// Vendors finrod's browser dependencies into ../static/vendor/ so the running
// container serves everything locally — no esm.sh / CDN dependency at runtime.
//
//   node build.mjs
//
// Outputs:
//   static/vendor/codemirror.js            — esbuild bundle (view+state+commands)
//   static/vendor/cropper.esm.js           — cropperjs ESM (dependency-free)
//   static/vendor/cropper.min.css          — cropperjs stylesheet
//   static/vendor/tesseract/tesseract.esm.min.js  — Tesseract.js main thread
//   static/vendor/tesseract/worker.min.js         — Tesseract.js worker
//   static/vendor/tesseract/tesseract-core*.{js,wasm}  — wasm OCR core
//   static/vendor/tesseract/lang/<lang>.traineddata.gz — default language packs
import * as esbuild from "esbuild";
import { createRequire } from "node:module";
import { cp, mkdir, rm, readdir, writeFile } from "node:fs/promises";
import { dirname, join } from "node:path";
import { fileURLToPath } from "node:url";

const require = createRequire(import.meta.url);
const here = dirname(fileURLToPath(import.meta.url));
const OUT = join(here, "..", "static", "vendor");

// Only Dutch is baked in — finrod's real use is OCR'ing Dutch source text.
// Any other OCR language falls back to on-demand download from Tesseract's
// default CDN. These packs are served locally but only fetched by the browser
// when OCR is actually used (the OCR code path dynamic-imports Tesseract).
const LANGS = ["nld"];
const TESSDATA_BASE = "https://tessdata.projectnaptha.com/4.0.0";

const pkgDir = (name) => dirname(require.resolve(`${name}/package.json`));

await rm(OUT, { recursive: true, force: true });
await mkdir(join(OUT, "tesseract", "lang"), { recursive: true });

// 1. CodeMirror — bundle the whole graph into one ESM file.
await esbuild.build({
  entryPoints: [join(here, "src", "codemirror.js")],
  bundle: true,
  format: "esm",
  minify: true,
  legalComments: "none",
  outfile: join(OUT, "codemirror.js"),
});
console.log("✓ codemirror.js");

// 2. cropperjs — ESM build + CSS, both dependency-free, just copy.
const cropper = pkgDir("cropperjs");
await cp(join(cropper, "dist", "cropper.esm.js"), join(OUT, "cropper.esm.js"));
await cp(join(cropper, "dist", "cropper.min.css"), join(OUT, "cropper.min.css"));
console.log("✓ cropperjs");

// 3. Tesseract.js — ship the prebuilt worker + wasm core (do NOT bundle; it
//    spawns a Web Worker and streams wasm, which expects real dist files).
const tess = pkgDir("tesseract.js");
await cp(join(tess, "dist", "tesseract.esm.min.js"), join(OUT, "tesseract", "tesseract.esm.min.js"));
await cp(join(tess, "dist", "worker.min.js"), join(OUT, "tesseract", "worker.min.js"));

const core = pkgDir("tesseract.js-core");
for (const f of await readdir(core)) {
  if (f.endsWith(".js") || f.endsWith(".wasm")) {
    await cp(join(core, f), join(OUT, "tesseract", f));
  }
}
console.log("✓ tesseract engine + core");

// Language packs — fetched at build time so the runtime image is self-contained.
for (const lang of LANGS) {
  const url = `${TESSDATA_BASE}/${lang}.traineddata.gz`;
  const res = await fetch(url);
  if (!res.ok) throw new Error(`Failed to fetch ${url}: ${res.status}`);
  const buf = Buffer.from(await res.arrayBuffer());
  await writeFile(join(OUT, "tesseract", "lang", `${lang}.traineddata.gz`), buf);
  console.log(`✓ tessdata ${lang} (${(buf.length / 1e6).toFixed(1)} MB)`);
}

console.log("vendor build complete →", OUT);