nb-swap.mjs

// Swap GuessWord's embedding payload: replace GWD.vec/min/max in guessword.html// with int8-quantized ConceptNet Numberbatch vectors from build/nb-vecs.f32.// Words listed in build/nb-report2.json:dropped (corpus junk with no NB vector)// are removed from the vocabulary (words/level/N rebuilt). Everything else in// the file stays byte-identical. Refuses to run on any remaining zero row.import fs from "node:fs";import path from "node:path";import { fileURLToPath } from "node:url";const ROOT = path.dirname(fileURLToPath(import.meta.url));const HTML = path.join(ROOT, "..", "guessword.html");const html = fs.readFileSync(HTML, "utf8");const lines = html.split("\n");const li = lines.findIndex(l => l.startsWith("const GWD="));if (li < 0) throw new Error("GWD line not found");const GWD = JSON.parse(lines[li].slice("const GWD=".length, -1));const N0 = GWD.N, D = GWD.D;const words0 = GWD.words.split("|");const f32all = new Float32Array(fs.readFileSync(path.join(ROOT, "nb-vecs.f32")).buffer);if (f32all.length !== N0 * D) throw new Error(`nb-vecs.f32 size ${f32all.length} != ${N0 * D}`);// drop corpus-junk words that have no Numberbatch vectorconst rep2 = JSON.parse(fs.readFileSync(path.join(ROOT, "nb-report2.json"), "utf8"));const dropSet = new Set(rep2.dropped);const keep = [];for (let i = 0; i < N0; i++) if (!dropSet.has(words0[i])) keep.push(i);const N = keep.length;console.log(`vocab ${N0} -> ${N} (dropped ${N0 - N})`);const words = keep.map(i => words0[i]);const level = keep.map(i => GWD.level[i]).join("");const f32 = new Float32Array(N * D);for (let k = 0; k < N; k++) f32.set(f32all.subarray(keep[k] * D, keep[k] * D + D), k * D);// zero-row check (should be impossible after fallback + drop)const missing = [];for (let i = 0; i < N; i++) {  let nz = false;  for (let d = 0; d < D; d++) if (f32[i * D + d] !== 0) { nz = true; break; }  if (!nz) missing.push(words[i]);}if (missing.length) {  console.error(`ABORT: ${missing.length} words have no Numberbatch vector:`);  console.error(missing.join(" "));  process.exit(1);}// per-dimension affine min/max, matching the runtime decoder:// value = min[d] + byte/255 * (max[d]-min[d])const min = new Float64Array(D).fill(Infinity);const max = new Float64Array(D).fill(-Infinity);for (let i = 0; i < N; i++) {  const off = i * D;  for (let d = 0; d < D; d++) {    const v = f32[off + d];    if (v < min[d]) min[d] = v;    if (v > max[d]) max[d] = v;  }}const bytes = new Uint8Array(N * D);for (let i = 0; i < N; i++) {  const off = i * D;  for (let d = 0; d < D; d++) {    const span = max[d] - min[d] || 1;    bytes[off + d] = Math.round((f32[off + d] - min[d]) / span * 255);  }}GWD.N = N;GWD.words = words.join("|");GWD.level = level;GWD.vec = Buffer.from(bytes).toString("base64");GWD.min = Array.from(min, v => +v.toFixed(8));GWD.max = Array.from(max, v => +v.toFixed(8));const newLine = "const GWD=" + JSON.stringify(GWD) + ";";// keep the file's original line ending style for that lineif (lines[li].endsWith("\r")) lines[li] = newLine + "\r"; else lines[li] = newLine;fs.writeFileSync(HTML, lines.join("\n"));console.log(`swapped: vec b64 ${GWD.vec.length} chars (expect ${Math.ceil(N * D / 3) * 4}), min/max ${D} each`);console.log("GWD now on line", li + 1);