pinkudex/lib/video/metadata.ts

import "server-only";
import path from "node:path";
import { spawn } from "node:child_process";
import fs from "node:fs/promises";
import { revalidatePath } from "next/cache";
import { rawDb } from "@/lib/db/client";
import { getAppSetting } from "@/lib/db/appSettings";
import { classifyGroup, compilePatterns } from "./partClassify";
import type { VideoFile } from "./index";

const PROBE_TIMEOUT_MS = 10_000;

export type PlaybackMode = "direct" | "transcode";

export interface StoredVideoMetadata {
  absPath: string;
  relPath: string;
  code: string;
  sizeBytes: number;
  mtimeMs: number;
  probedAt: number | null;
  probeError: string | null;
  durationSec: number | null;
  videoCodec: string | null;
  videoBFrames: number | null;
  width: number | null;
  height: number | null;
  videoBitrate: number | null;
  playbackMode: PlaybackMode | null;
  partKind: "part" | "variant" | "single" | null;
  partIndex: number | null;
  variantGroup: string | null;
}

interface VideoMetadataRow {
  abs_path: string;
  rel_path: string;
  code: string;
  size_bytes: number;
  mtime_ms: number;
  probed_at: number | null;
  probe_error: string | null;
  duration_sec: number | null;
  video_codec: string | null;
  video_b_frames: number | null;
  width: number | null;
  height: number | null;
  video_bitrate: number | null;
  playback_mode: string | null;
  part_kind: string | null;
  part_index: number | null;
  variant_group: string | null;
}

interface FfprobeJson {
  streams?: Array<{
    codec_name?: string;
    width?: number;
    height?: number;
    bit_rate?: string;
    has_b_frames?: number;
  }>;
  format?: {
    duration?: string;
    bit_rate?: string;
  };
}

function mapRow(row: VideoMetadataRow | undefined): StoredVideoMetadata | null {
  if (!row) return null;
  return {
    absPath: row.abs_path,
    relPath: row.rel_path,
    code: row.code,
    sizeBytes: row.size_bytes,
    mtimeMs: row.mtime_ms,
    probedAt: row.probed_at,
    probeError: row.probe_error,
    durationSec: row.duration_sec,
    videoCodec: row.video_codec,
    videoBFrames: row.video_b_frames,
    width: row.width,
    height: row.height,
    videoBitrate: row.video_bitrate,
    playbackMode: row.playback_mode === "direct" || row.playback_mode === "transcode" ? row.playback_mode : null,
    partKind: row.part_kind === "part" || row.part_kind === "variant" || row.part_kind === "single" ? row.part_kind : null,
    partIndex: row.part_index,
    variantGroup: row.variant_group,
  };
}

function parseFiniteNumber(value: unknown): number | null {
  if (value == null || value === "N/A") return null;
  const n = typeof value === "number" ? value : Number(value);
  return Number.isFinite(n) && n > 0 ? n : null;
}

function parseNonNegativeNumber(value: unknown): number | null {
  if (value == null || value === "N/A") return null;
  const n = typeof value === "number" ? value : Number(value);
  return Number.isFinite(n) && n >= 0 ? n : null;
}

function isStatMatch(row: StoredVideoMetadata, sizeBytes: number, mtimeMs: number): boolean {
  return row.sizeBytes === sizeBytes && Math.abs(row.mtimeMs - mtimeMs) < 1;
}

export function getStoredVideoMetadata(absPath: string): StoredVideoMetadata | null {
  return mapRow(rawDb.prepare(`SELECT * FROM video_metadata WHERE abs_path = ?`).get(absPath) as VideoMetadataRow | undefined);
}

export function listStoredVideoMetadataForCode(code: string | null | undefined): StoredVideoMetadata[] {
  if (!code) return [];
  const rows = rawDb.prepare(`
    SELECT * FROM video_metadata
    WHERE upper(code) = upper(?)
    ORDER BY rel_path ASC
  `).all(code) as VideoMetadataRow[];
  return rows.map((row) => mapRow(row)).filter((row): row is StoredVideoMetadata => row !== null);
}

export function serializeVideoMetadata(meta: StoredVideoMetadata | null) {
  if (!meta) return null;
  return {
    absPath: meta.absPath,
    relPath: meta.relPath,
    code: meta.code,
    sizeBytes: meta.sizeBytes,
    mtimeMs: meta.mtimeMs,
    probedAt: meta.probedAt,
    probeError: meta.probeError,
    durationSec: meta.durationSec,
    videoCodec: meta.videoCodec,
    videoBFrames: meta.videoBFrames,
    width: meta.width,
    height: meta.height,
    videoBitrate: meta.videoBitrate,
    playbackMode: meta.playbackMode,
    partKind: meta.partKind,
    partIndex: meta.partIndex,
    variantGroup: meta.variantGroup,
  };
}

export async function syncVideoMetadataIndex(files: VideoFile[]): Promise<void> {
  const found = new Set(files.map((file) => file.abs));
  const upsert = rawDb.prepare(`
    INSERT INTO video_metadata (abs_path, rel_path, code, size_bytes, mtime_ms, dir_path)
    VALUES (?, ?, ?, ?, ?, ?)
    ON CONFLICT(abs_path) DO UPDATE SET
      rel_path = excluded.rel_path,
      code = excluded.code,
      dir_path = excluded.dir_path,
      probed_at = CASE
        WHEN video_metadata.size_bytes != excluded.size_bytes OR video_metadata.mtime_ms != excluded.mtime_ms THEN NULL
        ELSE video_metadata.probed_at
      END,
      probe_error = CASE
        WHEN video_metadata.size_bytes != excluded.size_bytes OR video_metadata.mtime_ms != excluded.mtime_ms THEN NULL
        ELSE video_metadata.probe_error
      END,
      duration_sec = CASE
        WHEN video_metadata.size_bytes != excluded.size_bytes OR video_metadata.mtime_ms != excluded.mtime_ms THEN NULL
        ELSE video_metadata.duration_sec
      END,
      video_codec = CASE
        WHEN video_metadata.size_bytes != excluded.size_bytes OR video_metadata.mtime_ms != excluded.mtime_ms THEN NULL
        ELSE video_metadata.video_codec
      END,
      video_b_frames = CASE
        WHEN video_metadata.size_bytes != excluded.size_bytes OR video_metadata.mtime_ms != excluded.mtime_ms THEN NULL
        ELSE video_metadata.video_b_frames
      END,
      width = CASE
        WHEN video_metadata.size_bytes != excluded.size_bytes OR video_metadata.mtime_ms != excluded.mtime_ms THEN NULL
        ELSE video_metadata.width
      END,
      height = CASE
        WHEN video_metadata.size_bytes != excluded.size_bytes OR video_metadata.mtime_ms != excluded.mtime_ms THEN NULL
        ELSE video_metadata.height
      END,
      video_bitrate = CASE
        WHEN video_metadata.size_bytes != excluded.size_bytes OR video_metadata.mtime_ms != excluded.mtime_ms THEN NULL
        ELSE video_metadata.video_bitrate
      END,
      playback_mode = CASE
        WHEN video_metadata.size_bytes != excluded.size_bytes OR video_metadata.mtime_ms != excluded.mtime_ms THEN NULL
        ELSE video_metadata.playback_mode
      END,
      size_bytes = excluded.size_bytes,
      mtime_ms = excluded.mtime_ms
  `);
  const deleteStale = rawDb.prepare(`DELETE FROM video_metadata WHERE abs_path = ?`);
  const tx = rawDb.transaction(() => {
    for (const file of files) {
      const last = Math.max(file.abs.lastIndexOf("/"), file.abs.lastIndexOf("\\"));
      const dir = last >= 0 ? file.abs.slice(0, last) : "";
      upsert.run(file.abs, file.rel, file.code, file.size, file.mtime, dir);
    }
    const rows = rawDb.prepare(`SELECT abs_path FROM video_metadata`).all() as Array<{ abs_path: string }>;
    for (const row of rows) {
      if (!found.has(row.abs_path)) deleteStale.run(row.abs_path);
    }
  });
  tx();
  classifyAndPersist(files);

  // Probe-data refresh runs in the background. Awaiting here used to
  // block rescan responses for minutes on libraries with many drifted
  // files (e.g. after a bulk rename). Each per-file probe completion
  // calls revalidatePath internally so detail pages update as soon as
  // their own video is fresh — no batch-level waiting.
  void reprobeDirtyFiles(files);
}

const REPROBE_CONCURRENCY = 2;

async function reprobeDirtyFiles(files: VideoFile[]): Promise<void> {
  let dirty: Array<{ abs_path: string }>;
  try {
    dirty = rawDb
      .prepare(`SELECT abs_path FROM video_metadata WHERE probed_at IS NULL AND probe_error IS NULL`)
      .all() as Array<{ abs_path: string }>;
  } catch (e) {
    console.error("[video] reprobe-dirty query failed:", e);
    return;
  }
  if (dirty.length === 0) return;

  const dirtySet = new Set(dirty.map((r) => r.abs_path));
  const targets = files.filter((f) => dirtySet.has(f.abs));
  if (targets.length === 0) return;

  // Process in chunks of REPROBE_CONCURRENCY. ffprobe is mostly waiting
  // on disk; small parallelism is enough.
  let cursor = 0;
  const workers: Promise<void>[] = [];
  // Throttle revalidation calls: a burst of 1000 path invalidations
  // would itself thrash. Coalesce so each batch of N codes triggers
  // one revalidate per code, deduped within a short window.
  const codesSeen = new Set<string>();
  for (let i = 0; i < REPROBE_CONCURRENCY; i++) {
    workers.push((async () => {
      while (cursor < targets.length) {
        const idx = cursor++;
        const file = targets[idx];
        if (!file) break;
        try {
          await probeVideoMetadata(file);
          if (!codesSeen.has(file.code)) {
            codesSeen.add(file.code);
            try { revalidatePath("/id/[code]", "page"); } catch { /* ignore */ }
          }
        } catch (e) {
          console.error(`[video] reprobe failed for ${file.abs}:`, e);
        }
      }
    })());
  }
  await Promise.all(workers).catch(() => { /* swallowed */ });
}

/**
 * Recompute part/variant classification for every file based on the
 * current `partSuffixPatterns` setting. Independent of probe data; safe
 * to run on every scan.
 */
function classifyAndPersist(files: VideoFile[]): void {
  const sources = getAppSetting("partSuffixPatterns") ?? [];
  const patterns = compilePatterns(sources);
  const byCode = new Map<string, VideoFile[]>();
  for (const f of files) {
    const arr = byCode.get(f.code);
    if (arr) arr.push(f);
    else byCode.set(f.code, [f]);
  }
  const update = rawDb.prepare(`
    UPDATE video_metadata SET part_kind = ?, part_index = ?, variant_group = ?
    WHERE abs_path = ?
  `);
  const tx = rawDb.transaction(() => {
    for (const group of byCode.values()) {
      const inputs = group.map((f) => ({
        key: f.abs,
        stem: stemOf(f.filename),
      }));
      const results = classifyGroup(inputs, patterns);
      for (const r of results) {
        update.run(r.partKind, r.partIndex, r.variantGroup, r.key);
      }
    }
  });
  tx();
}

function stemOf(filename: string): string {
  const ext = path.extname(filename);
  return ext ? filename.slice(0, -ext.length) : filename;
}

export interface SubtitleStreamInfo {
  index: number;
  codec: string;
  language: string | null;
  title: string | null;
  isImageBased: boolean;
  isTextBased: boolean;
}

const TEXT_SUBTITLE_CODECS = new Set(["subrip", "ass", "ssa", "mov_text", "webvtt", "text"]);
const IMAGE_SUBTITLE_CODECS = new Set(["hdmv_pgs_subtitle", "dvd_subtitle", "dvb_subtitle", "dvbsub", "pgssub"]);

interface FfprobeStream {
  index?: number;
  codec_type?: string;
  codec_name?: string;
  tags?: { language?: string; title?: string };
}

/** Enumerate subtitle streams in a container. Computed on demand — not
 *  persisted, since users frequently remux subs in/out and a stale list
 *  is worse than re-probing. Returns [] on error or missing ffprobe. */
export async function runFfprobeSubtitles(absPath: string): Promise<SubtitleStreamInfo[]> {
  return new Promise((resolve) => {
    const proc = spawn("ffprobe", [
      "-v", "error",
      "-select_streams", "s",
      "-show_entries", "stream=index,codec_name,codec_type:stream_tags=language,title",
      "-of", "json",
      absPath,
    ]);
    let out = "";
    let settled = false;
    const settle = (val: SubtitleStreamInfo[]) => { if (!settled) { settled = true; clearTimeout(t); resolve(val); } };
    const t = setTimeout(() => { try { proc.kill("SIGKILL"); } catch {} settle([]); }, PROBE_TIMEOUT_MS);
    proc.stdout?.on("data", (d) => { out += d.toString(); });
    proc.on("error", () => settle([]));
    proc.on("close", (code) => {
      if (code !== 0) { settle([]); return; }
      try {
        const json = JSON.parse(out) as { streams?: FfprobeStream[] };
        const streams = (json.streams ?? []).filter((s) => s.codec_type === "subtitle");
        const result: SubtitleStreamInfo[] = streams.map((s, i) => {
          const codec = (s.codec_name ?? "unknown").toLowerCase();
          return {
            // Use the per-codec_type ordinal — that's what ffmpeg's
            // 0:s:N mapping wants, NOT the absolute stream index.
            index: i,
            codec,
            language: typeof s.tags?.language === "string" ? s.tags.language : null,
            title: typeof s.tags?.title === "string" ? s.tags.title : null,
            isImageBased: IMAGE_SUBTITLE_CODECS.has(codec),
            isTextBased: TEXT_SUBTITLE_CODECS.has(codec),
          };
        });
        settle(result);
      } catch {
        settle([]);
      }
    });
  });
}

async function runFfprobe(absPath: string, signal?: AbortSignal): Promise<{
  durationSec: number | null;
  videoCodec: string | null;
  videoBFrames: number | null;
  width: number | null;
  height: number | null;
  videoBitrate: number | null;
}> {
  return new Promise((resolve, reject) => {
    const proc = spawn("ffprobe", [
      "-v", "error",
      "-select_streams", "v:0",
      "-show_entries", "stream=codec_name,width,height,bit_rate,has_b_frames:format=duration,bit_rate",
      "-of", "json",
      absPath,
    ]);
    let out = "";
    let err = "";
    let settled = false;

    const settle = (fn: () => void) => {
      if (settled) return;
      settled = true;
      clearTimeout(timeoutId);
      if (signal && abortHandler) signal.removeEventListener("abort", abortHandler);
      fn();
    };
    const kill = (message: string) => {
      try { proc.kill("SIGKILL"); } catch {}
      settle(() => reject(new Error(message)));
    };

    const timeoutId = setTimeout(() => kill("ffprobe timed out"), PROBE_TIMEOUT_MS);
    const abortHandler = signal ? () => kill("ffprobe aborted") : null;
    if (signal && abortHandler) {
      if (signal.aborted) { kill("ffprobe aborted"); return; }
      signal.addEventListener("abort", abortHandler, { once: true });
    }

    proc.stdout?.on("data", (d) => { out += d.toString(); });
    proc.stderr?.on("data", (d) => { err += d.toString(); });
    proc.on("error", (e) => settle(() => reject(e)));
    proc.on("close", (code) => {
      settle(() => {
        if (code !== 0) {
          reject(new Error(err.trim() || `ffprobe exited ${code}`));
          return;
        }
        try {
          const json = JSON.parse(out) as FfprobeJson;
          const stream = json.streams?.[0] ?? {};
          const streamBitrate = parseFiniteNumber(stream.bit_rate);
          const formatBitrate = parseFiniteNumber(json.format?.bit_rate);
          resolve({
            durationSec: parseFiniteNumber(json.format?.duration),
            videoCodec: typeof stream.codec_name === "string" ? stream.codec_name : null,
            videoBFrames: parseNonNegativeNumber(stream.has_b_frames),
            width: parseFiniteNumber(stream.width),
            height: parseFiniteNumber(stream.height),
            videoBitrate: streamBitrate ?? formatBitrate,
          });
        } catch (e) {
          reject(e);
        }
      });
    });
  });
}

export async function probeVideoMetadata(file: VideoFile, signal?: AbortSignal): Promise<StoredVideoMetadata> {
  const stat = await fs.stat(file.abs);
  const existing = getStoredVideoMetadata(file.abs);
  if (existing && isStatMatch(existing, stat.size, stat.mtimeMs)) {
    if (existing.probeError || existing.probedAt != null) return existing;
  }

  const base = {
    absPath: file.abs,
    relPath: file.rel,
    code: file.code,
    sizeBytes: stat.size,
    mtimeMs: stat.mtimeMs,
    playbackMode: existing?.playbackMode ?? null,
  };

  try {
    const probed = await runFfprobe(file.abs, signal);
    rawDb.prepare(`
      INSERT INTO video_metadata (
        abs_path, rel_path, code, size_bytes, mtime_ms, probed_at, probe_error,
        duration_sec, video_codec, video_b_frames, width, height, video_bitrate, playback_mode
      ) VALUES (?, ?, ?, ?, ?, ?, NULL, ?, ?, ?, ?, ?, ?, ?)
      ON CONFLICT(abs_path) DO UPDATE SET
        rel_path = excluded.rel_path,
        code = excluded.code,
        size_bytes = excluded.size_bytes,
        mtime_ms = excluded.mtime_ms,
        probed_at = excluded.probed_at,
        probe_error = NULL,
        duration_sec = excluded.duration_sec,
        video_codec = excluded.video_codec,
        video_b_frames = excluded.video_b_frames,
        width = excluded.width,
        height = excluded.height,
        video_bitrate = excluded.video_bitrate,
        playback_mode = excluded.playback_mode
    `).run(
      base.absPath, base.relPath, base.code, base.sizeBytes, base.mtimeMs, Date.now(),
      probed.durationSec, probed.videoCodec, probed.videoBFrames, probed.width, probed.height, probed.videoBitrate, base.playbackMode,
    );
  } catch (e) {
    rawDb.prepare(`
      INSERT INTO video_metadata (
        abs_path, rel_path, code, size_bytes, mtime_ms, probed_at, probe_error, playback_mode
      ) VALUES (?, ?, ?, ?, ?, ?, ?, ?)
      ON CONFLICT(abs_path) DO UPDATE SET
        rel_path = excluded.rel_path,
        code = excluded.code,
        size_bytes = excluded.size_bytes,
        mtime_ms = excluded.mtime_ms,
        probed_at = excluded.probed_at,
        probe_error = excluded.probe_error,
        playback_mode = excluded.playback_mode
    `).run(
      base.absPath, base.relPath, base.code, base.sizeBytes, base.mtimeMs, Date.now(),
      e instanceof Error ? e.message.slice(0, 500) : "ffprobe failed",
      base.playbackMode,
    );
  }

  return getStoredVideoMetadata(file.abs) ?? {
    ...base,
    probedAt: null,
    probeError: "metadata unavailable",
    durationSec: null,
    videoCodec: null,
    videoBFrames: null,
    width: null,
    height: null,
    videoBitrate: null,
    partKind: null,
    partIndex: null,
    variantGroup: null,
  };
}

export function setVideoPlaybackMode(file: VideoFile, mode: PlaybackMode | null): void {
  rawDb.prepare(`
    INSERT INTO video_metadata (abs_path, rel_path, code, size_bytes, mtime_ms, playback_mode)
    VALUES (?, ?, ?, ?, ?, ?)
    ON CONFLICT(abs_path) DO UPDATE SET
      rel_path = excluded.rel_path,
      code = excluded.code,
      size_bytes = excluded.size_bytes,
      mtime_ms = excluded.mtime_ms,
      playback_mode = excluded.playback_mode
  `).run(file.abs, file.rel, file.code, file.size, file.mtime, mode);
}

export function formatDuration(sec: number | null | undefined): string | null {
  if (sec == null || !Number.isFinite(sec) || sec <= 0) return null;
  const total = Math.round(sec);
  const h = Math.floor(total / 3600);
  const m = Math.floor((total % 3600) / 60);
  const s = total % 60;
  if (h > 0) return `${h}:${String(m).padStart(2, "0")}:${String(s).padStart(2, "0")}`;
  return `${m}:${String(s).padStart(2, "0")}`;
}

export function formatBitrate(bps: number | null | undefined): string | null {
  if (bps == null || !Number.isFinite(bps) || bps <= 0) return null;
  if (bps >= 1_000_000) return `${(bps / 1_000_000).toFixed(1)} Mbps`;
  if (bps >= 1_000) return `${Math.round(bps / 1_000)} Kbps`;
  return `${Math.round(bps)} bps`;
}

export function formatBytes(bytes: number | null | undefined): string | null {
  if (bytes == null || !Number.isFinite(bytes) || bytes < 0) return null;
  const units = ["B", "KB", "MB", "GB", "TB"];
  let n = bytes;
  let i = 0;
  while (n >= 1024 && i < units.length - 1) {
    n /= 1024;
    i++;
  }
  return `${i === 0 ? Math.round(n) : n.toFixed(n >= 10 ? 1 : 2)} ${units[i]}`;
}

export function formatResolution(width: number | null | undefined, height: number | null | undefined): string | null {
  if (!width || !height) return null;
  return `${width}x${height}`;
}

export function formatCodec(codec: string | null | undefined): string | null {
  if (!codec) return null;
  const map: Record<string, string> = {
    h264: "H.264",
    hevc: "HEVC",
    h265: "HEVC",
    av1: "AV1",
    vp9: "VP9",
    mpeg4: "MPEG-4",
  };
  return map[codec.toLowerCase()] ?? codec.toUpperCase();
}

export function formatVideoSummary(meta: StoredVideoMetadata | null | undefined): string | null {
  if (!meta || meta.probeError) return null;
  const parts = [
    formatResolution(meta.width, meta.height),
    formatCodec(meta.videoCodec),
    formatBitrate(meta.videoBitrate),
    formatBytes(meta.sizeBytes),
    formatDuration(meta.durationSec),
  ].filter((part): part is string => Boolean(part));
  return parts.length > 0 ? parts.join(" · ") : null;
}