test(04-01): RED — gazetteer schema/license gates (G1-G11) + three-environment loader proof (L1-L3)

- G1-G6: dataset invariants — schema, slug-id uniqueness, canonical-name law ('KLCC' is an alias, never a name), DUKE-as-alias, KL bbox, the ODbL header traveling inside the JSON - G7-G9: the licensing gate — verbatim GPL-3.0 LICENSE, LICENSES.md, the README no longer claiming 'all rights reserved' - G10-G11: no logging/storage anywhere in the module (D-15); the loader builds its index inside a function, never at module top level - L1-L3: the phase's top mechanical risk — bare-Node loading with the SRCH-01/negative probes, the static JSON import law, esbuild INLINING the dataset + the wrangler dry-run build guard 13 cases fail while the dataset, loader, LICENSE files, and README fix are absent — the RED gate.
2026-10-24 00:32:57 +08:00
parent 6368bad3d5
commit f1b31ca7ed
2 changed files with 310 additions and 0 deletions
@@ -0,0 +1,175 @@
+#!/usr/bin/env node
+/*
+ * gazetteer.test.mjs — dataset invariants + licensing gates (D-04, D-06,
+ * D-07, D-08, D-15).
+ * --------------------------------------------------------------------
+ * RED phase (plan 04-01 task 3): the data-dependent cases fail while
+ * kl-gazetteer.json does not exist, and the licensing cases fail while
+ * LICENSE / LICENSES.md / the README correction are absent.
+ *
+ *   G1  entry schema — id/name/aliases/lat/lng/category/address
+ *   G2  id uniqueness + slug charset (/^[a-z0-9-]{1,64}$/)
+ *   G3  canonical names vs shorthand — "Suria KLCC" is the NAME, "KLCC"
+ *       is an ALIAS (RESEARCH Verification A3: the alias list IS the product)
+ *   G4  "DUKE" is an alias of the-duke; no entry is NAMED "DUKE"
+ *   G5  coordinates inside the KL metro bbox
+ *   G6  the ODbL header travels INSIDE the JSON
+ *   G7  LICENSE exists at the repo root and is the GPL-3.0 text
+ *   G8  LICENSES.md documents BOTH licences + the OSM attribution
+ *   G9  the README no longer claims "all rights reserved"
+ *   G10 the shared module has no logging and no storage (D-15/T-04-01/02)
+ *   G11 the loader is lazy — no index build at module top level (T-04-04)
+ *
+ * Run:  npm run search:test     (exit 0 = all green)
+ */
+import { strict as assert } from "node:assert";
+import { readFileSync, readdirSync } from "node:fs";
+import { join, dirname } from "node:path";
+import { fileURLToPath } from "node:url";
+import { test } from "./harness.mjs";
+
+const here = dirname(fileURLToPath(import.meta.url)); // search/test
+const srcDir = join(here, "..", "src"); // search/src
+const repoRoot = join(here, "..", ".."); // repo root
+
+/** Closed category set the seed contract pins. */
+const CATEGORIES = new Set([
+  "mall", "office-tower", "park", "station", "hospital", "road", "highway",
+  "landmark", "convention", "village", "education", "government",
+]);
+
+function loadGazetteerJson() {
+  const text = readFileSync(join(srcDir, "data", "kl-gazetteer.json"), "utf8");
+  return JSON.parse(text);
+}
+
+/** Strip block comments and full-line // comments. */
+function stripComments(text) {
+  return text.replace(/\/\*[\s\S]*?\*\//g, " ").replace(/^\s*\/\/.*$/gm, " ");
+}
+
+/** Every .ts file under search/src (recursive). */
+function srcTsFiles() {
+  const files = readdirSync(srcDir, { recursive: true });
+  const out = [];
+  for (const f of files) {
+    if (typeof f === "string" && f.endsWith(".ts")) out.push(join(srcDir, f));
+  }
+  return out;
+}
+
+test("G1 entry schema — every entry is complete, finite, and categorised", () => {
+  const { entries } = loadGazetteerJson();
+  assert.ok(Array.isArray(entries) && entries.length >= 30, `seed floor is 30 entries, got ${entries.length}`);
+  for (const e of entries) {
+    assert.ok(typeof e.id === "string" && e.id.length > 0, `id: ${e.id}`);
+    assert.ok(typeof e.name === "string" && e.name.length > 0, `name of ${e.id}`);
+    assert.ok(Array.isArray(e.aliases) && e.aliases.length > 0, `aliases of ${e.id}`);
+    assert.ok(e.aliases.includes(e.name), `${e.id}: aliases must include the canonical name`);
+    assert.ok(Number.isFinite(e.lat) && Number.isFinite(e.lng), `coords of ${e.id}`);
+    assert.ok(CATEGORIES.has(e.category), `category of ${e.id}: ${e.category}`);
+    assert.ok(typeof e.address === "string" && e.address.length > 0, `address of ${e.id}`);
+  }
+});
+
+test("G2 id uniqueness + slug charset", () => {
+  const { entries } = loadGazetteerJson();
+  const ids = entries.map((e) => e.id);
+  assert.equal(new Set(ids).size, ids.length, "ids must be unique — they are the deep-link currency (D-11)");
+  for (const id of ids) {
+    assert.match(id, /^[a-z0-9-]{1,64}$/, `id charset: ${id}`);
+  }
+});
+
+test("G3 canonical name vs shorthand — 'Suria KLCC' is the name, 'KLCC' is an alias", () => {
+  const { entries } = loadGazetteerJson();
+  const klcc = entries.find((e) => e.id === "suria-klcc");
+  assert.notEqual(klcc, undefined, "suria-klcc must exist");
+  assert.equal(klcc.name, "Suria KLCC");
+  assert.ok(klcc.aliases.includes("KLCC"), "'KLCC' must be an explicit alias — it is not the OSM name");
+  assert.ok(!entries.some((e) => e.name === "KLCC"), "no entry may be NAMED the shorthand");
+});
+
+test("G4 'DUKE' is an alias; no entry is named the acronym", () => {
+  const { entries } = loadGazetteerJson();
+  const duke = entries.find((e) => e.id === "the-duke");
+  assert.notEqual(duke, undefined, "the-duke must exist");
+  assert.ok(duke.aliases.includes("DUKE"), "the highway acronym rides as an alias");
+  assert.notEqual(duke.name, "DUKE", "DUKE is an acronym, not a landmark name");
+  for (const e of entries) {
+    assert.notEqual(e.name, "DUKE");
+  }
+});
+
+test("G5 coordinates sit inside the KL metro bbox", () => {
+  const { entries } = loadGazetteerJson();
+  for (const e of entries) {
+    assert.ok(e.lat >= 2.9 && e.lat <= 3.4, `lat of ${e.id}: ${e.lat}`);
+    assert.ok(e.lng >= 101.4 && e.lng <= 102.0, `lng of ${e.id}: ${e.lng}`);
+  }
+});
+
+test("G6 the ODbL header travels inside the JSON (ODbL §4.4 share-alike)", () => {
+  const data = loadGazetteerJson();
+  assert.ok(String(data._license).includes("ODbL"), "_license must name ODbL");
+  assert.ok(String(data._attribution).includes("OpenStreetMap"), "_attribution must credit OpenStreetMap");
+  assert.ok(
+    String(data._notice).includes("opendatacommons.org/licenses/odbl"),
+    "_notice must link the ODbL text",
+  );
+  assert.ok(data.version >= 1);
+});
+
+test("G7 LICENSE exists at the repo root and is the GPL-3.0 text (D-04)", () => {
+  const license = readFileSync(join(repoRoot, "LICENSE"), "utf8");
+  assert.ok(license.includes("GNU GENERAL PUBLIC LICENSE"));
+  assert.ok(license.includes("Version 3, 29 June 2007"));
+  assert.ok(license.includes("TERMS AND CONDITIONS"));
+});
+
+test("G8 LICENSES.md documents both licences + the OSM attribution", () => {
+  const doc = readFileSync(join(repoRoot, "LICENSES.md"), "utf8");
+  for (const needle of ["GPL-3.0", "ODbL", "OpenStreetMap"]) {
+    assert.ok(doc.includes(needle), `LICENSES.md must name ${needle}`);
+  }
+});
+
+test("G9 the README licence section is no longer self-contradictory", () => {
+  const readme = readFileSync(join(repoRoot, "README.md"), "utf8");
+  assert.ok(readme.includes("GPL-3.0"), "README must point at the GPL-3.0 LICENSE");
+  assert.ok(!readme.includes("all rights reserved"), "the private-repo claim must be gone");
+});
+
+test("G10 the shared module has no logging and no storage surface (D-15)", () => {
+  const banned = ["console.log", "localStorage", "sessionStorage", "indexedDB", "document.cookie"];
+  for (const file of srcTsFiles()) {
+    const stripped = stripComments(readFileSync(file, "utf8"));
+    for (const token of banned) {
+      assert.ok(!stripped.includes(token), `${file} must not contain ${token} — the query text is user payload`);
+    }
+  }
+});
+
+test("G11 the loader is lazy — the index is built inside a function, never at module scope", async () => {
+  const loaderPath = join(srcDir, "gazetteer.ts");
+  const stripped = stripComments(readFileSync(loaderPath, "utf8"));
+  const firstFn = stripped.search(/\bfunction\b/);
+  assert.ok(firstFn >= 0, "gazetteer.ts must declare functions");
+  const topLevel = stripped.slice(0, firstFn);
+  assert.ok(!topLevel.includes("new Map("), "no index construction at module top level (Workers 1s startup, T-04-04)");
+  assert.ok(!topLevel.includes("JSON."), "no JSON traversal at module top level");
+  assert.ok(!topLevel.includes("for ("), "no dataset walking at module top level");
+
+  const mod = await import("../src/gazetteer.ts");
+  for (const fn of ["getEntries", "getEntryById", "getIndex"]) {
+    assert.equal(typeof mod[fn], "function", `export ${fn} must be a function`);
+  }
+  const entries = mod.getEntries();
+  assert.ok(entries.length >= 30);
+  const byId = mod.getEntryById("suria-klcc");
+  assert.notEqual(byId, null, "honest lookup finds the seed anchor");
+  assert.equal(mod.getEntryById("not-a-real-id"), null, "unknown id -> honest null");
+  const index = mod.getIndex();
+  assert.ok(index instanceof Map);
+  assert.ok(index.has("klcc"), "the index is keyed by NORMALIZED aliases (D-09 symmetry)");
+});
@@ -0,0 +1,135 @@
+#!/usr/bin/env node
+/*
+ * loader-env.test.mjs — the THREE-ENVIRONMENT loader proof (D-07).
+ * ----------------------------------------------------------------
+ * THE phase's top mechanical risk (RESEARCH): the JSON dataset must be
+ * reachable identically from (1) the bare-Node corpus runner, (2) the
+ * Vite bundle, and (3) the wrangler/esbuild Worker bundle — Workers have
+ * no fs, so the dataset MUST be a static build-time import, never a
+ * runtime fetch.
+ *
+ *   L1  bare Node (type-stripping): the module graph loads, the dataset
+ *       parses, search('KLCC') resolves suria-klcc, out-of-area is refused
+ *   L2  static JSON import in the loader + (once plan 04-04 lands) the
+ *       Vite bundle carries the dataset marker; until then that half is
+ *       SKIPPED visibly while the static-import half is still asserted
+ *   L3  esbuild — the bundler family wrangler uses — INLINES the dataset
+ *       (its bundle carries the 'suria-klcc' marker), and the Worker
+ *       dry-run build guard exits 0
+ *
+ * Run:  npm run search:test     (exit 0 = all green)
+ */
+import { strict as assert } from "node:assert";
+import { spawnSync } from "node:child_process";
+import { existsSync, mkdirSync, readdirSync, readFileSync } from "node:fs";
+import path from "node:path";
+import { fileURLToPath } from "node:url";
+import { test } from "./harness.mjs";
+
+const here = path.dirname(fileURLToPath(import.meta.url)); // search/test
+const searchDir = path.resolve(here, ".."); // search/
+const repoRoot = path.resolve(searchDir, ".."); // repo root
+const esbuildBin = path.join(repoRoot, "node_modules", ".bin", "esbuild");
+const wranglerBin = path.join(repoRoot, "node_modules", ".bin", "wrangler");
+
+// Lazy index load — missing module => per-case ERR_MODULE_NOT_FOUND (the RED state).
+let indexPromise = null;
+function loadIndexModule() {
+  if (!indexPromise) {
+    indexPromise = import("../src/index.ts");
+  }
+  return indexPromise;
+}
+
+test("L1 bare Node (type-stripping) — the loader works with zero infrastructure", async () => {
+  const mod = await loadIndexModule();
+  assert.ok(mod.getEntries().length >= 30, "the dataset parses in bare Node");
+  const source = mod.createGazetteerSource();
+  assert.equal(source.name, "gazetteer");
+  const hits = await source.search({ text: "KLCC" });
+  assert.ok(hits.length > 0, "KLCC must resolve");
+  assert.equal(hits[0].entry.id, "suria-klcc", "SRCH-01: KLCC resolves to suria-klcc at rank 1");
+  assert.equal((await source.search({ text: "Kuching" })).length, 0, "out-of-area is refused honestly (D-13)");
+  assert.equal((await source.search({ text: "Penang" })).length, 0, "out-of-area is refused honestly (D-13)");
+  assert.equal(source.byId("suria-klcc")?.name, "Suria KLCC", "byId is synchronous and honest");
+  assert.equal(source.byId("../etc/passwd"), null, "byId rejects non-slug ids");
+});
+
+test("L2 static JSON import in the loader (Vite-bundle half guarded until 04-04)", async () => {
+  const loaderPath = path.join(searchDir, "src", "gazetteer.ts");
+  const loaderSrc = readFileSync(loaderPath, "utf8");
+  assert.match(
+    loaderSrc,
+    /import raw from "\.\/data\/kl-gazetteer\.json"/,
+    "the dataset must be a STATIC import (build-time inlined)",
+  );
+  assert.ok(!/fetch\(/.test(loaderSrc), "a runtime fetch would hard-fail in a Worker");
+  assert.ok(!/readFileSync/.test(loaderSrc), "Workers have no fs — never read the dataset at runtime");
+
+  // The Vite-bundle half is only assertable once plan 04-04 wires web/ to
+  // @araharah/search. Until then: visible skip, static-import half asserted.
+  const webSrcDir = path.join(repoRoot, "web", "src");
+  let webImportsSearch = false;
+  try {
+    for (const entry of readdirSync(webSrcDir, { recursive: true })) {
+      const rel = String(entry);
+      if (rel.endsWith(".ts") || rel.endsWith(".tsx")) {
+        if (readFileSync(path.join(webSrcDir, rel), "utf8").includes("@araharah/search")) {
+          webImportsSearch = true;
+          break;
+        }
+      }
+    }
+  } catch {
+    // no web/src — treat as not-yet-landed
+  }
+  if (!webImportsSearch) {
+    console.log("SKIP L2 (web-bundle marker) — web/ does not import @araharah/search yet (lands in plan 04-04); the static-import half is asserted above");
+    return;
+  }
+  const distAssets = path.join(repoRoot, "web", "dist", "assets");
+  assert.ok(existsSync(distAssets), "web/dist/assets must exist once web imports the search module");
+  let marker = false;
+  for (const file of readdirSync(distAssets)) {
+    if (file.endsWith(".js") && readFileSync(path.join(distAssets, file), "utf8").includes("suria-klcc")) {
+      marker = true;
+      break;
+    }
+  }
+  assert.ok(marker, "the Vite bundle must inline the gazetteer (build-time, never fetched)");
+});
+
+test("L3 esbuild inlines the gazetteer; the wrangler dry-run build guard exits 0", () => {
+  // — the inline half (provable at wave 1): bundle the search module with the
+  //   SAME bundler family wrangler uses and grep the output for the marker —
+  const bundleDir = "/tmp/opencode/araharah-search-bundle";
+  mkdirSync(bundleDir, { recursive: true });
+  const outfile = path.join(bundleDir, "index.js");
+  const bundle = spawnSync(
+    esbuildBin,
+    [path.join(searchDir, "src", "index.ts"), "--bundle", `--outfile=${outfile}`, "--log-level=error"],
+    { encoding: "utf8", timeout: 60000 },
+  );
+  assert.equal(bundle.status, 0, `esbuild bundle failed: ${bundle.stderr || bundle.stdout}`);
+  const bundled = readFileSync(outfile, "utf8");
+  assert.ok(
+    bundled.includes("suria-klcc"),
+    "a static JSON import MUST be inlined at build time — a runtime fetch would be a hard Worker failure",
+  );
+
+  // — the Worker dry-run half (build regression guard ONLY): the bundle-
+  //   marker grep for the WORKER bundle is plan 04-03 task 2's criterion,
+  //   after bot.ts imports createGazetteerSource() —
+  if (!existsSync(path.join(repoRoot, "web", "dist", "index.html"))) {
+    console.log("SKIP L3 (wrangler dry-run) — web/dist is not built; run npm run build first");
+    return;
+  }
+  const dryrunDir = "/tmp/opencode/araharah-worker-dryrun";
+  mkdirSync(dryrunDir, { recursive: true });
+  const dry = spawnSync(wranglerBin, ["deploy", "--dry-run", `--outdir=${dryrunDir}`], {
+    cwd: path.join(repoRoot, "bot"),
+    encoding: "utf8",
+    timeout: 180000,
+  });
+  assert.equal(dry.status, 0, `wrangler dry-run failed: ${(dry.stderr || dry.stdout).slice(-800)}`);
+});