pinkudex/lib/video/subtitleCache.ts

import "server-only";
import path from "node:path";
import fs from "node:fs";
import fsp from "node:fs/promises";
import crypto from "node:crypto";

/**
 * Bump on any change to srtToVtt, the ffmpeg arg recipe, or the
 * cache-key composition. Old entries become unreachable automatically.
 *  v1 → initial.
 *  v2 → added decodeSubtitleBuffer for non-UTF-8 SRTs/VTTs (cp936,
 *       shift-jis, big5, UTF-16). Existing UTF-8-only entries would
 *       still be correct but the version bump ensures any cached
 *       output produced with a buggy decode path is regenerated.
 */
export const CONVERTER_VERSION = 2;

const CACHE_DIR = path.join(process.cwd(), "data", "subtitle-cache");
fs.mkdirSync(CACHE_DIR, { recursive: true });

export type CacheKind = "embedded" | "srt" | "ass" | "ssa";

export interface CacheKeyInput {
  abs: string;
  size: number;
  mtimeMs: number;
  kind: CacheKind;
  /** ffmpeg stream index for embedded; ext for sidecar files. */
  streamOrExt: string | number;
}

export function cachePath(input: CacheKeyInput): string {
  const raw = [
    input.abs,
    input.size,
    Math.round(input.mtimeMs),
    input.kind,
    String(input.streamOrExt),
    CONVERTER_VERSION,
  ].join("|");
  const hash = crypto.createHash("sha1").update(raw).digest("hex");
  return path.join(CACHE_DIR, `${hash}.vtt`);
}

export async function readCache(file: string): Promise<Buffer | null> {
  try {
    const buf = await fsp.readFile(file);
    // Bump mtime so LRU pruning treats this entry as recently used.
    // Best effort: failure (read-only fs, locked file) is harmless.
    const now = Date.now() / 1000;
    fsp.utimes(file, now, now).catch(() => { /* ignore */ });
    return buf;
  } catch {
    return null;
  }
}

let writesSinceLastPrune = 0;
const PRUNE_WRITE_INTERVAL = 25;

export async function writeCache(file: string, data: Buffer | string): Promise<void> {
  // Atomic via rename — avoids partial files if the process is killed
  // mid-write or two requests race on the same key.
  const tmp = `${file}.${process.pid}.${Date.now()}.tmp`;
  try {
    await fsp.writeFile(tmp, data);
    await fsp.rename(tmp, file);
  } catch (e) {
    try { await fsp.unlink(tmp); } catch { /* ignore */ }
    throw e;
  }
  writesSinceLastPrune++;
  if (writesSinceLastPrune >= PRUNE_WRITE_INTERVAL) {
    writesSinceLastPrune = 0;
    void pruneSubtitleCacheIfNeeded();
  }
}

interface PruneResult {
  scanned: number;
  removed: number;
  beforeBytes: number;
  afterBytes: number;
}

/** LRU sweep keyed on file mtime. Walks `data/subtitle-cache/`,
 *  computes total size, and if it exceeds the configured limit,
 *  deletes the oldest-mtime entries until size drops below 80% of
 *  the cap. No-op when the limit setting is 0 (unlimited). */
export async function pruneSubtitleCacheIfNeeded(): Promise<PruneResult> {
  const { getAppSetting } = await import("@/lib/db/appSettings");
  const limitMb = Number(getAppSetting("subtitleCacheLimitMb"));
  const result: PruneResult = { scanned: 0, removed: 0, beforeBytes: 0, afterBytes: 0 };
  if (!Number.isFinite(limitMb) || limitMb <= 0) return result;
  const limitBytes = limitMb * 1024 * 1024;
  const lowWatermark = Math.floor(limitBytes * 0.8);

  let entries: import("node:fs").Dirent[];
  try {
    entries = await fsp.readdir(CACHE_DIR, { withFileTypes: true });
  } catch {
    return result;
  }
  type CacheEntry = { abs: string; size: number; mtimeMs: number };
  const items: CacheEntry[] = [];
  for (const e of entries) {
    if (!e.isFile() || !e.name.endsWith(".vtt")) continue;
    const abs = path.join(CACHE_DIR, e.name);
    try {
      const stat = await fsp.stat(abs);
      items.push({ abs, size: stat.size, mtimeMs: stat.mtimeMs });
      result.scanned++;
      result.beforeBytes += stat.size;
    } catch { /* file vanished mid-walk; skip */ }
  }
  if (result.beforeBytes <= limitBytes) {
    result.afterBytes = result.beforeBytes;
    return result;
  }
  // Oldest first.
  items.sort((a, b) => a.mtimeMs - b.mtimeMs);
  let running = result.beforeBytes;
  for (const it of items) {
    if (running <= lowWatermark) break;
    try {
      await fsp.unlink(it.abs);
      running -= it.size;
      result.removed++;
    } catch { /* concurrent delete; skip */ }
  }
  result.afterBytes = running;
  if (result.removed > 0) {
    console.log(
      `[subtitle-cache] pruned ${result.removed}/${result.scanned} files; ${(result.beforeBytes / 1_048_576).toFixed(1)}MB → ${(running / 1_048_576).toFixed(1)}MB`,
    );
  }
  return result;
}

/** Bootstrap entrypoint — fire one sweep on module load (delayed so
 *  other startup work isn't blocked). */
function scheduleBootstrapPrune(): void {
  setTimeout(() => { void pruneSubtitleCacheIfNeeded(); }, 5_000);
}
scheduleBootstrapPrune();