stoat-mbot/src/sources/ytdlp.ts

import { spawn } from "node:child_process";
import { connect } from "node:net";
import { randomUUID } from "node:crypto";
import { constants } from "node:fs";
import { access } from "node:fs/promises";
import type { Readable } from "node:stream";
import { config } from "../config.js";
import { logger } from "../logger.js";
import { UserFacingError, type Requester, type SearchResult, type SourceKind, type Track } from "../types.js";

const log = logger.child({ mod: "yt-dlp" });

/** Raw shape of the fields we consume from yt-dlp's JSON output. */
interface YtDlpEntry {
  id?: string;
  title?: string;
  duration?: number | null;
  uploader?: string | null;
  channel?: string | null;
  artist?: string | null;
  webpage_url?: string | null;
  url?: string | null;
  original_url?: string | null;
  thumbnail?: string | null;
  thumbnails?: Array<{ url?: string }> | null;
  is_live?: boolean | null;
  live_status?: string | null;
  extractor_key?: string | null;
  ie_key?: string | null;
  _type?: string;
  entries?: YtDlpEntry[] | null;
}

function baseArgs(): string[] {
  const args = ["--no-warnings", "--no-playlist-reverse", "--ignore-config", "--no-color"];
  if (config.YTDLP_COOKIES) args.push("--cookies", config.YTDLP_COOKIES);
  if (config.YTDLP_PROXY) args.push("--proxy", config.YTDLP_PROXY);
  for (const value of config.YTDLP_EXTRACTOR_ARGS?.split(";") ?? []) {
    const trimmed = value.trim();
    if (trimmed) args.push("--extractor-args", trimmed);
  }
  return args;
}

/** Proxy URLs often carry credentials, which have no business in logs. */
export function describeProxy(): string | null {
  if (!config.YTDLP_PROXY) return null;
  try {
    const url = new URL(config.YTDLP_PROXY);
    return `${url.protocol}//${url.username ? "***@" : ""}${url.host}`;
  } catch {
    return "(некорректный URL)";
  }
}

/**
 * A quick TCP probe of the proxy. The usual mistake is pointing a container at
 * 127.0.0.1, which is the container itself — better to say so at startup than to
 * let every search fail with a connection error.
 */
export function checkProxy(timeoutMs = 4000): Promise<string | null> {
  if (!config.YTDLP_PROXY) return Promise.resolve(null);
  let target: URL;
  try {
    target = new URL(config.YTDLP_PROXY);
  } catch {
    return Promise.resolve("YTDLP_PROXY не является корректным URL");
  }
  const port = Number(target.port) || (target.protocol === "https:" ? 443 : 1080);

  return new Promise((resolve) => {
    const socket = connect({ host: target.hostname, port });
    const done = (result: string | null) => {
      socket.destroy();
      resolve(result);
    };
    socket.setTimeout(timeoutMs);
    socket.once("connect", () => done(null));
    socket.once("timeout", () => done(`прокси ${target.hostname}:${port} не отвечает`));
    socket.once("error", (err: NodeJS.ErrnoException) => {
      const hint =
        target.hostname === "127.0.0.1" || target.hostname === "localhost"
          ? " — внутри контейнера это сам контейнер; используйте host.docker.internal"
          : "";
      done(`прокси ${target.hostname}:${port} недоступен (${err.code ?? err.message})${hint}`);
    });
  });
}

export type CookieStatus = "ok" | "read-only" | "missing";

/**
 * yt-dlp rewrites the cookie file after every run to persist rotated cookies,
 * so a read-only file quietly degrades back to anonymous access.
 */
export async function checkCookies(): Promise<CookieStatus | null> {
  if (!config.YTDLP_COOKIES) return null;
  try {
    await access(config.YTDLP_COOKIES, constants.R_OK | constants.W_OK);
    return "ok";
  } catch {
    try {
      await access(config.YTDLP_COOKIES, constants.R_OK);
      return "read-only";
    } catch {
      return "missing";
    }
  }
}

interface YtDlpRun {
  stdout: string;
  stderr: string;
  code: number | null;
}

function runYtDlp(args: string[], timeoutMs = 45_000): Promise<YtDlpRun> {
  return new Promise<YtDlpRun>((resolve, reject) => {
    const child = spawn(config.YTDLP_PATH, args, { windowsHide: true });
    let stdout = "";
    let stderr = "";
    const timer = setTimeout(() => {
      child.kill("SIGKILL");
      reject(new UserFacingError("yt-dlp не ответил вовремя"));
    }, timeoutMs);

    child.stdout.setEncoding("utf8");
    child.stdout.on("data", (chunk: string) => (stdout += chunk));
    child.stderr.setEncoding("utf8");
    child.stderr.on("data", (chunk: string) => (stderr += chunk));

    child.on("error", (err: NodeJS.ErrnoException) => {
      clearTimeout(timer);
      if (err.code === "ENOENT") {
        reject(new UserFacingError(`yt-dlp не найден (${config.YTDLP_PATH}). Проверьте YTDLP_PATH.`));
        return;
      }
      reject(err);
    });

    child.on("close", (code) => {
      clearTimeout(timer);
      if (code === 0 || stdout.trim().length > 0) {
        resolve({ stdout, stderr, code });
        return;
      }
      log.warn({ code, stderr: stderr.slice(0, 800) }, "yt-dlp failed");
      reject(new UserFacingError(firstUsefulError(stderr)));
    });
  });
}

/** yt-dlp messages people actually hit, in words that suggest what to do. */
const ERROR_PATTERNS: Array<[RegExp, string]> = [
  [/unsupported url/i, "Не понимаю эту ссылку — проверьте, что она открывается в браузере"],
  [/is not a valid url/i, "Это не похоже на ссылку"],
  [/video unavailable|this video is not available/i, "Видео недоступно"],
  [/private video/i, "Видео приватное"],
  [/members[- ]only|join this channel/i, "Видео только для участников канала"],
  [/confirm your age|age[- ]restricted/i, "Возрастное ограничение — нужны cookies аккаунта (YTDLP_COOKIES)"],
  [/confirm you'?re not a bot|sign in to confirm/i, "YouTube требует вход — добавьте cookies.txt (YTDLP_COOKIES)"],
  [/not available in your country|geo[- ]?restricted|blocked it in your country/i, "Недоступно в регионе сервера"],
  [/page needs to be reloaded/i, "YouTube сменил API — обновите yt-dlp (см. README)"],
  [/unable to download|network|timed out|connection/i, "Не удалось связаться с источником"],
];

/** Percent-encoded URLs are unreadable in a chat message. */
function decodeUrls(text: string): string {
  return text.replace(/https?:\/\/\S+/g, (url) => {
    try {
      return decodeURI(url);
    } catch {
      return url;
    }
  });
}

function firstUsefulError(stderr: string): string {
  const line = stderr
    .split(/\r?\n/)
    .map((l) => l.trim())
    .find((l) => l.toUpperCase().startsWith("ERROR"));
  if (!line) return "Не удалось получить трек";

  const raw = decodeUrls(line.replace(/^ERROR:\s*/i, "")).slice(0, 300);
  const known = ERROR_PATTERNS.find(([pattern]) => pattern.test(raw));
  return known ? known[1] : raw;
}

function sourceOf(entry: YtDlpEntry): SourceKind {
  const key = (entry.extractor_key ?? entry.ie_key ?? "").toLowerCase();
  if (key.includes("soundcloud")) return "soundcloud";
  if (key.includes("youtube")) return "youtube";
  const url = entry.webpage_url ?? entry.original_url ?? entry.url ?? "";
  if (url.includes("soundcloud.com")) return "soundcloud";
  if (url.includes("youtube.com") || url.includes("youtu.be")) return "youtube";
  return "direct";
}

function pickThumbnail(entry: YtDlpEntry): string | null {
  if (entry.thumbnail) return entry.thumbnail;
  const list = entry.thumbnails ?? [];
  const last = list.at(-1);
  return last?.url ?? null;
}

export function toTrack(entry: YtDlpEntry, requestedBy: Requester, fallbackSource?: SourceKind): Track {
  const isLive = Boolean(entry.is_live) || entry.live_status === "is_live";
  const url = entry.webpage_url ?? entry.original_url ?? entry.url ?? "";
  return {
    id: randomUUID(),
    title: entry.title?.trim() || "Без названия",
    author: entry.artist ?? entry.uploader ?? entry.channel ?? null,
    duration: isLive ? 0 : Math.max(0, Math.round(entry.duration ?? 0)),
    isLive,
    url,
    thumbnail: pickThumbnail(entry),
    source: fallbackSource ?? sourceOf(entry),
    requestedBy,
  };
}

function parseNdjson(stdout: string): YtDlpEntry[] {
  const entries: YtDlpEntry[] = [];
  for (const line of stdout.split(/\r?\n/)) {
    const trimmed = line.trim();
    if (!trimmed.startsWith("{")) continue;
    try {
      entries.push(JSON.parse(trimmed) as YtDlpEntry);
    } catch {
      // yt-dlp occasionally interleaves non-JSON noise; skip it.
    }
  }
  return entries;
}

export async function search(
  query: string,
  kind: "youtube" | "soundcloud",
  limit: number,
  requestedBy: Requester,
): Promise<Track[]> {
  const prefix = kind === "soundcloud" ? "scsearch" : "ytsearch";
  const { stdout, stderr } = await runYtDlp([
    ...baseArgs(),
    "--flat-playlist",
    "--dump-json",
    `${prefix}${limit}:${query}`,
  ]);
  const entries = parseNdjson(stdout);
  // yt-dlp can exit 0 with nothing to show (bot checks, region blocks). Without
  // this the panel would just render an empty list and say nothing at all.
  if (entries.length === 0 && stderr.trim()) {
    log.warn({ query, kind, stderr: stderr.slice(0, 500) }, "search returned nothing");
    throw new UserFacingError(firstUsefulError(stderr));
  }
  log.info({ query, kind, count: entries.length }, "search");
  return entries.map((entry) => toTrack(entry, requestedBy, kind));
}

/** Resolves a URL that may point at a single track, a playlist, or an album. */
export async function resolveUrl(
  url: string,
  requestedBy: Requester,
  maxTracks: number,
  options: { singleTrack?: boolean } = {},
): Promise<SearchResult> {
  const { stdout, stderr } = await runYtDlp([
    ...baseArgs(),
    "--flat-playlist",
    "--dump-single-json",
    ...(options.singleTrack ? ["--no-playlist"] : ["--playlist-end", String(maxTracks)]),
    url,
  ]);
  const root = parseNdjson(stdout)[0];
  if (!root) {
    // Typically a dead or malformed link; yt-dlp's own line says it best.
    log.warn({ url, stderr: stderr.slice(0, 500) }, "url resolved to nothing");
    throw new UserFacingError(
      stderr.trim()
        ? firstUsefulError(stderr)
        : "По этой ссылке ничего не открылось — проверьте, что она рабочая",
    );
  }

  if (root._type === "playlist" && Array.isArray(root.entries)) {
    const entries = root.entries.filter((e): e is YtDlpEntry => Boolean(e));
    if (entries.length === 0) throw new UserFacingError("Плейлист пуст или недоступен");
    return {
      tracks: entries.map((entry) => toTrack(entry, requestedBy)),
      playlist: {
        title: root.title?.trim() || "Плейлист",
        url: root.webpage_url ?? url,
        trackCount: entries.length,
      },
    };
  }

  return { tracks: [toTrack(root, requestedBy)], playlist: null };
}

/** Resolves a direct, time-limited media URL for a page URL (used when seeking). */
export async function resolveStreamUrl(pageUrl: string): Promise<string> {
  const { stdout } = await runYtDlp([...baseArgs(), "-f", "bestaudio/best", "--no-playlist", "-g", pageUrl]);
  const url = stdout.split(/\r?\n/).map((l) => l.trim()).find(Boolean);
  if (!url) throw new UserFacingError("Не удалось получить ссылку на аудиопоток");
  return url;
}

export interface AudioProcess {
  stream: Readable;
  kill(): void;
  /** Resolves with a reason when the download fails, or null when it was fine. */
  failure: Promise<string | null>;
}

/** Spawns yt-dlp writing the best audio to stdout, for piping straight into ffmpeg. */
export function openAudioStream(pageUrl: string): AudioProcess {
  const child = spawn(
    config.YTDLP_PATH,
    [...baseArgs(), "-f", "bestaudio/best", "--no-playlist", "--quiet", "-o", "-", pageUrl],
    { stdio: ["ignore", "pipe", "pipe"], windowsHide: true },
  );

  let stderr = "";
  let killed = false;
  child.stderr.setEncoding("utf8");
  child.stderr.on("data", (chunk: string) => {
    stderr = (stderr + chunk).slice(-2000);
  });

  const failure = new Promise<string | null>((resolve) => {
    child.on("close", (code) => {
      if (killed || code === 0 || code === null) {
        resolve(null);
        return;
      }
      log.warn({ code, stderr: stderr.slice(0, 500) }, "yt-dlp stream exited with error");
      resolve(firstUsefulError(stderr));
    });
    child.on("error", (err: Error) => resolve(err.message));
  });

  return {
    stream: child.stdout,
    failure,
    kill: () => {
      killed = true;
      if (child.exitCode === null) child.kill("SIGKILL");
    },
  };
}

export async function checkAvailable(): Promise<string | null> {
  try {
    const { stdout } = await runYtDlp(["--version"], 15_000);
    return stdout.trim() || null;
  } catch {
    return null;
  }
}