Aurora-SpotiRip/helpers/libraryScanner.tsx

import { ALL_FORMATS, BlobSource, Input } from "mediabunny";

export type ScannedLibraryTrack = {
  id: string;
  file: File;
  relativePath: string;
  title: string;
  artist: string;
  album?: string;
  albumArtist?: string;
  trackNumber?: number;
  discNumber?: number;
  genre?: string;
  date?: string;
  lyricsPresent: boolean;
  artworkPresent: boolean;
  isrc?: string;
  codec?: string;
  sampleRate?: number;
  channels?: number;
  bitrateKbps?: number;
  missing: string[];
  duplicateOf?: string;
};

type DirectoryHandle = FileSystemDirectoryHandle & {
  values(): AsyncIterableIterator<FileSystemFileHandle | FileSystemDirectoryHandle>;
};

const AUDIO_EXTENSIONS = new Set(["mp3", "wav", "flac", "m4a", "aac", "ogg", "opus"]);

function extension(name: string) {
  const dot = name.lastIndexOf(".");
  return dot >= 0 ? name.slice(dot + 1).toLowerCase() : "";
}

function cleanGuess(name: string) {
  return name.replace(/\.[^.]+$/, "").replace(/[_]+/g, " ").replace(/\s+/g, " ").trim();
}

function rawText(raw: Record<string, unknown> | undefined, key: string) {
  const value = raw?.[key];
  if (typeof value === "string") return value;
  if (Array.isArray(value) && typeof value[0] === "string") return value[0];
  return undefined;
}

async function inspectFile(file: File, relativePath: string): Promise<ScannedLibraryTrack> {
  const input = new Input({ source: new BlobSource(file), formats: ALL_FORMATS });
  try {
    const [tags, audioTrack] = await Promise.all([
      input.getMetadataTags(),
      input.getPrimaryAudioTrack(),
    ]);
    const [codec, sampleRate, channels, averageBitrate] = audioTrack
      ? await Promise.all([
          audioTrack.getCodec().catch(() => null),
          audioTrack.getSampleRate().catch(() => undefined),
          audioTrack.getNumberOfChannels().catch(() => undefined),
          audioTrack.getAverageBitrate().catch(() => null),
        ])
      : [null, undefined, undefined, null];

    const title = tags.title?.trim() || cleanGuess(file.name);
    const artist = tags.artist?.trim() || "";
    const album = tags.album?.trim() || undefined;
    const albumArtist = tags.albumArtist?.trim() || undefined;
    const raw = tags.raw as Record<string, unknown> | undefined;
    const isrc = rawText(raw, "TSRC") || rawText(raw, "ISRC") || rawText(raw, "isrc");

    const missing: string[] = [];
    if (!tags.title?.trim()) missing.push("title");
    if (!artist) missing.push("artist");
    if (!album) missing.push("album");
    if (!tags.trackNumber) missing.push("track");
    if (!tags.genre) missing.push("genre");
    if (!tags.images?.length) missing.push("artwork");

    return {
      id: relativePath + ":" + file.size + ":" + file.lastModified,
      file,
      relativePath,
      title,
      artist,
      album,
      albumArtist,
      trackNumber: tags.trackNumber,
      discNumber: tags.discNumber,
      genre: tags.genre,
      date: tags.date?.toISOString().slice(0, 10),
      lyricsPresent: Boolean(tags.lyrics?.trim()),
      artworkPresent: Boolean(tags.images?.length),
      isrc,
      codec: codec || undefined,
      sampleRate,
      channels,
      bitrateKbps: averageBitrate ? Math.round(averageBitrate / 1000) : undefined,
      missing,
    };
  } finally {
    input.dispose();
  }
}

export async function collectAudioFiles(
  directory: FileSystemDirectoryHandle,
  prefix = "",
): Promise<Array<{ file: File; relativePath: string }>> {
  const files: Array<{ file: File; relativePath: string }> = [];
  for await (const entry of (directory as DirectoryHandle).values()) {
    const relativePath = prefix ? prefix + "/" + entry.name : entry.name;
    if (entry.kind === "directory") {
      files.push(...(await collectAudioFiles(entry as FileSystemDirectoryHandle, relativePath)));
      continue;
    }
    if (!AUDIO_EXTENSIONS.has(extension(entry.name))) continue;
    files.push({
      file: await (entry as FileSystemFileHandle).getFile(),
      relativePath,
    });
  }
  return files;
}

export async function scanAudioLibrary(
  directory: FileSystemDirectoryHandle,
  onProgress?: (completed: number, total: number) => void,
) {
  const candidates = await collectAudioFiles(directory);
  const tracks: ScannedLibraryTrack[] = [];
  for (let index = 0; index < candidates.length; index += 1) {
    const candidate = candidates[index];
    try {
      tracks.push(await inspectFile(candidate.file, candidate.relativePath));
    } catch {
      tracks.push({
        id: candidate.relativePath + ":" + candidate.file.size + ":" + candidate.file.lastModified,
        file: candidate.file,
        relativePath: candidate.relativePath,
        title: cleanGuess(candidate.file.name),
        artist: "",
        lyricsPresent: false,
        artworkPresent: false,
        missing: ["unreadable metadata"],
      });
    }
    onProgress?.(index + 1, candidates.length);
  }

  const identities = new Map<string, ScannedLibraryTrack>();
  for (const track of tracks) {
    const identity = track.isrc?.trim().toUpperCase() ||
      track.artist.toLowerCase().trim() + "::" +
      track.title.toLowerCase().trim() + "::" +
      (track.album || "").toLowerCase().trim();
    const existing = identities.get(identity);
    if (existing) track.duplicateOf = existing.relativePath;
    else identities.set(identity, track);
  }
  return tracks;
}