@jjlmoya/utils-streaming 1.16.0 → 1.18.0

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@jjlmoya/utils-streaming",
-    "version": "1.16.0",
+    "version": "1.18.0",
     "type": "module",
     "main": "./src/index.ts",
     "types": "./src/index.ts",
@@ -28,7 +28,7 @@
         "check": "astro check",
         "type-check": "astro check",
         "test": "vitest run",
-        "preversion": "npm run lint && npm run test",
+        "preversion": "npm run lint && npm run test && npm run build",
         "postversion": "git push && git push --tags",
         "patch": "npm version patch",
         "minor": "npm version minor",

package/src/tests/seo_parity.test.ts

@@ -0,0 +1,60 @@
+import { describe, it, expect } from 'vitest';
+import { ALL_ENTRIES } from '../entries';
+import type { KnownLocale } from '../types';
+
+interface ExpectedCounts {
+  seo: number;
+  faq: number;
+  howTo: number;
+}
+
+function countItems(arr: unknown[] | undefined): number {
+  return arr?.length ?? 0;
+}
+
+async function verifyLocaleParity(
+  entry: typeof ALL_ENTRIES[number],
+  loc: KnownLocale,
+  expected: ExpectedCounts,
+): Promise<void> {
+  const locContent = await entry.i18n[loc]?.();
+  expect(locContent, `Locale ${loc} missing content`).toBeDefined();
+
+  const locSeoCount = countItems(locContent?.seo);
+  const locFaqCount = countItems(locContent?.faq);
+  const locHowToCount = countItems(locContent?.howTo);
+
+  expect(
+    locSeoCount,
+    `Locale ${loc} SEO sections count (${locSeoCount}) must match EN (${expected.seo})`,
+  ).toBe(expected.seo);
+  expect(
+    locFaqCount,
+    `Locale ${loc} FAQ items count (${locFaqCount}) must match EN (${expected.faq})`,
+  ).toBe(expected.faq);
+  expect(
+    locHowToCount,
+    `Locale ${loc} HowTo steps count (${locHowToCount}) must match EN (${expected.howTo})`,
+  ).toBe(expected.howTo);
+}
+
+describe('SEO & i18n Structural Parity Suite', () => {
+  ALL_ENTRIES.forEach((entry) => {
+    describe(`Tool: ${entry.id}`, () => {
+      it('all 15 locales should have identical SEO section counts and types as English', async () => {
+        const enContent = await entry.i18n.en?.();
+        expect(enContent).toBeDefined();
+        const expected: ExpectedCounts = {
+          seo: countItems(enContent?.seo),
+          faq: countItems(enContent?.faq),
+          howTo: countItems(enContent?.howTo),
+        };
+
+        const locales = Object.keys(entry.i18n) as KnownLocale[];
+        for (const loc of locales) {
+          await verifyLocaleParity(entry, loc, expected);
+        }
+      });
+    });
+  });
+});

package/src/tests/spanish_leakage.test.ts

@@ -0,0 +1,153 @@
+import { describe, expect, it } from 'vitest';
+import { ALL_ENTRIES } from '../entries';
+
+const STRUCTURAL_KEYS = new Set([
+  'id',
+  'slug',
+  'url',
+  'icon',
+  'image',
+  'imageUrl',
+  'keywords',
+  'category',
+  'tags',
+  'toolId',
+  'type',
+  'locale',
+  'language',
+  'direction',
+]);
+
+const TRANSLATABLE_KEYS = [
+  'title',
+  'description',
+  'faqTitle',
+  'faq',
+  'howTo',
+  'seo',
+  'schemas',
+] as const;
+
+const SPANISH_MARKERS = [
+  ['sangre', /\bsangre\b/gi],
+  ['molino', /\bmolino\b/gi],
+  ['grano', /\bgrano\b/gi],
+  ['paladar', /\bpaladar\b/gi],
+  ['cuchillas', /\bcuchillas\b/gi],
+  ['trozos', /\btrozos\b/gi],
+  ['frescura', /\bfrescura\b/gi],
+  ['ingresa', /\bingresa\b/gi],
+  ['selecciona', /\bselecciona\b/gi],
+  ['herramienta', /\bherramienta\b/gi],
+  ['según', /\bsegún\b/gi],
+  ['después', /\bdespués\b/gi],
+  ['puedes', /\bpuedes\b/gi],
+  ['debes', /\bdebes\b/gi],
+  ['tus', /\btus\b/gi],
+  ['a la vez', /\ba la vez\b/gi],
+  ['los datos', /\blos datos\b/gi],
+  ['las opciones', /\blas opciones\b/gi],
+  ['el resultado', /\bel resultado\b/gi],
+  ['método de extracción', /\bmétodo de extracción\b/gi],
+  ['uniformidad del molino', /\buniformidad del molino\b/gi],
+  ['café recién tostado', /\bcafé recién tostado\b/gi],
+] as const;
+
+type UnknownRecord = Record<string, unknown>;
+
+function normalize(value: string): string {
+  return value
+    .replace(/<[^>]*>/g, ' ')
+    .replace(/&nbsp;/gi, ' ')
+    .replace(/\s+/g, ' ')
+    .trim()
+    .toLocaleLowerCase('es');
+}
+
+function isTechnicalInvariant(text: string): boolean {
+  return [
+    'data:image/svg+xml;base64',
+    'background-image: url',
+    '.layout-playground {',
+    'const samplerate =',
+  ].some((pattern) => text.includes(pattern)) || (text.includes('presets') && text.includes('hz'));
+}
+
+function collectString(value: string, output: string[]): void {
+  const text = normalize(value);
+  if (!isTechnicalInvariant(text) && text.length >= 20) output.push(text);
+}
+
+function collectObject(value: UnknownRecord, output: string[]): void {
+  Object.entries(value).forEach(([childKey, childValue]) => {
+    collectText(childValue, output, childKey);
+  });
+}
+
+function collectText(value: unknown, output: string[], key?: string): void {
+  if (key && STRUCTURAL_KEYS.has(key)) return;
+  if (typeof value === 'string') return collectString(value, output);
+  if (Array.isArray(value)) return value.forEach((item) => collectText(item, output));
+  if (value && typeof value === 'object') collectObject(value as UnknownRecord, output);
+}
+
+async function loadText(loader: unknown): Promise<string[]> {
+  if (typeof loader !== 'function') return [];
+  const module = await (loader as () => Promise<unknown>)();
+  const output: string[] = [];
+  if (module && typeof module === 'object') {
+    const record = module as UnknownRecord;
+    for (const key of TRANSLATABLE_KEYS) {
+      collectText(record[key], output, key);
+    }
+  }
+  return output;
+}
+
+function findSpanishMarkers(text: string[]): string[] {
+  const corpus = text.join(' ');
+  return SPANISH_MARKERS.flatMap(([label, pattern]) =>
+    pattern.test(corpus) ? [label] : [],
+  );
+}
+
+function findCopiedFragments(spanish: string[], translated: string[]): string[] {
+  const corpus = translated.join(' ');
+  return spanish
+    .filter((fragment) => fragment.length >= 80 && corpus.includes(fragment))
+    .sort((a, b) => b.length - a.length)
+    .slice(0, 3);
+}
+
+describe('Locales must not contain copied Spanish content', () => {
+  for (const entry of ALL_ENTRIES) {
+    it(`${entry.id} has no untranslated Spanish blocks`, async () => {
+      const spanish = await loadText(entry.i18n.es);
+      const failures: string[] = [];
+
+      for (const [locale, loader] of Object.entries(entry.i18n)) {
+        if (locale === 'es') continue;
+
+        const translated = await loadText(loader);
+        const copiedFragments = findCopiedFragments(spanish, translated);
+        const markerHits = findSpanishMarkers(translated);
+
+        if (copiedFragments.length > 0 || markerHits.length >= 2) {
+          const details = [
+            copiedFragments.length > 0
+              ? `copied fragments: ${copiedFragments
+                  .map((fragment) => JSON.stringify(fragment.slice(0, 120)))
+                  .join(', ')}`
+              : '',
+            markerHits.length >= 2 ? `Spanish markers: ${markerHits.join(', ')}` : '',
+          ]
+            .filter(Boolean)
+            .join('; ');
+          failures.push(`${locale}: ${details}`);
+        }
+      }
+
+      expect(failures, failures.join('\n')).toEqual([]);
+    });
+  }
+});

package/src/tests/translation_copy.test.ts

@@ -0,0 +1,124 @@
+import { describe, expect, it } from 'vitest';
+import { ALL_ENTRIES } from '../entries';
+
+const COPY_THRESHOLD = 0.9;
+
+const STRUCTURAL_KEYS = new Set([
+  '@context',
+  '@type',
+  'applicationCategory',
+  'columns',
+  'highlight',
+  'icon',
+  'level',
+  'operatingSystem',
+  'position',
+  'positive',
+  'price',
+  'priceCurrency',
+  'slug',
+  'trend',
+  'type',
+  'url',
+  'value',
+  'variant',
+]);
+
+function normalizeText(value: string): string {
+  return value
+    .replace(/<[^>]*>/g, ' ')
+    .replace(/&(?:amp|lt|gt|quot|apos|nbsp);/gi, ' ')
+    .replace(/[\u2018\u2019]/g, "'")
+    .replace(/[\u201c\u201d]/g, '"')
+    .replace(/\s+/g, ' ')
+    .trim()
+    .toLocaleLowerCase();
+}
+
+function collectText(value: unknown, path: string, parts: string[]): void {
+  if (typeof value === 'string') {
+    const normalized = normalizeText(value);
+    if (normalized.length >= 2) parts.push(normalized);
+    return;
+  }
+
+  if (Array.isArray(value)) {
+    value.forEach((item, index) => collectText(item, `${path}[${index}]`, parts));
+    return;
+  }
+
+  if (!value || typeof value !== 'object') return;
+
+  Object.entries(value).forEach(([key, child]) => {
+    if (STRUCTURAL_KEYS.has(key)) return;
+    collectText(child, `${path}.${key}`, parts);
+  });
+}
+
+function localeCorpus(content: unknown): string {
+  if (!content || typeof content !== 'object') return '';
+
+  const record = content as Record<string, unknown>;
+  const parts: string[] = [];
+  collectText(record.title, 'title', parts);
+  collectText(record.description, 'description', parts);
+  collectText(record.faqTitle, 'faqTitle', parts);
+  collectText(record.faq, 'faq', parts);
+  collectText(record.seo, 'seo', parts);
+  collectText(record.schemas, 'schemas', parts);
+  return parts.join(' ');
+}
+
+function tokenCounts(text: string): Map<string, number> {
+  const counts = new Map<string, number>();
+  for (const token of text.match(/[\p{L}\p{N}]+/gu) ?? []) {
+    counts.set(token, (counts.get(token) ?? 0) + 1);
+  }
+  return counts;
+}
+
+function copySimilarity(left: string, right: string): number {
+  const leftCounts = tokenCounts(left);
+  const rightCounts = tokenCounts(right);
+  const leftTotal = [...leftCounts.values()].reduce((sum, count) => sum + count, 0);
+  const rightTotal = [...rightCounts.values()].reduce((sum, count) => sum + count, 0);
+  if (leftTotal === 0 || rightTotal === 0) return 0;
+
+  let shared = 0;
+  for (const [token, count] of leftCounts) {
+    shared += Math.min(count, rightCounts.get(token) ?? 0);
+  }
+  return (2 * shared) / (leftTotal + rightTotal);
+}
+
+describe('Locales must not copy another locale wholesale', () => {
+  ALL_ENTRIES.forEach((entry) => {
+    it(`${entry.id} is not at least ${COPY_THRESHOLD * 100}% identical to another locale`, async () => {
+      const corpora = new Map<string, string>();
+
+      for (const [locale, loader] of Object.entries(entry.i18n)) {
+        if (!loader) continue;
+        corpora.set(locale, localeCorpus(await loader()));
+      }
+
+      const locales = [...corpora.keys()];
+      const violations: string[] = [];
+
+      for (let leftIndex = 0; leftIndex < locales.length; leftIndex += 1) {
+        for (let rightIndex = leftIndex + 1; rightIndex < locales.length; rightIndex += 1) {
+          const left = locales[leftIndex];
+          const right = locales[rightIndex];
+          const similarity = copySimilarity(corpora.get(left) ?? '', corpora.get(right) ?? '');
+
+          if (similarity >= COPY_THRESHOLD) {
+            violations.push(`${left} ↔ ${right}: ${(similarity * 100).toFixed(1)}%`);
+          }
+        }
+      }
+
+      expect(violations, `Locale copy threshold exceeded in ${entry.id}`).toEqual([]);
+    });
+  });
+});
+
+