
	}
	if !isAllowedSearchSource(opts.Source) {
		return lastFMOptions{}, fmt.Errorf("%s %q", errLastFMInvalidSource, opts.Source)
	}
	if opts.FallbackSource != "" && !isAllowedSearchSource(opts.FallbackSource) {
		return lastFMOptions{}, fmt.Errorf("%s %q", errLastFMInvalidSource, opts.FallbackSource)
	}
	return opts, nil
}

func isValidLastFMPlaylistURL(raw string) bool {
	u, err := url.Parse(strings.TrimSpace(raw))
	if err != nil || u == nil || u.Host == "" {
		return false
	}
	s := strings.ToLower(strings.TrimSpace(u.Scheme))
	if s != "http" && s != "https" {
		return false
	}
	h := strings.ToLower(strings.TrimPrefix(strings.TrimSpace(u.Host), "www."))
	if h != "last.fm" && !strings.HasSuffix(h, ".last.fm") {
		return false
	}
	p := strings.ToLower(strings.TrimSpace(u.Path))
	return strings.Contains(p, "/playlists/")
}

func fetchLastFMPlaylist(ctx context.Context, verifySSL bool, playlistURL string) (string, []lastFMTrack, error) {
	parsed, err := url.Parse(playlistURL)
	if err != nil || parsed.Scheme == "" || parsed.Host == "" {
		return "", nil, fmt.Errorf("invalid playlist url")
	}
	if !isValidLastFMPlaylistURL(playlistURL) {
		return "", nil, fmt.Errorf("invalid playlist url")
	}
	client := netutil.NewHTTPClient(30*time.Second, verifySSL, 0)

	page1, err := fetchLastFMPlaylistPage(ctx, client, parsed, 1)
	if err != nil {
		return fetchLastFMPlaylistViaMirror(ctx, verifySSL, playlistURL)
	}
	title, total, err := extractLastFMPlaylistInfo(page1)
	if err != nil {
		return fetchLastFMPlaylistViaMirror(ctx, verifySSL, playlistURL)
	}
	tracks := extractLastFMTitleArtistPairs(page1)
	if total <= len(tracks) || total <= 50 {
		if len(tracks) > total && total > 0 {
			tracks = tracks[:total]
		}
		return title, tracks, nil
	}

	remaining := total - 50
	lastPage := 1 + remaining/50
	if remaining%50 != 0 {
		lastPage++
	}
	for page := 2; page <= lastPage; page++ {
		body, fetchErr := fetchLastFMPlaylistPage(ctx, client, parsed, page)
		if fetchErr != nil {
			return "", nil, fetchErr
		}
		tracks = append(tracks, extractLastFMTitleArtistPairs(body)...)
	}
	if len(tracks) > total {
		tracks = tracks[:total]
	}
	return title, tracks, nil
}

func fetchLastFMPlaylistViaMirror(ctx context.Context, verifySSL bool, playlistURL string) (string, []lastFMTrack, error) {
	client := netutil.NewHTTPClient(30*time.Second, verifySSL, 0)
	all := make([]lastFMTrack, 0, 200)
	title := ""

	for page := 1; page <= 50; page++ {
		body, err := fetchLastFMPlaylistMirrorPage(ctx, client, playlistURL, page)
		if err != nil {
			if page == 1 {
				return "", nil, err
			}
			break
		}
		pageTitle, tracks := extractLastFMTracksFromMirrorMarkdown(body)
		if title == "" && strings.TrimSpace(pageTitle) != "" {
			title = pageTitle
		}
		if len(tracks) == 0 {
			break
		}
		all = append(all, tracks...)
		if !strings.Contains(strings.ToLower(body), "show more") {
			break
		}
	}

	if len(all) == 0 {
		return "", nil, fmt.Errorf("could not parse playlist tracks from last.fm")
	}
	if strings.TrimSpace(title) == "" {
		title = "Last.fm Playlist"
	}
	return title, all, nil
}

func fetchLastFMPlaylistMirrorPage(ctx context.Context, client *http.Client, playlistURL string, page int) (string, error) {
	u, err := url.Parse(playlistURL)
	if err != nil {
		return "", err
	}
	if page > 1 {
		q := u.Query()
		q.Set("page", strconv.Itoa(page))
		u.RawQuery = q.Encode()
	}
	raw := u.String()
	raw = strings.TrimPrefix(raw, "https://")
	raw = strings.TrimPrefix(raw, "http://")
	mirrorURL := "https://r.jina.ai/http://" + raw

	req, err := http.NewRequestWithContext(ctx, http.MethodGet, mirrorURL, nil)
	if err != nil {
		return "", err
	}
	req.Header.Set("User-Agent", "streamrip-go/0")
	resp, err := client.Do(req)
	if err != nil {
		return "", err
	}
	defer func() { _ = resp.Body.Close() }()
	if resp.StatusCode < 200 || resp.StatusCode >= 300 {
		return "", fmt.Errorf("lastfm mirror request failed: status %d", resp.StatusCode)
	}
	b, err := io.ReadAll(resp.Body)
	if err != nil {
		return "", err
	}
	return string(b), nil
}

func fetchLastFMPlaylistPage(ctx context.Context, client *http.Client, parsed *url.URL, page int) (string, error) {
	u := *parsed
	if page > 1 {
		q := u.Query()
		q.Set("page", strconv.Itoa(page))
		u.RawQuery = q.Encode()
	}
	req, err := http.NewRequestWithContext(ctx, http.MethodGet, u.String(), nil)
	if err != nil {
		return "", err
	}
	req.Header.Set("User-Agent", "streamrip-go/0")

	resp, err := client.Do(req)
	if err != nil {
		return "", err
	}
	defer func() { _ = resp.Body.Close() }()
	if resp.StatusCode < 200 || resp.StatusCode >= 300 {
		return "", fmt.Errorf("lastfm request failed: status %d", resp.StatusCode)
	}
	b, err := io.ReadAll(resp.Body)
	if err != nil {
		return "", err
	}
	return string(b), nil
}

func extractLastFMPlaylistInfo(page string) (string, int, error) {
	titleMatch := lastFMPlaylistTitleRe.FindStringSubmatch(page)
	if len(titleMatch) < 2 {
		return "", 0, fmt.Errorf("could not parse playlist title")
	}
	totalMatch := lastFMTotalTracksRe.FindStringSubmatch(page)
	if len(totalMatch) < 2 {
		return "", 0, fmt.Errorf("could not parse total track count")
	}
	total, err := strconv.Atoi(totalMatch[1])
	if err != nil {
		return "", 0, fmt.Errorf("invalid total track count")
	}
	return html.UnescapeString(strings.TrimSpace(titleMatch[1])), total, nil
}

func extractLastFMTitleArtistPairs(page string) []lastFMTrack {
	dataPairs := lastFMDataTrackArtistRe.FindAllStringSubmatch(page, -1)
	if len(dataPairs) > 0 {
		out := make([]lastFMTrack, 0, len(dataPairs))
		for _, m := range dataPairs {
			title := html.UnescapeString(strings.TrimSpace(jsonutil.FirstNonEmpty(m[1], m[2])))
			artist := html.UnescapeString(strings.TrimSpace(jsonutil.FirstNonEmpty(m[3], m[4])))
			if title == "" || artist == "" {
				continue
			}
			out = append(out, lastFMTrack{Title: title, Artist: artist})
		}
		if len(out) > 0 {
			return out
		}
	}

	titles := lastFMTitleTagsRe.FindAllStringSubmatch(page, -1)
	out := make([]lastFMTrack, 0, len(titles)/2)
	for i := 0; i+1 < len(titles); i += 2 {
		titleRaw := strings.TrimSpace(jsonutil.FirstNonEmpty(titles[i][1], titles[i][2]))
		artistRaw := strings.TrimSpace(jsonutil.FirstNonEmpty(titles[i+1][1], titles[i+1][2]))
		if strings.EqualFold(titleRaw, "Play on YouTube") || strings.EqualFold(artistRaw, "Play on YouTube") {
			continue
		}
		title := html.UnescapeString(titleRaw)
		artist := html.UnescapeString(artistRaw)
		if title == "" || artist == "" {
			continue
		}
		out = append(out, lastFMTrack{Title: title, Artist: artist})
	}
	return out
}

func extractLastFMTracksFromMirrorMarkdown(md string) (string, []lastFMTrack) {
	lines := strings.Split(strings.ReplaceAll(md, "\r\n", "\n"), "\n")
	title := ""
	tracks := make([]lastFMTrack, 0, 100)
	for _, line := range lines {
		line = strings.TrimSpace(line)
		if title == "" {
			if m := lastFMMirrorTitleRe.FindStringSubmatch(line); len(m) >= 2 {
				title = strings.TrimSpace(html.UnescapeString(m[1]))
			}
		}
		if !strings.HasPrefix(line, "|") || !strings.Contains(strings.ToLower(line), "play track") {
			continue
		}
		cols := splitMarkdownTableRow(line)
		if len(cols) < 6 {
			continue
		}
		trackName := markdownLinkText(cols[3])
		artist := markdownLinkText(cols[4])
		if strings.TrimSpace(trackName) == "" || strings.TrimSpace(artist) == "" {
			continue
		}
		tracks = append(tracks, lastFMTrack{Title: html.UnescapeString(strings.TrimSpace(trackName)), Artist: html.UnescapeString(strings.TrimSpace(artist))})
	}
	return title, tracks
}

func splitMarkdownTableRow(line string) []string {
	trimmed := strings.TrimSpace(line)
	trimmed = strings.TrimPrefix(trimmed, "|")
	trimmed = strings.TrimSuffix(trimmed, "|")
	parts := strings.Split(trimmed, "|")
	out := make([]string, 0, len(parts))
	for _, p := range parts {
		out = append(out, strings.TrimSpace(p))
	}
	return out
}

func markdownLinkText(cell string) string {
	m := lastFMMirrorLinkTextRe.FindStringSubmatch(cell)
	if len(m) >= 2 {
		return m[1]
	}
	return strings.TrimSpace(cell)
}

func resolveLastFMTracks(ctx context.Context, mainApp *app.Main, opts lastFMOptions, tracks []lastFMTrack) ([]resolvedLastFMTrack, error) {
	primary, err := mainApp.GetLoggedInProvider(ctx, opts.Source)
	if err != nil {
		return nil, fmt.Errorf("%s login error: %w", opts.Source, err)
	}
	var fallback provider.Client
	if opts.FallbackSource != "" && opts.FallbackSource != opts.Source {
		fallback, err = mainApp.GetLoggedInProvider(ctx, opts.FallbackSource)
		if err != nil {
			return nil, fmt.Errorf("%s login error: %w", opts.FallbackSource, err)
		}
	}

	found := 0
	failed := 0
	resolved := make([]resolvedLastFMTrack, 0, len(tracks))
	for i, tr := range tracks {
		query := strings.TrimSpace(tr.Title + " " + tr.Artist)
		id, source, searchErr := searchLastFMTrack(ctx, opts, primary, fallback, query)
		if searchErr != nil {
			failed++
			fmt.Printf("[%d/%d] search failed: %s (%v)\n", i+1, len(tracks), query, searchErr)
			continue
		}
		if id == "" {
			failed++
			fmt.Printf("[%d/%d] no result: %s\n", i+1, len(tracks), query)
			continue
		}
		resolved = append(resolved, resolvedLastFMTrack{Source: source, ID: id, Query: query})
		found++
		fmt.Printf("[%d/%d] found: %s (%s)\n", i+1, len(tracks), query, source)
	}
	fmt.Printf("lastfm resolve complete: %d found, %d failed\n", found, failed)
	return resolved, nil
}

func fetchSoundcloudOEmbed(ctx context.Context, verifySSL bool, trackURL string) (map[string]any, error) {
	parsed, err := url.Parse(trackURL)
	if err != nil || parsed.Scheme == "" || parsed.Host == "" {
		return nil, fmt.Errorf("invalid soundcloud url")
	}

	q := url.Values{}
	q.Set("format", "json")
	q.Set("url", trackURL)
	endpoint := "https://soundcloud.com/oembed?" + q.Encode()

	client := netutil.NewHTTPClient(20*time.Second, verifySSL, 0)
	req, err := http.NewRequestWithContext(ctx, http.MethodGet, endpoint, nil)
	if err != nil {
		return nil, err
	}
	req.Header.Set("User-Agent", "streamrip-go/0.1")

	resp, err := client.Do(req)
	if err != nil {
		return nil, err
	}
	defer func() { _ = resp.Body.Close() }()
	if resp.StatusCode < 200 || resp.StatusCode >= 300 {
		return nil, fmt.Errorf("soundcloud oembed failed: status %d", resp.StatusCode)
	}
	body, err := io.ReadAll(resp.Body)
	if err != nil {
		return nil, err
	}
	out := map[string]any{}
	if err = json.Unmarshal(body, &out); err != nil {
		return nil, err
	}
	return out, nil
}

func searchLastFMTrack(ctx context.Context, opts lastFMOptions, primary provider.Client, fallback provider.Client, query string) (string, string, error) {
	pages, err := primary.Search(ctx, "track", query, 1)
	if err == nil {
		results := normalizeSearchResults(opts.Source, "track", pages)
		if len(results) > 0 {
			return results[0].ID, opts.Source, nil
		}
	}
	if fallback != nil {
		pages, fbErr := fallback.Search(ctx, "track", query, 1)
		if fbErr != nil {
			if err != nil {
				return "", "", fmt.Errorf("primary=%v fallback=%v", err, fbErr)
			}
			return "", "", fbErr
		}
		results := normalizeSearchResults(opts.FallbackSource, "track", pages)
		if len(results) > 0 {
			return results[0].ID, opts.FallbackSource, nil
		}
	}
	if err != nil {
		return "", "", err
	}
	return "", "", nil
}