Press n or j to go to the next uncovered block, b, p or k for the previous block.
| 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 | 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 48x 48x 48x 48x 48x 140x 48x 192x 192x 50x 144x 8x 48x 48x 78x 78x 48x 48x 39x 39x 39x 254x 254x 254x 48x 48x 48x 17x 17x 16x 16x 16x 83x 48x 48x 48x 48x 48x 47x 47x 47x 47x 47x 46x 46x 48x 48x 48x 48x 48x 3x 48x 48x 4x 48x 48x 48x 5x 48x 48x 48x 1x 48x 48x 48x 3x 45x 1x 48x 48x 48x 48x 1x 48x 48x 2x 48x 48x 13x 48x 30x 30x 30x 30x 30x 30x 30x 30x 30x 30x 30x 30x 30x 240x 30x 30x 30x 30x 30x 30x 30x 30x 30x 30x 30x 30x 6x 24x 3x 1x 2x 24x 9x 4x 9x 4x 24x 4x 24x | // SPDX-FileCopyrightText: 2024-2026 Hack23 AB
// SPDX-License-Identifier: Apache-2.0
/**
* @module Utils/ContentValidator
* @description Post-generation content quality validation for news articles.
*
* Validates generated article HTML for minimum word counts, placeholder text,
* required structural HTML elements, language consistency, meta tag alignment,
* read-time accuracy, and keyword localization. Produces a structured quality report.
*/
import { ArticleCategory } from '../types/index.js';
import { stripScriptBlocks } from './html-sanitize.js';
// ─── Types ────────────────────────────────────────────────────────────────────
/** Quality metrics collected during content validation */
export interface ContentValidationMetrics {
/** Approximate word count of plain text extracted from HTML */
wordCount: number;
/** Whether all required structural HTML elements are present */
htmlValid: boolean;
/** Whether placeholder/template markers were detected in the output */
hasPlaceholders: boolean;
/** Computed read-time based on actual word count (words / 250, min 1) */
computedReadTime: number;
/** Claimed read-time extracted from the article (0 if not found) */
claimedReadTime: number;
/** Whether the html lang attribute matches the expected language */
langAttributeValid: boolean;
/** Whether the dir attribute is correctly set for RTL languages */
dirAttributeValid: boolean;
/** Whether meta tags (title, og:title, twitter:title) are synchronized */
metaTagsSynced: boolean;
/** Whether keywords contain at least some localized terms for non-English articles */
keywordsLocalized: boolean;
}
/** Result returned by {@link validateArticleContent} */
export interface ContentValidationResult {
/** true when no errors were found (warnings are allowed) */
valid: boolean;
/** Non-fatal quality notices */
warnings: string[];
/** Fatal quality failures */
errors: string[];
/** Collected numeric and boolean quality metrics */
metrics: ContentValidationMetrics;
}
/** Quality metrics collected during translation completeness validation */
export interface TranslationValidationMetrics {
/** Ratio of ASCII printable characters to total text (0–1). High values in CJK articles suggest untranslated content. */
asciiRatio: number;
/** Ratio of CJK characters to total text (0–1). Low values in ja/ko/zh articles suggest untranslated content. */
cjkCharRatio: number;
/** Whether the `dir="rtl"` attribute is present on the `<html>` element */
hasRtlDir: boolean;
/** Whether Unicode bidi control characters or `‎`/`‏` markers are present */
hasBidiMarkers: boolean;
/** English phrases found in non-English articles that likely should have been translated */
untranslatedPhrases: readonly string[];
}
/**
* Result returned by {@link validateTranslationCompleteness}.
*
* Translation validation is non-blocking — it produces warnings only.
*/
export interface TranslationValidationResult {
/** true when no translation quality issues were detected */
valid: boolean;
/** Non-fatal translation quality notices */
warnings: string[];
/** Collected translation quality metrics */
metrics: TranslationValidationMetrics;
}
// ─── Constants ────────────────────────────────────────────────────────────────
/** Minimum word counts (plain text) required per article category */
const MIN_WORD_COUNTS: Readonly<Record<string, number>> = {
[ArticleCategory.WEEK_AHEAD]: 500,
[ArticleCategory.MONTH_AHEAD]: 500,
[ArticleCategory.BREAKING_NEWS]: 300,
[ArticleCategory.COMMITTEE_REPORTS]: 400,
[ArticleCategory.PROPOSITIONS]: 300,
[ArticleCategory.MOTIONS]: 300,
[ArticleCategory.WEEK_IN_REVIEW]: 300,
[ArticleCategory.MONTH_IN_REVIEW]: 600,
} as const;
/** Default minimum word count when category is not listed */
const DEFAULT_MIN_WORDS = 300;
/** Words per minute for read-time calculation */
const WORDS_PER_MINUTE = 250;
/** Maximum read-time deviation (minutes) before a warning is triggered */
const READ_TIME_TOLERANCE = 2;
/** RTL language codes requiring dir="rtl" */
const RTL_LANGUAGES: ReadonlySet<string> = new Set(['ar', 'he']);
/** Patterns that indicate un-replaced template markers */
const PLACEHOLDER_PATTERNS: ReadonlyArray<RegExp> = [
/\{\{[^}]+\}\}/u,
/\[TODO[^\]]*\]/iu,
/\bPLACEHOLDER\b/u,
] as const;
/** Required structural HTML elements that every article must contain */
const REQUIRED_HTML_ELEMENTS: ReadonlyArray<{
selector: string | readonly string[];
label: string;
}> = [
{
selector: ['class="site-header__langs"', 'class="language-switcher"'],
label: 'language switcher nav',
},
{ selector: 'class="article-top-nav"', label: 'article-top-nav (back button)' },
{ selector: 'class="site-header"', label: 'site-header' },
{ selector: '<main id="main"', label: 'main content wrapper' },
] as const;
/**
* Localized keyword indicators per language.
* If a non-English article's keywords contain at least one of these
* language-specific terms, keyword localization is considered acceptable.
* This catches the common issue of English-only keywords in translated articles.
*/
const LOCALIZED_KEYWORD_INDICATORS: Readonly<Record<string, ReadonlyArray<string>>> = {
sv: ['parlamentet', 'lagstiftning', 'EU', 'europeisk', 'utskott', 'omröstning', 'förordning'],
da: ['parlamentet', 'lovgivning', 'udvalg', 'afstemning', 'forordning', 'europæisk'],
no: ['parlamentet', 'lovgivning', 'komité', 'avstemning', 'forordning', 'europeisk'],
fi: ['parlamentti', 'lainsäädäntö', 'valiokunta', 'äänestys', 'asetus', 'eurooppalainen'],
de: ['Parlament', 'Gesetzgebung', 'Ausschuss', 'Abstimmung', 'Verordnung', 'europäisch'],
fr: ['parlement', 'législation', 'commission', 'vote', 'règlement', 'européen'],
es: ['parlamento', 'legislación', 'comisión', 'votación', 'reglamento', 'europeo'],
nl: ['parlement', 'wetgeving', 'commissie', 'stemming', 'verordening', 'Europees'],
ar: ['البرلمان', 'التشريع', 'اللجنة', 'التصويت', 'الأوروبي', 'القرار'],
he: ['הפרלמנט', 'חקיקה', 'ועדה', 'הצבעה', 'האירופי', 'תקנה'],
ja: ['議会', '立法', '委員会', '投票', '規則', '欧州'],
ko: ['의회', '입법', '위원회', '투표', '규정', '유럽'],
zh: ['议会', '立法', '委员会', '投票', '条例', '欧洲'],
} as const;
/** CJK language codes requiring ideographic character density checks */
const CJK_LANGUAGES: ReadonlySet<string> = new Set(['ja', 'ko', 'zh']);
/**
* ASCII ratio threshold above which CJK articles are considered likely untranslated.
* Real CJK content typically has <50 % ASCII (mostly HTML entities and punctuation).
*/
const CJK_ASCII_RATIO_THRESHOLD = 0.85;
/**
* Minimum CJK character ratio expected in properly translated CJK articles.
* Below this value the content is likely still English.
*/
const CJK_CHAR_RATIO_THRESHOLD = 0.05;
/**
* Common English phrases that should not appear (un-translated) in non-English articles.
* These are generic header/label/placeholder phrases that a proper translation would replace.
*/
const ENGLISH_PLACEHOLDER_PHRASES: ReadonlyArray<string> = [
'European Parliament',
'Read more',
'Table of Contents',
'Key Takeaways',
'Executive Summary',
'Click here',
'Learn more',
'Subscribe',
] as const;
// ─── Helpers ──────────────────────────────────────────────────────────────────
// stripScriptBlocks is imported from html-sanitize.ts
/**
* Extract plain text from the `<main>` element of an article and count words.
*
* Restricts counting to the main content area so that JSON-LD scripts,
* navigation, and header/footer boilerplate do not inflate the word count.
* Falls back to the full document when no `<main>` element is found.
*
* @param html - Raw HTML string
* @returns Approximate word count of the main content area
*/
function countWordsInHtml(html: string): number {
const mainMatch = /<main[^>]*>([\s\S]*?)<\/main>/u.exec(html);
const source = mainMatch?.[1] ?? html;
const plainText = stripScriptBlocks(source)
.replace(/<[^>]+>/gu, ' ')
.replace(/\s+/gu, ' ')
.trim();
Iif (!plainText) return 0;
return plainText.split(' ').length;
}
/**
* Detect whether any un-replaced template placeholder patterns remain in the content.
*
* @param html - HTML string to inspect
* @returns true if at least one placeholder pattern is found
*/
function detectPlaceholders(html: string): boolean {
return PLACEHOLDER_PATTERNS.some((pattern) => pattern.test(html));
}
/**
* Check that all required structural HTML elements are present.
*
* @param html - HTML string to inspect
* @returns Array of labels for missing elements (empty when all present)
*/
function findMissingElements(html: string): string[] {
return REQUIRED_HTML_ELEMENTS.filter((el) => {
const sel = el.selector;
if (Array.isArray(sel)) {
return !sel.some((s) => html.includes(s));
}
return !html.includes(sel as string);
}).map((el) => el.label);
}
/**
* Extract the value of the `lang` attribute from the `<html>` tag.
*
* @param html - HTML string to inspect
* @returns The lang value or empty string if not found
*/
function extractLangAttribute(html: string): string {
const match = /<html[^>]*\slang="([^"]+)"/iu.exec(html);
return match?.[1] ?? '';
}
/**
* Extract the value of the `dir` attribute from the `<html>` tag.
*
* @param html - HTML string to inspect
* @returns The dir value or empty string if not found
*/
function extractDirAttribute(html: string): string {
const match = /<html[^>]*\sdir="([^"]+)"/iu.exec(html);
return match?.[1] ?? '';
}
/**
* Extract the claimed read-time from the article.
* Looks for patterns like "5 min read", "3分で読了", "5分钟阅读", etc.
*
* @param html - HTML string to inspect
* @returns Claimed read time in minutes or 0 if not found
*/
function extractClaimedReadTime(html: string): number {
// Look for read-time inside the article meta section
const readTimeMatch =
/class="article-read-time"[^>]*>([^<]*)/iu.exec(html) ??
/article-read-time[^>]*>([^<]*)/iu.exec(html);
if (!readTimeMatch?.[1]) return 0;
const text = readTimeMatch[1].trim();
// Extract the numeric portion — handles "5 min read", "5分で読了", "٥ دقائق قراءة"
const numMatch = /(\d+)/u.exec(text);
return numMatch?.[1] ? parseInt(numMatch[1], 10) : 0;
}
/**
* Extract meta tag content by name or property.
*
* @param html - HTML string
* @param attr - Attribute name ('name' or 'property')
* @param value - Attribute value to match (e.g. 'og:title')
* @returns The content attribute value or empty string
*/
function extractMetaContent(html: string, attr: string, value: string): string {
// Handle both orderings: <meta name="x" content="y"> and <meta content="y" name="x">
const pattern1 = new RegExp(`<meta\\s+${attr}="${value}"\\s+content="([^"]*)"`, 'iu');
const pattern2 = new RegExp(`<meta\\s+content="([^"]*)"\\s+${attr}="${value}"`, 'iu');
return pattern1.exec(html)?.[1] ?? pattern2.exec(html)?.[1] ?? '';
}
/**
* Extract the page title from the `<title>` tag.
*
* @param html - HTML string
* @returns Title text or empty string
*/
function extractTitle(html: string): string {
const match = /<title>([^<]*)<\/title>/iu.exec(html);
return match?.[1]?.trim() ?? '';
}
/**
* Check whether keywords contain language-specific localized terms.
* For English articles, always returns true.
* For non-English articles, checks if at least one keyword matches
* a known localized indicator for that language.
*
* @param html - HTML string to extract keywords from
* @param language - Expected language code
* @returns true if keywords appear localized for the given language
*/
function checkKeywordLocalization(html: string, language: string): boolean {
if (language === 'en') return true;
const keywordsMeta = extractMetaContent(html, 'name', 'keywords');
if (!keywordsMeta) return true; // No keywords = no localization issue
const indicators = LOCALIZED_KEYWORD_INDICATORS[language];
Iif (!indicators) return true; // Unknown language = skip check
const keywordsLower = keywordsMeta.toLowerCase();
return indicators.some((indicator) => keywordsLower.includes(indicator.toLowerCase()));
}
/**
* Check whether meta tags (title, OG, Twitter) are synchronized.
*
* @param html - HTML string to inspect
* @returns true if the core meta tags are reasonably aligned
*/
function checkMetaTagSync(html: string): boolean {
const pageTitle = extractTitle(html);
const ogTitle = extractMetaContent(html, 'property', 'og:title');
const twitterTitle = extractMetaContent(html, 'name', 'twitter:title');
// If OG or Twitter title is present, it should match the page title
// (stripping the " | EU Parliament Monitor" suffix from page title)
const coreTitle = pageTitle.replace(/\s*\|\s*EU Parliament Monitor$/iu, '').trim();
if (ogTitle && ogTitle !== coreTitle) return false;
Iif (twitterTitle && twitterTitle !== coreTitle) return false;
// Also check description alignment
const description = extractMetaContent(html, 'name', 'description');
const ogDescription = extractMetaContent(html, 'property', 'og:description');
const twitterDescription = extractMetaContent(html, 'name', 'twitter:description');
if (ogDescription && description && ogDescription !== description) return false;
Iif (twitterDescription && description && twitterDescription !== description) return false;
return true;
}
// ─── Public API ───────────────────────────────────────────────────────────────
/**
* Validate the quality of a generated article.
*
* Checks performed:
* - Minimum word count threshold for the given article type
* - Presence of un-replaced placeholder/template markers
* - Existence of required structural HTML elements
* - Language attribute consistency (`lang` and `dir`)
* - Read-time accuracy (computed vs claimed)
* - Meta tag synchronization (title/OG/Twitter alignment)
* - Keyword localization for non-English articles
*
* @param html - Complete HTML string of the generated article
* @param language - Language code of the article (e.g. `"en"`, `"de"`, `"ar"`)
* @param articleType - Article category string (e.g. `"week-ahead"`)
* @returns Structured validation result with errors, warnings and metrics
*/
export function validateArticleContent(
html: string,
language: string,
articleType: string
): ContentValidationResult {
const warnings: string[] = [];
const errors: string[] = [];
// Word count check
const wordCount = countWordsInHtml(html);
const minWords =
MIN_WORD_COUNTS[articleType] !== undefined ? MIN_WORD_COUNTS[articleType] : DEFAULT_MIN_WORDS;
if (wordCount < minWords) {
warnings.push(
`Content too short: ${wordCount} words (minimum ${minWords} for "${articleType}")`
);
}
// Placeholder detection
const hasPlaceholders = detectPlaceholders(html);
if (hasPlaceholders) {
errors.push('Un-replaced template placeholder(s) detected in generated content');
}
// Required HTML elements
const missingElements = findMissingElements(html);
const htmlValid = missingElements.length === 0;
if (!htmlValid) {
errors.push(`Missing required HTML element(s): ${missingElements.join(', ')}`);
}
// Read-time accuracy
const computedReadTime = Math.max(1, Math.ceil(wordCount / WORDS_PER_MINUTE));
const claimedReadTime = extractClaimedReadTime(html);
if (claimedReadTime > 0 && Math.abs(computedReadTime - claimedReadTime) > READ_TIME_TOLERANCE) {
warnings.push(
`Read-time mismatch: claimed ${claimedReadTime} min but content is ~${computedReadTime} min (${wordCount} words)`
);
}
// Language attribute check
const langAttr = extractLangAttribute(html);
const langAttributeValid = langAttr === language;
if (!langAttributeValid && langAttr) {
warnings.push(
`Language attribute mismatch: <html lang="${langAttr}"> but expected "${language}"`
);
} else if (!langAttr) {
warnings.push('Missing lang attribute on <html> element');
}
// Dir attribute check for RTL languages
const dirAttr = extractDirAttribute(html);
const isRtl = RTL_LANGUAGES.has(language);
const dirAttributeValid = isRtl ? dirAttr === 'rtl' : dirAttr !== 'rtl';
if (isRtl && dirAttr !== 'rtl') {
warnings.push(
`RTL language "${language}" should have dir="rtl" but found dir="${dirAttr || '(none)'}"`
);
}
// Meta tag synchronization
const metaTagsSynced = checkMetaTagSync(html);
if (!metaTagsSynced) {
warnings.push(
'Meta tag mismatch: title, og:title, twitter:title, or descriptions are not synchronized'
);
}
// Keyword localization
const keywordsLocalized = checkKeywordLocalization(html, language);
if (!keywordsLocalized) {
warnings.push(
`Keywords for "${language}" article appear to be entirely in English — consider localizing`
);
}
return {
valid: errors.length === 0,
warnings,
errors,
metrics: {
wordCount,
htmlValid,
hasPlaceholders,
computedReadTime,
claimedReadTime,
langAttributeValid,
dirAttributeValid,
metaTagsSynced,
keywordsLocalized,
},
};
}
// ─── Translation validation helpers ───────────────────────────────────────────
/**
* Extract plain body text from `<main>` for character-class analysis.
* Strips all HTML tags and normalises whitespace.
*
* @param html - Raw HTML string
* @returns Plain text content from the main element
*/
function extractMainPlainText(html: string): string {
const mainMatch = /<main[^>]*>([\s\S]*?)<\/main>/u.exec(html);
const source = mainMatch?.[1] ?? html;
return stripScriptBlocks(source)
.replace(/<[^>]+>/gu, ' ')
.replace(/&(?:[a-z][a-z0-9]+|#\d+|#x[0-9a-f]+);/giu, ' ')
.replace(/\s+/gu, ' ')
.trim();
}
/**
* Compute the ratio of ASCII printable characters (0x20–0x7E) in a string.
*
* @param text - Plain text string
* @returns Ratio from 0 to 1 (1 = all ASCII)
*/
function computeAsciiRatio(text: string): number {
Iif (text.length === 0) return 0;
const asciiCount = (text.match(/[\x20-\x7E]/gu) ?? []).length;
// Use Array.from to correctly count Unicode characters (handles surrogate pairs)
const charCount = Array.from(text).length;
return asciiCount / charCount;
}
/**
* Compute the ratio of CJK Unified Ideograph characters in a string.
* Covers CJK Unified Ideographs (U+4E00–U+9FFF), Extension A (U+3400–U+4DBF),
* Hiragana, Katakana, and Hangul Syllables.
*
* @param text - Plain text string
* @returns Ratio from 0 to 1
*/
function computeCjkCharRatio(text: string): number {
Iif (text.length === 0) return 0;
// CJK Unified Ideographs + Extension A, Hiragana, Katakana, Hangul Syllables
// Note: Extension B (U+20000–U+2A6DF) is omitted as it triggers unsafe-regex lint
// and is extremely rare in EU Parliament content.
const cjkPattern = /[\u4E00-\u9FFF\u3400-\u4DBF\u3040-\u309F\u30A0-\u30FF\uAC00-\uD7AF]/gu;
const matches = text.match(cjkPattern);
// Use Array.from to correctly count Unicode characters (handles surrogate pairs)
const charCount = Array.from(text).length;
return (matches?.length ?? 0) / charCount;
}
/**
* Detect common English phrases that should have been translated in non-English articles.
*
* @param text - Plain text content
* @returns Array of detected untranslated English phrases
*/
function findUntranslatedPhrases(text: string): string[] {
const lowerText = text.toLowerCase();
return ENGLISH_PLACEHOLDER_PHRASES.filter((phrase) => lowerText.includes(phrase.toLowerCase()));
}
/**
* Check whether Unicode bidirectional control characters or HTML bidi markers are present.
*
* @param html - Raw HTML string
* @returns true if bidi markers or control characters are found
*/
function detectBidiMarkers(html: string): boolean {
// Unicode bidi control characters: LRM, RLM, LRE, RLE, PDF, LRO, RLO, LRI, RLI, FSI, PDI
const bidiControlPattern = /[\u200E\u200F\u202A-\u202E\u2066-\u2069]/u;
// HTML entities: ‎ ‏
const bidiEntityPattern = /&(?:lrm|rlm);/iu;
return bidiControlPattern.test(html) || bidiEntityPattern.test(html);
}
/**
* Validate translation completeness and cultural adaptation for a generated article.
*
* Checks performed:
* - **RTL languages (ar, he)**: Verify `dir="rtl"` on `<html>`, detect bidi control markers
* - **CJK languages (ja, ko, zh)**: Check that content has sufficient CJK character density
* (high ASCII ratio suggests the article was not actually translated)
* - **All non-English**: Detect common English phrases that should have been translated
*
* This function is purely analytical — no AI calls. It produces warnings only
* and never blocks article generation.
*
* @param html - Complete HTML string of the generated article
* @param lang - Language code of the article (e.g. `"ar"`, `"ja"`, `"en"`)
* @returns Structured translation validation result with warnings and metrics
*/
export function validateTranslationCompleteness(
html: string,
lang: string
): TranslationValidationResult {
const warnings: string[] = [];
const plainText = extractMainPlainText(html);
const asciiRatio = computeAsciiRatio(plainText);
const cjkCharRatio = computeCjkCharRatio(plainText);
const htmlDir = extractDirAttribute(html);
const hasRtlDir = htmlDir === 'rtl';
const hasBidiMarkers = detectBidiMarkers(html);
const untranslatedPhrases = findUntranslatedPhrases(plainText);
// Skip validation warnings for English — it is the source language,
// but still compute and return real metrics for telemetry/reporting.
if (lang === 'en') {
return {
valid: true,
warnings: [],
metrics: {
asciiRatio,
cjkCharRatio,
hasRtlDir,
hasBidiMarkers,
untranslatedPhrases,
},
};
}
// ── RTL validation ──────────────────────────────────────────────────────
if (RTL_LANGUAGES.has(lang) && !hasRtlDir) {
if (!htmlDir) {
warnings.push(
`Translation quality: RTL language "${lang}" missing dir="rtl" on <html> element`
);
} else {
warnings.push(
`Translation quality: RTL language "${lang}" expected dir="rtl" on <html> element but found dir="${htmlDir}"`
);
}
}
// ── CJK density check ──────────────────────────────────────────────────
if (CJK_LANGUAGES.has(lang) && plainText.length > 0) {
if (asciiRatio > CJK_ASCII_RATIO_THRESHOLD) {
warnings.push(
`Translation quality: ${lang.toUpperCase()} article has ${(asciiRatio * 100).toFixed(0)}% ASCII characters — content may be untranslated`
);
}
if (cjkCharRatio < CJK_CHAR_RATIO_THRESHOLD) {
warnings.push(
`Translation quality: ${lang.toUpperCase()} article has only ${(cjkCharRatio * 100).toFixed(1)}% CJK characters — expected native script content`
);
}
}
// ── Untranslated English phrase detection ───────────────────────────────
if (untranslatedPhrases.length > 0) {
warnings.push(
`Translation quality: found ${untranslatedPhrases.length} likely untranslated English phrase(s): ${untranslatedPhrases.slice(0, 3).join(', ')}`
);
}
return {
valid: warnings.length === 0,
warnings,
metrics: {
asciiRatio,
cjkCharRatio,
hasRtlDir,
hasBidiMarkers,
untranslatedPhrases,
},
};
}
|