riksdagsmonitor 0.8.14 → 0.8.15

package/dist/scripts/extract-news-metadata.js

@@ -0,0 +1,107 @@
+/**
+ * Extract News Article Metadata to JSON Database
+ *
+ * Parses all news article HTML files and extracts Schema.org JSON-LD
+ * metadata into a single data/news-articles.json file.
+ *
+ * Usage: node --experimental-strip-types scripts/extract-news-metadata.ts
+ *
+ * @module scripts/extract-news-metadata
+ */
+import { readFileSync, writeFileSync, readdirSync, mkdirSync } from 'fs';
+import { join, dirname, relative, sep } from 'path';
+import { fileURLToPath } from 'url';
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+const ROOT = join(__dirname, '..');
+function collectNewsFiles(dir) {
+    const result = [];
+    for (const entry of readdirSync(dir, { withFileTypes: true })) {
+        if (entry.isDirectory()) {
+            result.push(...collectNewsFiles(join(dir, entry.name)));
+        }
+        else if (entry.name.endsWith('.html') && !entry.name.startsWith('index')) {
+            result.push(join(dir, entry.name));
+        }
+    }
+    return result;
+}
+function extractMetadata() {
+    const newsDir = join(ROOT, 'news');
+    const allFilePaths = collectNewsFiles(newsDir);
+    const files = allFilePaths.map((fp) => relative(newsDir, fp).split(sep).join('/'));
+    const articles = [];
+    for (const file of files) {
+        const content = readFileSync(join(newsDir, file), 'utf-8');
+        // Extract JSON-LD blocks
+        const jsonLdRegex = /<script type="application\/ld\+json">([\s\S]*?)<\/script>/g;
+        let match;
+        let articleData = null;
+        while ((match = jsonLdRegex.exec(content)) !== null) {
+            try {
+                const parsed = JSON.parse(match[1]);
+                if (parsed['@type'] === 'NewsArticle') {
+                    articleData = parsed;
+                    break;
+                }
+            }
+            catch {
+                // Skip malformed JSON-LD
+            }
+        }
+        if (!articleData) {
+            console.warn(`WARN: No NewsArticle JSON-LD in ${file}`);
+            continue;
+        }
+        // Extract Open Graph image
+        const ogImageMatch = content.match(/property="og:image" content="([^"]+)"/);
+        const ogImage = ogImageMatch?.[1] ?? '';
+        // Extract language from filename
+        const langMatch = file.match(/-([a-z]{2})\.html$/);
+        const lang = langMatch?.[1] ?? 'en';
+        // Extract slug (filename without language suffix)
+        const slug = file.replace(/-[a-z]{2}\.html$/, '');
+        articles.push({
+            slug,
+            file,
+            lang,
+            headline: articleData.headline ?? '',
+            description: articleData.description ?? '',
+            datePublished: articleData.datePublished ?? '',
+            dateModified: articleData.dateModified ?? '',
+            articleSection: articleData.articleSection ?? '',
+            wordCount: articleData.wordCount || (() => {
+                const stripped = content.replace(/<[^>]+>/g, ' ');
+                return stripped.split(/\s+/).filter((w) => w.length > 0).length;
+            })(),
+            inLanguage: articleData.inLanguage ?? lang,
+            keywords: articleData.keywords ?? '',
+            image: ogImage,
+            url: articleData.mainEntityOfPage?.['@id'] ?? '',
+        });
+    }
+    // Sort by date descending, then by language
+    articles.sort((a, b) => {
+        const dateCompare = b.datePublished.localeCompare(a.datePublished);
+        if (dateCompare !== 0)
+            return dateCompare;
+        return a.lang.localeCompare(b.lang);
+    });
+    const db = {
+        version: '1.0.0',
+        generatedAt: new Date().toISOString(),
+        totalArticles: articles.length,
+        uniqueSlugs: [...new Set(articles.map((a) => a.slug))].length,
+        languages: [...new Set(articles.map((a) => a.lang))].sort(),
+        articles,
+    };
+    mkdirSync(join(ROOT, 'data'), { recursive: true });
+    writeFileSync(join(ROOT, 'data', 'news-articles.json'), JSON.stringify(db, null, 2));
+    console.log('Generated data/news-articles.json:');
+    console.log('  Total articles:', db.totalArticles);
+    console.log('  Unique slugs:', db.uniqueSlugs);
+    console.log('  Languages:', db.languages.join(', '));
+    console.log('  File size:', (JSON.stringify(db).length / 1024).toFixed(1), 'KB');
+}
+extractMetadata();
+//# sourceMappingURL=extract-news-metadata.js.map

package/dist/scripts/extract-news-metadata.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"extract-news-metadata.js","sourceRoot":"","sources":["../../scripts/extract-news-metadata.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,YAAY,EAAE,aAAa,EAAE,WAAW,EAAE,SAAS,EAAE,MAAM,IAAI,CAAC;AACzE,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,GAAG,EAAE,MAAM,MAAM,CAAC;AACpD,OAAO,EAAE,aAAa,EAAE,MAAM,KAAK,CAAC;AAEpC,MAAM,UAAU,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAClD,MAAM,SAAS,GAAG,OAAO,CAAC,UAAU,CAAC,CAAC;AACtC,MAAM,IAAI,GAAG,IAAI,CAAC,SAAS,EAAE,IAAI,CAAC,CAAC;AAyCnC,SAAS,gBAAgB,CAAC,GAAW;IACnC,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,GAAG,EAAE,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC;QAC9D,IAAI,KAAK,CAAC,WAAW,EAAE,EAAE,CAAC;YACxB,MAAM,CAAC,IAAI,CAAC,GAAG,gBAAgB,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAC1D,CAAC;aAAM,IAAI,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;YAC3E,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;QACrC,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,eAAe;IACtB,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;IACnC,MAAM,YAAY,GAAG,gBAAgB,CAAC,OAAO,CAAC,CAAC;IAC/C,MAAM,KAAK,GAAG,YAAY,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,QAAQ,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;IAEnF,MAAM,QAAQ,GAAsB,EAAE,CAAC;IAEvC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,OAAO,CAAC,CAAC;QAE3D,yBAAyB;QACzB,MAAM,WAAW,GAAG,4DAA4D,CAAC;QACjF,IAAI,KAA6B,CAAC;QAClC,IAAI,WAAW,GAAyB,IAAI,CAAC;QAE7C,OAAO,CAAC,KAAK,GAAG,WAAW,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YACpD,IAAI,CAAC;gBACH,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAE,CAAkB,CAAC;gBACtD,IAAI,MAAM,CAAC,OAAO,CAAC,KAAK,aAAa,EAAE,CAAC;oBACtC,WAAW,GAAG,MAAM,CAAC;oBACrB,MAAM;gBACR,CAAC;YACH,CAAC;YAAC,MAAM,CAAC;gBACP,yBAAyB;YAC3B,CAAC;QACH,CAAC;QAED,IAAI,CAAC,WAAW,EAAE,CAAC;YACjB,OAAO,CAAC,IAAI,CAAC,mCAAmC,IAAI,EAAE,CAAC,CAAC;YACxD,SAAS;QACX,CAAC;QAED,2BAA2B;QAC3B,MAAM,YAAY,GAAG,OAAO,CAAC,KAAK,CAAC,uCAAuC,CAAC,CAAC;QAC5E,MAAM,OAAO,GAAW,YAAY,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAEhD,iCAAiC;QACjC,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,oBAAoB,CAAC,CAAC;QACnD,MAAM,IAAI,GAAW,SAAS,EAAE,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC;QAE5C,kDAAkD;QAClD,MAAM,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,kBAAkB,EAAE,EAAE,CAAC,CAAC;QAElD,QAAQ,CAAC,IAAI,CAAC;YACZ,IAAI;YACJ,IAAI;YACJ,IAAI;YACJ,QAAQ,EAAE,WAAW,CAAC,QAAQ,IAAI,EAAE;YACpC,WAAW,EAAE,WAAW,CAAC,WAAW,IAAI,EAAE;YAC1C,aAAa,EAAE,WAAW,CAAC,aAAa,IAAI,EAAE;YAC9C,YAAY,EAAE,WAAW,CAAC,YAAY,IAAI,EAAE;YAC5C,cAAc,EAAE,WAAW,CAAC,cAAc,IAAI,EAAE;YAChD,SAAS,EAAE,WAAW,CAAC,SAAS,IAAI,CAAC,GAAG,EAAE;gBACxC,MAAM,QAAQ,GAAG,OAAO,CAAC,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC,CAAC;gBAClD,OAAO,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAS,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC;YAC1E,CAAC,CAAC,EAAE;YACJ,UAAU,EAAE,WAAW,CAAC,UAAU,IAAI,IAAc;YACpD,QAAQ,EAAE,WAAW,CAAC,QAAQ,IAAI,EAAE;YACpC,KAAK,EAAE,OAAiB;YACxB,GAAG,EAAE,WAAW,CAAC,gBAAgB,EAAE,CAAC,KAAK,CAAC,IAAI,EAAE;SACjD,CAAC,CAAC;IACL,CAAC;IAED,4CAA4C;IAC5C,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;QACrB,MAAM,WAAW,GAAG,CAAC,CAAC,aAAa,CAAC,aAAa,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC;QACnE,IAAI,WAAW,KAAK,CAAC;YAAE,OAAO,WAAW,CAAC;QAC1C,OAAO,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IACtC,CAAC,CAAC,CAAC;IAEH,MAAM,EAAE,GAAiB;QACvB,OAAO,EAAE,OAAO;QAChB,WAAW,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QACrC,aAAa,EAAE,QAAQ,CAAC,MAAM;QAC9B,WAAW,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM;QAC7D,SAAS,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE;QAC3D,QAAQ;KACT,CAAC;IAEF,SAAS,CAAC,IAAI,CAAC,IAAI,EAAE,MAAM,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IACnD,aAAa,CAAC,IAAI,CAAC,IAAI,EAAE,MAAM,EAAE,oBAAoB,CAAC,EAAE,IAAI,CAAC,SAAS,CAAC,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;IAErF,OAAO,CAAC,GAAG,CAAC,oCAAoC,CAAC,CAAC;IAClD,OAAO,CAAC,GAAG,CAAC,mBAAmB,EAAE,EAAE,CAAC,aAAa,CAAC,CAAC;IACnD,OAAO,CAAC,GAAG,CAAC,iBAAiB,EAAE,EAAE,CAAC,WAAW,CAAC,CAAC;IAC/C,OAAO,CAAC,GAAG,CAAC,cAAc,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IACrD,OAAO,CAAC,GAAG,CAAC,cAAc,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC,MAAM,GAAG,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;AACnF,CAAC;AAED,eAAe,EAAE,CAAC"}

package/dist/scripts/extract-vocabulary.d.ts

@@ -0,0 +1,15 @@
+/**
+ * @module Intelligence/Terminology
+ * @category Intelligence Operations / Supporting Infrastructure
+ * @name Vocabulary Extraction - Political Terminology Pattern Analysis
+ *
+ * @description
+ * Advanced terminology extraction system analyzing translated news articles across
+ * all 14 supported languages to identify and catalog political terminology patterns.
+ *
+ * @author Hack23 AB (Linguistic Intelligence Team)
+ * @license Apache-2.0
+ * @version 2.5.0
+ */
+export {};
+//# sourceMappingURL=extract-vocabulary.d.ts.map

package/dist/scripts/extract-vocabulary.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"extract-vocabulary.d.ts","sourceRoot":"","sources":["../../scripts/extract-vocabulary.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG"}

package/dist/scripts/extract-vocabulary.js

@@ -0,0 +1,255 @@
+/**
+ * @module Intelligence/Terminology
+ * @category Intelligence Operations / Supporting Infrastructure
+ * @name Vocabulary Extraction - Political Terminology Pattern Analysis
+ *
+ * @description
+ * Advanced terminology extraction system analyzing translated news articles across
+ * all 14 supported languages to identify and catalog political terminology patterns.
+ *
+ * @author Hack23 AB (Linguistic Intelligence Team)
+ * @license Apache-2.0
+ * @version 2.5.0
+ */
+import { readFileSync, readdirSync } from 'fs';
+import { basename, join } from 'path';
+// ---------------------------------------------------------------------------
+// Types
+// ---------------------------------------------------------------------------
+/** Map of language codes to language names */
+const LANGUAGES = {
+    en: 'English', sv: 'Swedish', da: 'Danish', no: 'Norwegian', fi: 'Finnish',
+    de: 'German', fr: 'French', es: 'Spanish', nl: 'Dutch',
+    ar: 'Arabic', he: 'Hebrew', ja: 'Japanese', ko: 'Korean', zh: 'Chinese',
+};
+// Track skipped files for warning summary
+const skippedFiles = [];
+// ---------------------------------------------------------------------------
+// Term extraction
+// ---------------------------------------------------------------------------
+/**
+ * Extract political terms from HTML content using structure-based approach.
+ */
+function extractTerms(content, _lang) {
+    const terms = {};
+    // Extract titles (main political terminology)
+    const h3Pattern = /<h3>(.*?)<\/h3>/g;
+    const h3Matches = [];
+    let h3Match;
+    while ((h3Match = h3Pattern.exec(content)) !== null) {
+        const cleanText = h3Match[1].replace(/<[^>]+>/g, '').trim();
+        if (cleanText)
+            h3Matches.push(cleanText);
+    }
+    terms.titles = h3Matches.slice(0, 10);
+    // Extract "What to Watch" heading (any language) - structure-based
+    const h2Pattern = /<h2[^>]*>([^<]+)<\/h2>/g;
+    const h2Matches = [];
+    let h2Match;
+    while ((h2Match = h2Pattern.exec(content)) !== null) {
+        const text = h2Match[1].trim();
+        if (text.length > 5 && text.length < 100) {
+            h2Matches.push(text);
+        }
+    }
+    if (h2Matches.length > 0) {
+        terms.watchLabel = h2Matches[0];
+    }
+    // Extract structured labels from <strong>…:</strong> (language-agnostic)
+    const strongLabelPattern = /<strong>\s*([^:<]+?)\s*:\s*<\/strong>/g;
+    const strongLabels = [];
+    let strongMatch;
+    while ((strongMatch = strongLabelPattern.exec(content)) !== null) {
+        const label = strongMatch[1].trim();
+        if (label.length > 0 && label.length < 50) {
+            strongLabels.push(label);
+        }
+    }
+    if (strongLabels[0])
+        terms.committeeLabel = strongLabels[0];
+    if (strongLabels[1])
+        terms.documentLabel = strongLabels[1];
+    // Extract article type from title
+    const titleMatch = content.match(/<h1>([^<]+)<\/h1>/);
+    if (titleMatch)
+        terms.mainTitle = titleMatch[1].trim();
+    return terms;
+}
+// ---------------------------------------------------------------------------
+// Article analysis
+// ---------------------------------------------------------------------------
+/**
+ * Analyze all news articles.
+ */
+function analyzeArticles(directory = 'news', datePrefix = null) {
+    const results = {};
+    for (const lang of Object.keys(LANGUAGES)) {
+        results[lang] = {
+            language: LANGUAGES[lang],
+            code: lang,
+            samples: [],
+        };
+    }
+    try {
+        const files = readdirSync(directory).filter((f) => {
+            if (!f.endsWith('.html'))
+                return false;
+            if (datePrefix && !f.includes(datePrefix))
+                return false;
+            return true;
+        });
+        console.log(`\nScanning ${files.length} HTML files in ${directory}/`);
+        if (datePrefix) {
+            console.log(`Filtering by date prefix: "${datePrefix}"\n`);
+        }
+        for (const file of files) {
+            const match = file.match(/-([a-z]{2})\.html$/);
+            if (!match) {
+                skippedFiles.push({ file, reason: 'No language code in filename' });
+                continue;
+            }
+            const lang = match[1];
+            if (!results[lang]) {
+                skippedFiles.push({ file, reason: `Unknown language code: ${lang}` });
+                continue;
+            }
+            try {
+                const content = readFileSync(join(directory, file), 'utf-8');
+                const terms = extractTerms(content, lang);
+                // Determine article type
+                let articleType = 'general';
+                if (file.includes('committee'))
+                    articleType = 'committee-reports';
+                else if (file.includes('proposition'))
+                    articleType = 'propositions';
+                else if (file.includes('motion'))
+                    articleType = 'motions';
+                else if (file.includes('evening'))
+                    articleType = 'evening-analysis';
+                else if (file.includes('week-ahead'))
+                    articleType = 'week-ahead';
+                results[lang].samples.push({
+                    file: basename(file),
+                    type: articleType,
+                    terms,
+                });
+            }
+            catch (error) {
+                skippedFiles.push({ file, reason: `Read error: ${error.message}` });
+            }
+        }
+    }
+    catch (error) {
+        console.error(`Error reading directory: ${error.message}`);
+        process.exit(1);
+    }
+    return results;
+}
+// ---------------------------------------------------------------------------
+// Report
+// ---------------------------------------------------------------------------
+/**
+ * Generate vocabulary report.
+ */
+function generateReport(results) {
+    console.log('\n========================================');
+    console.log('Political Vocabulary Analysis Report');
+    console.log('========================================\n');
+    for (const [code, data] of Object.entries(results)) {
+        if (data.samples.length === 0)
+            continue;
+        console.log(`\n## ${data.language} (${code.toUpperCase()})`);
+        console.log(`Samples analyzed: ${data.samples.length}`);
+        // Collect unique labels
+        const watchLabels = new Set();
+        const committeeLabels = new Set();
+        const documentLabels = new Set();
+        const _mainTitles = new Set();
+        for (const sample of data.samples) {
+            if (sample.terms.watchLabel)
+                watchLabels.add(sample.terms.watchLabel);
+            if (sample.terms.committeeLabel)
+                committeeLabels.add(sample.terms.committeeLabel);
+            if (sample.terms.documentLabel)
+                documentLabels.add(sample.terms.documentLabel);
+            if (sample.terms.mainTitle)
+                _mainTitles.add(sample.terms.mainTitle);
+        }
+        if (watchLabels.size > 0)
+            console.log(`  "What to Watch": ${Array.from(watchLabels).join(', ')}`);
+        if (committeeLabels.size > 0)
+            console.log(`  "Committee": ${Array.from(committeeLabels).join(', ')}`);
+        if (documentLabels.size > 0)
+            console.log(`  "Document": ${Array.from(documentLabels).join(', ')}`);
+        // Show sample titles from any articles (prioritize committee reports)
+        const committeeReports = data.samples.filter((s) => s.type === 'committee-reports');
+        const sampleWithTitles = committeeReports.find((s) => s.terms.titles && s.terms.titles.length > 0) ||
+            data.samples.find((s) => s.terms.titles && s.terms.titles.length > 0);
+        if (sampleWithTitles && sampleWithTitles.terms.titles && sampleWithTitles.terms.titles.length > 0) {
+            console.log(`  Sample titles: ${sampleWithTitles.terms.titles.slice(0, 3).join(', ')}`);
+        }
+    }
+    // Warning summary
+    if (skippedFiles.length > 0) {
+        console.log('\n\n⚠️  WARNING: Skipped Files Summary');
+        console.log('=====================================');
+        console.log(`Total skipped: ${skippedFiles.length}\n`);
+        // Group by reason
+        const byReason = {};
+        for (const { file, reason } of skippedFiles) {
+            if (!byReason[reason])
+                byReason[reason] = [];
+            byReason[reason].push(file);
+        }
+        for (const [reason, files] of Object.entries(byReason)) {
+            console.log(`${reason}: ${files.length} file(s)`);
+            if (files.length <= 5) {
+                files.forEach((f) => console.log(`  - ${f}`));
+            }
+            else {
+                files.slice(0, 3).forEach((f) => console.log(`  - ${f}`));
+                console.log(`  ... and ${files.length - 3} more`);
+            }
+            console.log();
+        }
+    }
+    console.log('\n========================================');
+    console.log('Analysis complete!');
+    console.log('========================================\n');
+}
+// ---------------------------------------------------------------------------
+// CLI
+// ---------------------------------------------------------------------------
+const args = process.argv.slice(2);
+let datePrefix = null;
+let directory = 'news';
+for (let i = 0; i < args.length; i++) {
+    if (args[i] === '--date-prefix' && args[i + 1]) {
+        datePrefix = args[i + 1];
+        i++;
+    }
+    else if (args[i] === '--directory' && args[i + 1]) {
+        directory = args[i + 1];
+        i++;
+    }
+    else if (args[i] === '--help' || args[i] === '-h') {
+        console.log(`
+Usage: node scripts/extract-vocabulary.js [options]
+
+Options:
+  --date-prefix <prefix>   Filter files by date prefix (e.g., "2026-02-")
+  --directory <path>       Directory to scan (default: "news")
+  --help, -h              Show this help message
+
+Examples:
+  node scripts/extract-vocabulary.js
+  node scripts/extract-vocabulary.js --date-prefix 2026-02-
+  node scripts/extract-vocabulary.js --directory news --date-prefix 2026-03-
+`);
+        process.exit(0);
+    }
+}
+// Run analysis
+const results = analyzeArticles(directory, datePrefix);
+generateReport(results);
+//# sourceMappingURL=extract-vocabulary.js.map

package/dist/scripts/extract-vocabulary.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"extract-vocabulary.js","sourceRoot":"","sources":["../../scripts/extract-vocabulary.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAEH,OAAO,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,IAAI,CAAC;AAC/C,OAAO,EAAE,QAAQ,EAAE,IAAI,EAAE,MAAM,MAAM,CAAC;AAItC,8EAA8E;AAC9E,QAAQ;AACR,8EAA8E;AAE9E,8CAA8C;AAC9C,MAAM,SAAS,GAAuC;IACpD,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,WAAW,EAAE,EAAE,EAAE,SAAS;IAC1E,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,OAAO;IACtD,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,UAAU,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,SAAS;CACxE,CAAC;AA6BF,0CAA0C;AAC1C,MAAM,YAAY,GAAkB,EAAE,CAAC;AAEvC,8EAA8E;AAC9E,kBAAkB;AAClB,8EAA8E;AAE9E;;GAEG;AACH,SAAS,YAAY,CAAC,OAAe,EAAE,KAAe;IACpD,MAAM,KAAK,GAAmB,EAAE,CAAC;IAEjC,8CAA8C;IAC9C,MAAM,SAAS,GAAG,kBAAkB,CAAC;IACrC,MAAM,SAAS,GAAa,EAAE,CAAC;IAC/B,IAAI,OAA+B,CAAC;IACpC,OAAO,CAAC,OAAO,GAAG,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACpD,MAAM,SAAS,GAAG,OAAO,CAAC,CAAC,CAAE,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QAC7D,IAAI,SAAS;YAAE,SAAS,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;IAC3C,CAAC;IACD,KAAK,CAAC,MAAM,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IAEtC,mEAAmE;IACnE,MAAM,SAAS,GAAG,yBAAyB,CAAC;IAC5C,MAAM,SAAS,GAAa,EAAE,CAAC;IAC/B,IAAI,OAA+B,CAAC;IACpC,OAAO,CAAC,OAAO,GAAG,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACpD,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QAChC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,MAAM,GAAG,GAAG,EAAE,CAAC;YACzC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACvB,CAAC;IACH,CAAC;IACD,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzB,KAAK,CAAC,UAAU,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC;IAClC,CAAC;IAED,yEAAyE;IACzE,MAAM,kBAAkB,GAAG,wCAAwC,CAAC;IACpE,MAAM,YAAY,GAAa,EAAE,CAAC;IAClC,IAAI,WAAmC,CAAC;IACxC,OAAO,CAAC,WAAW,GAAG,kBAAkB,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACjE,MAAM,KAAK,GAAG,WAAW,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QACrC,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,KAAK,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;YAC1C,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAC3B,CAAC;IACH,CAAC;IAED,IAAI,YAAY,CAAC,CAAC,CAAC;QAAE,KAAK,CAAC,cAAc,GAAG,YAAY,CAAC,CAAC,CAAC,CAAC;IAC5D,IAAI,YAAY,CAAC,CAAC,CAAC;QAAE,KAAK,CAAC,aAAa,GAAG,YAAY,CAAC,CAAC,CAAC,CAAC;IAE3D,kCAAkC;IAClC,MAAM,UAAU,GAAG,OAAO,CAAC,KAAK,CAAC,mBAAmB,CAAC,CAAC;IACtD,IAAI,UAAU;QAAE,KAAK,CAAC,SAAS,GAAG,UAAU,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;IAExD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,8EAA8E;AAC9E,mBAAmB;AACnB,8EAA8E;AAE9E;;GAEG;AACH,SAAS,eAAe,CAAC,YAAoB,MAAM,EAAE,aAA4B,IAAI;IACnF,MAAM,OAAO,GAAoB,EAAqB,CAAC;IAEvD,KAAK,MAAM,IAAI,IAAI,MAAM,CAAC,IAAI,CAAC,SAAS,CAAe,EAAE,CAAC;QACxD,OAAO,CAAC,IAAI,CAAC,GAAG;YACd,QAAQ,EAAE,SAAS,CAAC,IAAI,CAAC;YACzB,IAAI,EAAE,IAAI;YACV,OAAO,EAAE,EAAE;SACZ,CAAC;IACJ,CAAC;IAED,IAAI,CAAC;QACH,MAAM,KAAK,GAAG,WAAW,CAAC,SAAS,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE;YAChD,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC;gBAAE,OAAO,KAAK,CAAC;YACvC,IAAI,UAAU,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,UAAU,CAAC;gBAAE,OAAO,KAAK,CAAC;YACxD,OAAO,IAAI,CAAC;QACd,CAAC,CAAC,CAAC;QAEH,OAAO,CAAC,GAAG,CAAC,cAAc,KAAK,CAAC,MAAM,kBAAkB,SAAS,GAAG,CAAC,CAAC;QACtE,IAAI,UAAU,EAAE,CAAC;YACf,OAAO,CAAC,GAAG,CAAC,8BAA8B,UAAU,KAAK,CAAC,CAAC;QAC7D,CAAC;QAED,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,oBAAoB,CAAC,CAAC;YAC/C,IAAI,CAAC,KAAK,EAAE,CAAC;gBACX,YAAY,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,8BAA8B,EAAE,CAAC,CAAC;gBACpE,SAAS;YACX,CAAC;YAED,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAW,CAAC;YAChC,IAAI,CAAC,OAAO,CAAC,IAAgB,CAAC,EAAE,CAAC;gBAC/B,YAAY,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,0BAA0B,IAAI,EAAE,EAAE,CAAC,CAAC;gBACtE,SAAS;YACX,CAAC;YAED,IAAI,CAAC;gBACH,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,SAAS,EAAE,IAAI,CAAC,EAAE,OAAO,CAAC,CAAC;gBAC7D,MAAM,KAAK,GAAG,YAAY,CAAC,OAAO,EAAE,IAAgB,CAAC,CAAC;gBAEtD,yBAAyB;gBACzB,IAAI,WAAW,GAAG,SAAS,CAAC;gBAC5B,IAAI,IAAI,CAAC,QAAQ,CAAC,WAAW,CAAC;oBAAE,WAAW,GAAG,mBAAmB,CAAC;qBAC7D,IAAI,IAAI,CAAC,QAAQ,CAAC,aAAa,CAAC;oBAAE,WAAW,GAAG,cAAc,CAAC;qBAC/D,IAAI,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC;oBAAE,WAAW,GAAG,SAAS,CAAC;qBACrD,IAAI,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC;oBAAE,WAAW,GAAG,kBAAkB,CAAC;qBAC/D,IAAI,IAAI,CAAC,QAAQ,CAAC,YAAY,CAAC;oBAAE,WAAW,GAAG,YAAY,CAAC;gBAEjE,OAAO,CAAC,IAAgB,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC;oBACrC,IAAI,EAAE,QAAQ,CAAC,IAAI,CAAC;oBACpB,IAAI,EAAE,WAAW;oBACjB,KAAK;iBACN,CAAC,CAAC;YACL,CAAC;YAAC,OAAO,KAAc,EAAE,CAAC;gBACxB,YAAY,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,eAAgB,KAAe,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;YACjF,CAAC;QACH,CAAC;IACH,CAAC;IAAC,OAAO,KAAc,EAAE,CAAC;QACxB,OAAO,CAAC,KAAK,CAAC,4BAA6B,KAAe,CAAC,OAAO,EAAE,CAAC,CAAC;QACtE,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,8EAA8E;AAC9E,SAAS;AACT,8EAA8E;AAE9E;;GAEG;AACH,SAAS,cAAc,CAAC,OAAwB;IAC9C,OAAO,CAAC,GAAG,CAAC,4CAA4C,CAAC,CAAC;IAC1D,OAAO,CAAC,GAAG,CAAC,sCAAsC,CAAC,CAAC;IACpD,OAAO,CAAC,GAAG,CAAC,4CAA4C,CAAC,CAAC;IAE1D,KAAK,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAsC,EAAE,CAAC;QACxF,IAAI,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,SAAS;QAExC,OAAO,CAAC,GAAG,CAAC,QAAQ,IAAI,CAAC,QAAQ,KAAK,IAAI,CAAC,WAAW,EAAE,GAAG,CAAC,CAAC;QAC7D,OAAO,CAAC,GAAG,CAAC,qBAAqB,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;QAExD,wBAAwB;QACxB,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;QACtC,MAAM,eAAe,GAAG,IAAI,GAAG,EAAU,CAAC;QAC1C,MAAM,cAAc,GAAG,IAAI,GAAG,EAAU,CAAC;QACzC,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;QAEtC,KAAK,MAAM,MAAM,IAAI,IAAI,CAAC,OAAO,EAAE,CAAC;YAClC,IAAI,MAAM,CAAC,KAAK,CAAC,UAAU;gBAAE,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC;YACtE,IAAI,MAAM,CAAC,KAAK,CAAC,cAAc;gBAAE,eAAe,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,cAAc,CAAC,CAAC;YAClF,IAAI,MAAM,CAAC,KAAK,CAAC,aAAa;gBAAE,cAAc,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,aAAa,CAAC,CAAC;YAC/E,IAAI,MAAM,CAAC,KAAK,CAAC,SAAS;gBAAE,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;QACtE,CAAC;QAED,IAAI,WAAW,CAAC,IAAI,GAAG,CAAC;YAAE,OAAO,CAAC,GAAG,CAAC,sBAAsB,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAClG,IAAI,eAAe,CAAC,IAAI,GAAG,CAAC;YAAE,OAAO,CAAC,GAAG,CAAC,kBAAkB,KAAK,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACtG,IAAI,cAAc,CAAC,IAAI,GAAG,CAAC;YAAE,OAAO,CAAC,GAAG,CAAC,iBAAiB,KAAK,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAEnG,sEAAsE;QACtE,MAAM,gBAAgB,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,mBAAmB,CAAC,CAAC;QACpF,MAAM,gBAAgB,GACpB,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,IAAI,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC;YACzE,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,IAAI,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAExE,IAAI,gBAAgB,IAAI,gBAAgB,CAAC,KAAK,CAAC,MAAM,IAAI,gBAAgB,CAAC,KAAK,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAClG,OAAO,CAAC,GAAG,CAAC,oBAAoB,gBAAgB,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAC1F,CAAC;IACH,CAAC;IAED,kBAAkB;IAClB,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,OAAO,CAAC,GAAG,CAAC,wCAAwC,CAAC,CAAC;QACtD,OAAO,CAAC,GAAG,CAAC,uCAAuC,CAAC,CAAC;QACrD,OAAO,CAAC,GAAG,CAAC,kBAAkB,YAAY,CAAC,MAAM,IAAI,CAAC,CAAC;QAEvD,kBAAkB;QAClB,MAAM,QAAQ,GAA6B,EAAE,CAAC;QAC9C,KAAK,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,YAAY,EAAE,CAAC;YAC5C,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC;gBAAE,QAAQ,CAAC,MAAM,CAAC,GAAG,EAAE,CAAC;YAC7C,QAAQ,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC9B,CAAC;QAED,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,EAAE,CAAC;YACvD,OAAO,CAAC,GAAG,CAAC,GAAG,MAAM,KAAK,KAAK,CAAC,MAAM,UAAU,CAAC,CAAC;YAClD,IAAI,KAAK,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;gBACtB,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC;YAChD,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC;gBAC1D,OAAO,CAAC,GAAG,CAAC,aAAa,KAAK,CAAC,MAAM,GAAG,CAAC,OAAO,CAAC,CAAC;YACpD,CAAC;YACD,OAAO,CAAC,GAAG,EAAE,CAAC;QAChB,CAAC;IACH,CAAC;IAED,OAAO,CAAC,GAAG,CAAC,4CAA4C,CAAC,CAAC;IAC1D,OAAO,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAC;IAClC,OAAO,CAAC,GAAG,CAAC,4CAA4C,CAAC,CAAC;AAC5D,CAAC;AAED,8EAA8E;AAC9E,MAAM;AACN,8EAA8E;AAE9E,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;AACnC,IAAI,UAAU,GAAkB,IAAI,CAAC;AACrC,IAAI,SAAS,GAAG,MAAM,CAAC;AAEvB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;IACrC,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,eAAe,IAAI,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;QAC/C,UAAU,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC;QAC1B,CAAC,EAAE,CAAC;IACN,CAAC;SAAM,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,aAAa,IAAI,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;QACpD,SAAS,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC;QACzB,CAAC,EAAE,CAAC;IACN,CAAC;SAAM,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,QAAQ,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACpD,OAAO,CAAC,GAAG,CAAC;;;;;;;;;;;;CAYf,CAAC,CAAC;QACC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;AACH,CAAC;AAED,eAAe;AACf,MAAM,OAAO,GAAG,eAAe,CAAC,SAAS,EAAE,UAAU,CAAC,CAAC;AACvD,cAAc,CAAC,OAAO,CAAC,CAAC"}

package/dist/scripts/fix-article-navigation.d.ts

@@ -0,0 +1,37 @@
+#!/usr/bin/env -S npx tsx
+/**
+ * Fix Article Navigation: Language Switcher + Back-to-News Top Nav
+ *
+ * This script ensures ALL news articles have:
+ * 1. A language switcher nav (14 languages) after <body>
+ * 2. An article-top-nav div with a localized back-to-news link before the article
+ *
+ * It auto-discovers all articles in the news/ directory and processes them
+ * idempotently — safe to run multiple times.
+ *
+ * Usage:
+ *   npx tsx scripts/fix-article-navigation.ts
+ *   npx tsx scripts/fix-article-navigation.ts --dry-run
+ *
+ * @author Hack23 AB
+ * @license Apache-2.0
+ */
+import type { Language } from './types/language.js';
+export interface ProcessResult {
+    addedSwitcher: boolean;
+    addedTopnav: boolean;
+    fixedTopnav: boolean;
+}
+/**
+ * Transform article HTML string: ensures language-switcher and article-top-nav
+ * with a back-to-news link are present. Idempotent — safe to call multiple times.
+ *
+ * @param content   Original HTML string
+ * @param baseSlug  Article base slug without directory prefix (e.g. "2026-01-01-article")
+ * @param lang      Language code for this variant
+ * @returns Updated HTML string and flags indicating what changed
+ */
+export declare function transformContent(content: string, baseSlug: string, lang: Language): {
+    content: string;
+} & ProcessResult;
+//# sourceMappingURL=fix-article-navigation.d.ts.map

package/dist/scripts/fix-article-navigation.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"fix-article-navigation.d.ts","sourceRoot":"","sources":["../../scripts/fix-article-navigation.ts"],"names":[],"mappings":";AACA;;;;;;;;;;;;;;;;GAgBG;AAKH,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AA8BpD,MAAM,WAAW,aAAa;IAC5B,aAAa,EAAE,OAAO,CAAC;IACvB,WAAW,EAAE,OAAO,CAAC;IACrB,WAAW,EAAE,OAAO,CAAC;CACtB;AAED;;;;;;;;GAQG;AACH,wBAAgB,gBAAgB,CAC9B,OAAO,EAAE,MAAM,EACf,QAAQ,EAAE,MAAM,EAChB,IAAI,EAAE,QAAQ,GACb;IAAE,OAAO,EAAE,MAAM,CAAA;CAAE,GAAG,aAAa,CA6ErC"}

package/dist/scripts/fix-article-navigation.js

@@ -0,0 +1,198 @@
+#!/usr/bin/env -S npx tsx
+/**
+ * Fix Article Navigation: Language Switcher + Back-to-News Top Nav
+ *
+ * This script ensures ALL news articles have:
+ * 1. A language switcher nav (14 languages) after <body>
+ * 2. An article-top-nav div with a localized back-to-news link before the article
+ *
+ * It auto-discovers all articles in the news/ directory and processes them
+ * idempotently — safe to run multiple times.
+ *
+ * Usage:
+ *   npx tsx scripts/fix-article-navigation.ts
+ *   npx tsx scripts/fix-article-navigation.ts --dry-run
+ *
+ * @author Hack23 AB
+ * @license Apache-2.0
+ */
+import fs from 'fs';
+import path from 'path';
+import { fileURLToPath } from 'url';
+import { ALL_LANG_CODES, FOOTER_LABELS } from './article-template/constants.js';
+import { getNewsIndexFilename, generateArticleLanguageSwitcher } from './article-template/helpers.js';
+// ── Helpers ───────────────────────────────────────────────────────────────
+function extractLang(filename) {
+    const name = filename.replace(/\.html$/, '');
+    for (const lang of ALL_LANG_CODES) {
+        if (name.endsWith(`-${lang}`))
+            return lang;
+    }
+    return null;
+}
+function extractBase(filename) {
+    const name = filename.replace(/\.html$/, '');
+    for (const lang of ALL_LANG_CODES) {
+        if (name.endsWith(`-${lang}`))
+            return name.slice(0, -(lang.length + 1));
+    }
+    return null;
+}
+function generateTopNav(lang) {
+    const label = FOOTER_LABELS[lang].backToNews;
+    const index = getNewsIndexFilename(lang);
+    return `\n<div class="article-top-nav">\n  <a href="${index}" class="back-to-news">\n    ← ${label}\n  </a>\n</div>\n`;
+}
+/**
+ * Transform article HTML string: ensures language-switcher and article-top-nav
+ * with a back-to-news link are present. Idempotent — safe to call multiple times.
+ *
+ * @param content   Original HTML string
+ * @param baseSlug  Article base slug without directory prefix (e.g. "2026-01-01-article")
+ * @param lang      Language code for this variant
+ * @returns Updated HTML string and flags indicating what changed
+ */
+export function transformContent(content, baseSlug, lang) {
+    let result = content;
+    let addedSwitcher = false;
+    let addedTopnav = false;
+    let fixedTopnav = false;
+    // ── 1. Language switcher ──────────────────────────────────────────
+    const hasSwitcher = result.includes('language-switcher');
+    if (!hasSwitcher) {
+        const switcherHtml = generateArticleLanguageSwitcher(baseSlug, lang);
+        // Prefer inserting AFTER the skip-link so it remains the first focusable element.
+        const skipLinkPattern = /(<a[^>]*class="skip-link"[^>]*>[\s\S]*?<\/a>)/;
+        if (skipLinkPattern.test(result)) {
+            result = result.replace(skipLinkPattern, `$1\n${switcherHtml}`);
+        }
+        else {
+            result = result.replace(/(<body[^>]*>)/, `$1\n${switcherHtml}`);
+        }
+        addedSwitcher = true;
+    }
+    else {
+        // Update existing switcher to have all 14 languages.
+        // Use [^\S\n]* (spaces/tabs, not newlines) to consume any indentation before <nav>,
+        // so the replacement string's own leading spaces don't accumulate on repeated runs.
+        const newSwitcher = generateArticleLanguageSwitcher(baseSlug, lang);
+        result = result.replace(/[^\S\n]*<nav class="language-switcher"[^>]*>[\s\S]*?<\/nav>/, newSwitcher);
+    }
+    // ── 2. article-top-nav ────────────────────────────────────────────
+    const hasTopnav = result.includes('article-top-nav');
+    // If top-nav exists but is missing back-to-news link, replace it
+    if (hasTopnav) {
+        const topNavHasBackLink = /<div class="article-top-nav">[\s\S]*?class="back-to-news"[\s\S]*?<\/div>/.test(result);
+        if (!topNavHasBackLink) {
+            result = result.replace(/<div class="article-top-nav">[\s\S]*?<\/div>/, generateTopNav(lang).trim());
+            fixedTopnav = true;
+        }
+    }
+    if (!hasTopnav) {
+        const topNavHtml = generateTopNav(lang);
+        let inserted = false;
+        // Pattern A: insert after closing </nav> of language-switcher, before article/div.news-article
+        // Note: the article/div tag may have attributes before class= (e.g. id="main-content")
+        // and may carry additional classes (e.g. "news-article article-type-propositions")
+        if (result.includes('</nav>')) {
+            const navPattern = /((<\/nav>)([\s]*)(<(?:article|div)\b[^>]*?\bclass="(?:news-article(?:\s[^"]*)?|container)"))/s;
+            const match = navPattern.exec(result);
+            if (match) {
+                const endOfNav = match.index + match[2].length;
+                const whitespace = match[3];
+                const articleTag = match[4];
+                const afterFull = result.slice(match.index + match[0].length);
+                result = result.slice(0, endOfNav) + topNavHtml + whitespace + articleTag + afterFull;
+                inserted = true;
+            }
+        }
+        // Pattern B: insert directly before <article/div with class="news-article ..." or "container">
+        // Allows arbitrary attributes (e.g. id="main-content") before class=
+        if (!inserted) {
+            const articlePattern = /(<(?:article|div)\b[^>]*?\bclass="(?:news-article(?:\s[^"]*)?|container)")/;
+            const match = articlePattern.exec(result);
+            if (match) {
+                result = result.slice(0, match.index) + topNavHtml + '\n' + result.slice(match.index);
+                inserted = true;
+            }
+        }
+        if (inserted)
+            addedTopnav = true;
+    }
+    return { content: result, addedSwitcher, addedTopnav, fixedTopnav };
+}
+function processArticle(filepath, baseSlug, lang, dryRun) {
+    const original = fs.readFileSync(filepath, 'utf-8');
+    const { content, addedSwitcher, addedTopnav, fixedTopnav } = transformContent(original, baseSlug, lang);
+    // ── Write if changed ──────────────────────────────────────────────
+    if (content !== original && !dryRun) {
+        fs.writeFileSync(filepath, content, 'utf-8');
+    }
+    return { addedSwitcher, addedTopnav, fixedTopnav };
+}
+function discoverArticles(newsDir) {
+    const articles = {};
+    const files = fs.readdirSync(newsDir).sort();
+    for (const name of files) {
+        if (!name.endsWith('.html') || name.startsWith('index'))
+            continue;
+        const lang = extractLang(name);
+        const base = extractBase(name);
+        if (lang && base) {
+            if (!articles[base])
+                articles[base] = {};
+            articles[base][lang] = path.join(newsDir, name);
+        }
+    }
+    return articles;
+}
+function main() {
+    const dryRun = process.argv.includes('--dry-run');
+    if (dryRun) {
+        console.log('=== DRY RUN — no files will be modified ===\n');
+    }
+    const scriptDir = path.dirname(fileURLToPath(import.meta.url));
+    const newsDir = path.resolve(scriptDir, '..', 'news');
+    if (!fs.existsSync(newsDir)) {
+        console.error(`ERROR: news directory not found at ${newsDir}`);
+        process.exit(1);
+    }
+    console.log('=== Fix Article Navigation ===');
+    console.log(`News directory: ${newsDir}\n`);
+    const articles = discoverArticles(newsDir);
+    const slugs = Object.keys(articles).sort();
+    console.log(`Discovered ${slugs.length} unique article slugs\n`);
+    let total = 0;
+    let switchersAdded = 0;
+    let topnavsAdded = 0;
+    let topnavsFixed = 0;
+    for (const baseSlug of slugs) {
+        const langFiles = articles[baseSlug];
+        for (const lang of ALL_LANG_CODES) {
+            const filepath = langFiles[lang];
+            if (!filepath)
+                continue;
+            total++;
+            const { addedSwitcher, addedTopnav, fixedTopnav } = processArticle(filepath, baseSlug, lang, dryRun);
+            if (addedSwitcher)
+                switchersAdded++;
+            if (addedTopnav)
+                topnavsAdded++;
+            if (fixedTopnav)
+                topnavsFixed++;
+        }
+    }
+    console.log('=== Summary ===');
+    console.log(`Total files processed: ${total}`);
+    console.log(`Language switchers added: ${switchersAdded}`);
+    console.log(`Top nav (article-top-nav) added: ${topnavsAdded}`);
+    console.log(`Top nav fixed (missing back-to-news link): ${topnavsFixed}`);
+    if (dryRun) {
+        console.log('\n(Dry run — no files were modified)');
+    }
+    console.log('\n✓ Done!');
+}
+if (import.meta.url === `file://${process.argv[1]}`) {
+    main();
+}
+//# sourceMappingURL=fix-article-navigation.js.map