av-club/sidecar/ffmpeg/run.ts

import { join } from "node:path";
import { mkdtemp, rm } from "node:fs/promises";
import { tmpdir } from "node:os";
import { JOB_TIMEOUT_MS } from "../config";
import { invalid, unprocessable, ApiError } from "../errors";
import { contentPath, putAsset, readAsset, type Asset } from "../assets/store";
import {
  AUDIO_FILTERS,
  FORMATS,
  INPUT_OPTIONS,
  OUTPUT_OPTIONS,
  PRESETS,
  VIDEO_FILTERS,
  type FfmpegParam,
} from "./vocabulary";

/**
 * Running ffmpeg.
 *
 * Everything the caller supplies is a NAME or a NUMBER from the vocabulary. Paths are constructed
 * here from an asset id, and the process is spawned with an argv array so there is no shell to quote
 * for. See docs/SECURITY.md.
 */

export type FfmpegRequest = {
  input?: unknown;
  preset?: unknown;
  inputOptions?: unknown;
  outputOptions?: unknown;
  filters?: unknown;
  format?: unknown;
};

const BITRATE = /^[0-9]+(\.[0-9]+)?[kKmM]?$/;

function value(spec: FfmpegParam, raw: unknown, path: string): string {
  switch (spec.kind) {
    case "int":
      if (typeof raw !== "number" || !Number.isInteger(raw)) throw invalid(`${path} must be an integer`, path);
      if (raw < spec.min || raw > spec.max) throw invalid(`${path} must be between ${spec.min} and ${spec.max}`, path);
      return String(raw);
    case "number":
      if (typeof raw !== "number" || !Number.isFinite(raw)) throw invalid(`${path} must be a number`, path);
      if (raw < spec.min || raw > spec.max) throw invalid(`${path} must be between ${spec.min} and ${spec.max}`, path);
      return String(raw);
    case "enum":
      if (typeof raw !== "string" || !spec.values.includes(raw)) {
        throw invalid(`${path} must be one of: ${spec.values.join(", ")}`, path);
      }
      return raw;
    case "bitrate":
      if (typeof raw !== "string" || !BITRATE.test(raw)) {
        throw invalid(`${path} must be a bitrate like "2M" or "128k"`, path);
      }
      return raw;
  }
}

function options(table: Record<string, FfmpegParam>, raw: unknown, path: string): string[] {
  if (raw === undefined || raw === null) return [];
  if (typeof raw !== "object" || Array.isArray(raw)) throw invalid(`${path} must be an object`, path);

  const args: string[] = [];
  for (const [key, entry] of Object.entries(raw as Record<string, unknown>)) {
    const spec = table[key];
    if (!spec) {
      throw invalid(`Unknown option ${JSON.stringify(key)}. Accepted here: ${Object.keys(table).join(", ")}`, `${path}.${key}`);
    }
    args.push(`-${key}`, value(spec, entry, `${path}.${key}`));
  }
  return args;
}

/**
 * A filter chain, built from structured objects rather than parsed from a string.
 *
 * ffmpeg's filter syntax has its own escaping rules, and parsing it well enough to validate it is
 * harder than accepting `{ name: args }`. The names and parameters are ffmpeg's; only the
 * punctuation is ours — and every filter that can read a file or load a plugin is simply absent from
 * the table, so it cannot be named.
 */
function filterChain(table: Record<string, Record<string, FfmpegParam>>, raw: unknown, path: string): string {
  if (raw === undefined || raw === null) return "";
  if (!Array.isArray(raw)) throw invalid(`${path} must be an array of filters`, path);

  return raw
    .map((entry, i) => {
      const at = `${path}[${i}]`;
      if (!entry || typeof entry !== "object") throw invalid(`${at} must be an object`, at);

      const pairs = Object.entries(entry as Record<string, unknown>);
      if (pairs.length !== 1) throw invalid(`${at} must name exactly one filter`, at);

      const [name, rawArgs] = pairs[0]!;
      const spec = table[name];
      if (!spec) throw invalid(`Unknown filter ${JSON.stringify(name)}. See docs/FFMPEG.md.`, `${at}.${name}`);

      const args = Object.entries((rawArgs ?? {}) as Record<string, unknown>).map(([key, v]) => {
        const argSpec = spec[key];
        if (!argSpec) {
          throw invalid(`Unknown parameter ${JSON.stringify(key)} for ${name}`, `${at}.${name}.${key}`);
        }
        return `${key}=${value(argSpec, v, `${at}.${name}.${key}`)}`;
      });

      return args.length ? `${name}=${args.join(":")}` : name;
    })
    .join(",");
}

/** Run ffmpeg on an asset and store the result as a new asset. The input is never modified. */
export async function runFfmpeg(body: FfmpegRequest): Promise<Asset> {
  if (typeof body.input !== "string") throw invalid("input must be an asset id", "input");

  const source = await readAsset(body.input);
  if (source.kind === "image") {
    throw invalid(`Asset ${source.id} is an image. Use /v1/sharp for images.`, "input");
  }

  // A preset supplies the options; anything given explicitly overrides it. So "web-1080 but crf 20"
  // does not mean writing out every flag again.
  let format = "mp4";
  let presetOptions: string[] = [];
  let presetFilters = "";

  if (body.preset !== undefined) {
    if (typeof body.preset !== "string" || !PRESETS[body.preset]) {
      throw invalid(`preset must be one of: ${Object.keys(PRESETS).join(", ")}`, "preset");
    }
    const preset = PRESETS[body.preset]!;
    format = preset.format;
    presetOptions = options(OUTPUT_OPTIONS, preset.options, "preset");
    presetFilters = preset.filters?.v
      ? filterChain(VIDEO_FILTERS, preset.filters.v.map((f) => ({ [f.name]: f.args })), "preset.filters")
      : "";
  }

  if (body.format !== undefined) {
    if (typeof body.format !== "string" || !FORMATS[body.format]) {
      throw invalid(`format must be one of: ${Object.keys(FORMATS).join(", ")}`, "format");
    }
    format = body.format;
  }

  const inputArgs = options(INPUT_OPTIONS, body.inputOptions, "inputOptions");
  const outputArgs = options(OUTPUT_OPTIONS, body.outputOptions, "outputOptions");

  const filters = (body.filters ?? {}) as Record<string, unknown>;
  const videoChain = filterChain(VIDEO_FILTERS, filters.v, "filters.v") || presetFilters;
  const audioChain = filterChain(AUDIO_FILTERS, filters.a, "filters.a");

  // One directory per job, removed however this ends — otherwise disk fills with the failures.
  const workDir = await mkdtemp(join(tmpdir(), "avclub-"));
  const output = join(workDir, `out.${FORMATS[format]!.extension}`);

  try {
    const argv = [
      "ffmpeg",
      "-nostdin",
      "-y",
      "-v", "error",
      // Nothing but local files. Kills SSRF and the concat:/subfile: tricks outright — the second
      // lock behind "the caller never names a path".
      "-protocol_whitelist", "file",
      ...inputArgs,
      "-i", contentPath(source.id),
      ...(videoChain ? ["-vf", videoChain] : []),
      ...(audioChain ? ["-af", audioChain] : []),
      ...presetOptions,
      // After the preset, so an explicit option wins.
      ...outputArgs,
      "-f", format,
      output,
    ];

    const proc = Bun.spawn(argv, { stdout: "pipe", stderr: "pipe", stdin: "ignore" });

    // A crafted file can make an encoder run effectively forever.
    let timedOut = false;
    const timer = setTimeout(() => {
      timedOut = true;
      proc.kill();
    }, JOB_TIMEOUT_MS);

    const [stderr, code] = await Promise.all([new Response(proc.stderr).text(), proc.exited]);
    clearTimeout(timer);

    if (timedOut) {
      throw new ApiError("timeout", `Encoding took longer than ${JOB_TIMEOUT_MS / 1000}s and was stopped.`);
    }
    if (code !== 0) {
      throw unprocessable(`ffmpeg failed: ${stderr.trim().split("\n").at(-1) ?? `exit ${code}`}`);
    }

    return await putAsset(await Bun.file(output).arrayBuffer());
  } finally {
    await rm(workDir, { recursive: true, force: true });
  }
}