filebrozer/http/metadata.go

package fbhttp

import (
    "strings"
)

// canonicalMap maps various user-entered or localized keys to canonical tag names.
// Canonical names are English, PascalCase where appropriate (e.g., AlbumArtist).
var canonicalMap = map[string]string{
    // title
    "title": "Title",
    "titre": "Title",
    "song": "Title",

    // artist (main)
    "artist": "Artist",
    "artiste": "Artist",
    "artists": "Artist",

    // album
    "album": "Album",

    // album artist
    "albumartist": "AlbumArtist",
    "album artist": "AlbumArtist",
    "album_artist": "AlbumArtist",
    "artistesdelalbum": "AlbumArtist",
    "artistealbum": "AlbumArtist",

    // composer / author
    "composer": "Composer",
    "auteur": "Composer",

    // track
    "track": "Track",
    "tracknumber": "Track",
    "trackno": "Track",
    "piste": "Track",

    // disc
    "disc": "Disc",
    "discnumber": "Disc",
    "disque": "Disc",

    // genre
    "genre": "Genre",

    // year/date
    "year": "Year",
    "annee": "Year",
    "année": "Year",
    "date": "Date",

    // comment
    "comment": "Comment",
    "commentaire": "Comment",

    // lyrics
    "lyrics": "Lyrics",
}

// ffmpegKey maps canonical names to ffmpeg metadata keys.
// FFmpeg expects lower-case keys; AlbumArtist becomes album_artist, Year maps to date.
var ffmpegKey = map[string]string{
    "Title":       "title",
    "Artist":      "artist",
    "Album":       "album",
    "AlbumArtist": "album_artist",
    "Composer":    "composer",
    "Track":       "track",
    "TrackNumber": "track",
    "Disc":        "disc",
    "DiscNumber":  "disc",
    "Genre":       "genre",
    "Date":        "date",
    "Year":        "date", // prefer date for ffmpeg
    "Comment":     "comment",
    "Lyrics":      "lyrics",
}

// ffmpegMultiKey maps multi-valued canonical names to ffmpeg/Vorbis keys.
// For FLAC/Vorbis, use uppercase ARTISTS/ALBUMARTISTS for proper multi-value support.
var ffmpegMultiKey = map[string]string{
    "Artists":      "ARTISTS",
    "AlbumArtists": "ALBUMARTISTS",
}

// allowedCanonicals is the set of canonical tags we accept for writes.
var allowedCanonicals = func() map[string]struct{} {
    m := make(map[string]struct{}, len(ffmpegKey)+len(ffmpegMultiKey))
    for k := range ffmpegKey {
        m[k] = struct{}{}
    }
    for k := range ffmpegMultiKey {
        m[k] = struct{}{}
    }
    return m
}()

// normalizeKey reduces a user-provided key to a lookup token: lowercased, trimmed
// and with common separators removed.
func normalizeKey(s string) string {
    s = strings.TrimSpace(s)
    s = strings.ToLower(s)
    // normalize separators so that "Album Artist", "album-artist", "album_artist" match
    s = strings.ReplaceAll(s, "_", " ")
    s = strings.ReplaceAll(s, "-", " ")
    s = strings.Join(strings.Fields(s), " ") // collapse multiple spaces
    return s
}

// normalizeAndMapToFFmpeg takes arbitrary incoming tags, applies key normalization
// and synonym mapping, filters to allowed canonical keys, and returns a map of
// ffmpeg-ready metadata keys with their values (empty values are dropped).
func normalizeAndMapToFFmpeg(in map[string]string) map[string]string {
    out := map[string]string{}
    for k, v := range in {
        if strings.TrimSpace(v) == "" {
            continue
        }
        token := normalizeKey(k)
        canonical, ok := canonicalMap[token]
        if !ok {
            // If user already provided a canonical name, accept it
            // (e.g., Title, Artist, AlbumArtist)
            c2 := strings.TrimSpace(k)
            if _, allowed := allowedCanonicals[c2]; allowed {
                canonical = c2
            } else {
                // unrecognized key: skip
                continue
            }
        }
        ffk, ok := ffmpegKey[canonical]
        if !ok {
            continue
        }
        out[ffk] = v
    }
    return out
}
// normalizeMultiToFFmpeg maps multi-valued canonical keys to ffmpeg keys, preserving arrays.
// Uses ffmpegMultiKey for proper Vorbis comment names (ARTISTS, ALBUMARTISTS).
func normalizeMultiToFFmpeg(in map[string][]string) map[string][]string {
    out := map[string][]string{}
    for k, arr := range in {
        token := normalizeKey(k)
        canonical, ok := canonicalMap[token]
        if !ok {
            // if already canonical, keep as-is
            if _, allowed := allowedCanonicals[k]; allowed {
                canonical = k
            } else {
                continue
            }
        }
        // First check if this is a multi-valued tag
        ffk, ok := ffmpegMultiKey[canonical]
        if !ok {
            // Fall back to regular ffmpeg key
            ffk, ok = ffmpegKey[canonical]
            if !ok {
                continue
            }
        }
        vals := []string{}
        for _, v := range arr {
            v = strings.TrimSpace(v)
            if v == "" {
                continue
            }
            vals = append(vals, v)
        }
        if len(vals) > 0 {
            out[ffk] = vals
        }
    }
    return out
}

// mapClearCanonicalsToFFmpeg converts a list of canonical names to ffmpeg keys
// intended to be cleared by setting them to empty value.
func mapClearCanonicalsToFFmpeg(keys []string) map[string]struct{} {
    out := map[string]struct{}{}
    for _, k := range keys {
        token := normalizeKey(k)
        canonical, ok := canonicalMap[token]
        if !ok {
            // if already canonical and allowed, keep
            if _, allowed := allowedCanonicals[k]; allowed {
                canonical = k
            } else {
                continue
            }
        }
        if ffk, ok := ffmpegKey[canonical]; ok {
            out[ffk] = struct{}{}
        }
    }
    return out
}

// mapClearsToID3Frames converts canonical clear keys to ID3v2 frame IDs
func mapClearsToID3Frames(keys []string) []string {
    out := []string{}
    for _, k := range keys {
        token := normalizeKey(k)
        canonical, ok := canonicalMap[token]
        if !ok {
            if _, allowed := allowedCanonicals[k]; allowed {
                canonical = k
            } else {
                continue
            }
        }
        switch canonical {
        case "Title":
            out = append(out, "TIT2")
        case "Artist", "Artists":
            out = append(out, "TPE1")
        case "Album":
            out = append(out, "TALB")
        case "AlbumArtist", "AlbumArtists":
            out = append(out, "TPE2")
        case "Composer":
            out = append(out, "TCOM")
        case "Track", "TrackNumber":
            out = append(out, "TRCK")
        case "Disc", "DiscNumber":
            out = append(out, "TPOS")
        case "Genre":
            out = append(out, "TCON")
        case "Date", "Year":
            // Prefer TDRC; TYER used historically
            out = append(out, "TDRC", "TYER")
        case "Comment":
            out = append(out, "COMM")
        case "Lyrics":
            out = append(out, "USLT")
        }
    }
    return out
}

func mapClearsToTXXX(keys []string) []string {
    out := []string{}
    for _, k := range keys {
        token := normalizeKey(k)
        switch token {
        case "artists":
            out = append(out, "Artists", "ARTISTS")
        case "albumartists":
            out = append(out, "AlbumArtists", "ALBUMARTISTS")
        }
    }
    return out
}