euparliamentmonitor 0.9.11 → 0.9.13

package/README.md

@@ -136,7 +136,7 @@ The published site is the audience-facing companion to this npm/TypeScript packa
 
 **MCP Server Integration**: The project uses the
 [European-Parliament-MCP-Server](https://github.com/Hack23/European-Parliament-MCP-Server)
-v1.3.3 for accessing real EU Parliament data via the Model Context Protocol.
+v1.3.6 for accessing real EU Parliament data via the Model Context Protocol.
 
 - **MCP Server Status**: ✅ Fully operational — 60+ EP data tools available
   (feeds, direct lookups, analytical tools, intelligence correlation)
@@ -432,7 +432,7 @@ import type { ArticleCategory, LanguageCode } from 'euparliamentmonitor/types';
 
 ## 🔌 Data Sources
 
-**Primary — European Parliament MCP Server** ([Hack23/European-Parliament-MCP-Server](https://github.com/Hack23/European-Parliament-MCP-Server) v1.3.3+, fully operational):
+**Primary — European Parliament MCP Server** ([Hack23/European-Parliament-MCP-Server](https://github.com/Hack23/European-Parliament-MCP-Server) v1.3.6+, fully operational):
 
 - 🗳️ Plenary sessions, voting records, roll-call votes
 - 📜 Adopted texts, motions, resolutions, urgency files

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "euparliamentmonitor",
-  "version": "0.9.11",
+  "version": "0.9.13",
   "type": "module",
   "description": "European Parliament Intelligence Platform - Monitor political activity with systematic transparency",
   "main": "scripts/index.js",
@@ -63,6 +63,9 @@
     "build:check-tests": "tsc --project tsconfig.test.json --noEmit",
     "copy-vendor": "node scripts/copy-vendor.js",
     "validate-analysis": "node scripts/validate-analysis-completeness.js",
+    "discover:untranslated-briefs": "node scripts/discover-untranslated-briefs.js",
+    "validate:translations": "node scripts/validate-brief-translations.js",
+    "validate:manifest-seo": "node scripts/validate-manifest-seo.js",
     "sync:templates": "node scripts/templates/sync-template-frontmatter.js",
     "sync:templates:check": "node scripts/templates/sync-template-frontmatter.js --check",
     "prior-run-diff": "node scripts/aggregator/prior-run-diff.js",
@@ -74,6 +77,7 @@
     "optimize-css": "node scripts/optimize-css.js",
     "minify-assets": "node scripts/minify-assets.js",
     "validate-ep-api": "npx tsx src/utils/validate-ep-api.ts",
+    "mcp:probe": "npx tsx src/utils/mcp-probe.ts",
     "lint:prompts": "node scripts/lint-prompts.js",
     "htmlhint": "sh -c 'htmlhint *.html; set -- news/*.html; if [ -e \"$1\" ]; then htmlhint \"$@\"; else echo \"No news/*.html files to lint\"; fi'",
     "serve": "python3 -m http.server 8080",
@@ -149,7 +153,7 @@
     "@playwright/test": "1.60.0",
     "@types/d3": "7.4.3",
     "@types/markdown-it": "^14.1.2",
-    "@types/node": "25.7.0",
+    "@types/node": "25.8.0",
     "@types/papaparse": "5.5.2",
     "@typescript-eslint/eslint-plugin": "8.59.3",
     "@typescript-eslint/parser": "8.59.3",
@@ -159,7 +163,7 @@
     "chartjs-plugin-annotation": "3.1.0",
     "clean-css": "^5.3.3",
     "d3": "7.9.0",
-    "eslint": "10.3.0",
+    "eslint": "10.4.0",
     "eslint-config-prettier": "10.1.8",
     "eslint-plugin-jsdoc": "62.9.0",
     "eslint-plugin-security": "4.0.0",
@@ -168,9 +172,9 @@
     "html-minifier-terser": "^7.2.0",
     "htmlhint": "1.9.2",
     "husky": "9.1.7",
-    "jscpd": "4.2.0",
+    "jscpd": "4.2.2",
     "knip": "^6.7.0",
-    "lint-staged": "17.0.4",
+    "lint-staged": "17.0.5",
     "mermaid": "11.15.0",
     "papaparse": "5.5.3",
     "prettier": "3.8.3",
@@ -187,7 +191,7 @@
     "node": ">=26"
   },
   "dependencies": {
-    "european-parliament-mcp-server": "1.3.3",
+    "european-parliament-mcp-server": "1.3.6",
     "markdown-it": "^14.1.1",
     "markdown-it-anchor": "^9.2.0",
     "markdown-it-attrs": "^4.3.1",

package/scripts/aggregator/analysis-aggregator.js

@@ -18,6 +18,10 @@ import { buildKeyTakeaways, KEY_TAKEAWAYS_SECTION_ID, KEY_TAKEAWAYS_SECTION_TITL
 import { flattenManifestFiles as _flattenManifestFiles, latestGateResult as _latestGateResult, resolveArticleType as _resolveArticleType, resolveRunId as _resolveRunId, } from './manifest/index.js';
 import { READER_GUIDE_SECTION_ID, READER_GUIDE_SECTION_IDS, READER_GUIDE_SECTION_TITLE, } from './reader-guide-constants.js';
 export { READER_GUIDE_SECTION_ID, READER_GUIDE_SECTION_IDS, READER_GUIDE_SECTION_TITLE, } from './reader-guide-constants.js';
+const TRADECRAFT_EXCLUDED_FILES = new Set([
+    'analysis/methodologies/executive-brief-translation-guide.md',
+    'analysis/templates/executive-brief-translation-template.md',
+]);
 /**
  * Normalise `manifest.files` into a flat list of `runRelPath` strings.
  *
@@ -94,8 +98,9 @@ export function discoverTradecraftFiles(repoRoot) {
             continue;
         const entries = fs.readdirSync(dir, { withFileTypes: true });
         for (const entry of entries) {
-            if (entry.isFile() && entry.name.endsWith('.md')) {
-                result.push(`${sub}/${entry.name}`);
+            const rel = `${sub}/${entry.name}`;
+            if (entry.isFile() && entry.name.endsWith('.md') && !TRADECRAFT_EXCLUDED_FILES.has(rel)) {
+                result.push(rel);
             }
         }
     }

package/scripts/aggregator/article-generator.js

@@ -444,7 +444,7 @@ function getMetadataEntry(map, lang) {
         return descriptor.value;
     }
     const en = Object.getOwnPropertyDescriptor(map, 'en')?.value;
-    return en ?? { title: '', description: '', keywords: [] };
+    return en ?? { title: '', description: '', keywords: [], source: 'template' };
 }
 /**
  * Count the number of articles the site currently publishes, derived
@@ -699,6 +699,7 @@ function applyCliOverrides(base, titleOverride, descriptionOverride) {
                 title: titleOverride ?? entry.title,
                 description: descriptionOverride ?? entry.description,
                 keywords: entry.keywords,
+                source: titleOverride || descriptionOverride ? 'manifest' : entry.source,
             },
             enumerable: true,
             writable: true,

package/scripts/aggregator/article-metadata.d.ts

@@ -4,6 +4,19 @@ export interface ResolvedMetadataEntry {
     readonly title: string;
     readonly description: string;
     readonly keywords: readonly string[];
+    /**
+     * `"localized-brief"` when the title/description came from a translated
+     * `executive-brief_<lang>.md`; `"english-brief"` when the locale fell
+     * through to the English brief; `"english-editorial"` when the locale
+     * used an aggregated-Markdown / artefact source; `"template"` when only
+     * the localized type/date template was available. For `lang === 'en'`
+     * the value is always `"english-brief"` or `"english-editorial"` or
+     * `"template"` (no fall-through). Lets downstream consumers — the
+     * news-index, the static-site fallback note, the manifest-SEO
+     * validator — record the asymmetry called out in
+     * [`.github/prompts/04-article-generation.md`](../../.github/prompts/04-article-generation.md) § 6.2 priority 3.
+     */
+    readonly source: 'manifest' | 'localized-brief' | 'english-brief' | 'english-editorial' | 'template';
 }
 /** Fully resolved metadata — one entry per supported language. */
 export type ResolvedMetadata = LanguageMap<ResolvedMetadataEntry>;
@@ -113,6 +126,33 @@ export declare function truncateDescription(text: string): string;
  * @returns Truncated title with trailing ellipsis when clipped
  */
 export declare function truncateTitle(text: string): string;
+/**
+ * Return the first complete sentence from a prose paragraph, suitable
+ * for use as a fallback editorial title when the artefact H1 is
+ * categorical (e.g. `# EU Parliament Committee Reports`) and the
+ * resolver must derive `<title>` from the BLUF / lede summary instead.
+ *
+ * A "sentence" is the prefix up to the first sentence-terminator
+ * (`. `, `! `, `? `, `; `) inside the `[HEADLINE_SOFT_MIN,
+ * TITLE_MAX_LENGTH]` window. Common abbreviations (`Q1.`, `Q2.`,
+ * `H1.`, `H2.`, `Mr.`, `Mrs.`, `e.g.`, `i.e.`, `vs.`) are skipped
+ * so they don't terminate the sentence prematurely. When no
+ * acceptable terminator exists in the window, returns the entire
+ * input unchanged so {@link truncateTitle} can handle clause-boundary
+ * truncation downstream.
+ *
+ * This produces journalistically clean titles even for the
+ * propositions / committee-reports cases where the BLUF paragraph
+ * opens with a single long sentence that exceeds 140 chars —
+ * `truncateTitle` then breaks on a clause boundary, and the result is
+ * still grammatical because the input was a sentence prefix rather
+ * than an arbitrary paragraph slice.
+ *
+ * @param paragraph - Prose paragraph (post-{@link stripInlineMarkdown})
+ * @returns First sentence, or the original paragraph when none can be
+ *   identified within the soft-min window
+ */
+export declare function extractFirstSentence(paragraph: string): string;
 /**
  * Return the first Markdown H1 (`# …`) in the supplied text, stripped of
  * the leading `#` and trailing anchor syntax. Returns an empty string when
@@ -123,29 +163,19 @@ export declare function truncateTitle(text: string): string;
  */
 export declare function extractFirstH1(markdown: string): string;
 /**
- * Walk every line of the Markdown source and return the first line that
- * survives {@link shouldSkipDescriptionLine}. Inline Markdown decorations
- * are stripped and the result is truncated to fit `<meta description>`.
+ * Walk every line of the Markdown source and return the first paragraph
+ * that survives {@link shouldSkipDescriptionLine}. Consecutive non-blank
+ * prose lines are joined with a single space so hard-wrapped ledes
+ * (column-95 conventional wrap) produce a clean 140-180-character
+ * description rather than just the first 60-90-char line.
+ *
+ * Inline Markdown decorations are stripped and the result is truncated
+ * to fit `<meta description>`.
  *
  * @param markdown - Markdown source
  * @returns Prose description, or empty string when nothing qualifies
  */
 export declare function extractStrongProseLine(markdown: string): string;
-/**
- * Walk the body of an editorial artefact and, when it contains a `## …`
- * heading whose text matches one of `EDITORIAL_LEDE_HEADINGS`,
- * return the first prose paragraph that follows that heading. This is
- * the journalist's lede ("60-Second Read", "TL;DR", "BLUF — …", …) and
- * is exactly the sentence that should power `<meta description>` and
- * the OG/Twitter description fields.
- *
- * Returns the empty string when no lede heading is found or no qualifying
- * prose follows it. Inline Markdown is stripped and the result is
- * truncated to fit `<meta description>`.
- *
- * @param markdown - Editorial artefact source
- * @returns Lede paragraph, or empty string when none matched
- */
 export declare function extractLedeAfterHeading(markdown: string): string;
 /**
  * Return `true` when an artefact-H1 begins with one of the
@@ -217,6 +247,48 @@ export declare function extractArtifactHighlight(runDir: string, articleType: st
     readonly headline: string;
     readonly summary: string;
 } | null;
+/**
+ * Return `true` when a top-level `.md` filename looks like a translated
+ * sibling of a canonical editorial artefact (e.g.
+ * `executive-brief_ar.md`). These files must be excluded from the
+ * top-level fallback scan in {@link extractArtifactHighlight} because
+ * their localized H1s evade the English-only generic-heading detector
+ * and would otherwise hijack the English SEO surfaces.
+ *
+ * @param filename - Run-relative `.md` filename (no path separators)
+ * @returns `true` when the file is a translated sibling brief
+ */
+export declare function isTranslatedSiblingBrief(filename: string): boolean;
+/**
+ * Mine the FIRST named priority finding from an executive-brief–style
+ * artefact body. Looks for a section heading from
+ * {@link PRIORITY_FINDING_SECTION_HEADINGS} and returns the first dossier
+ * name + descriptive paragraph found inside it. Supports the three
+ * canonical Stage-B authoring patterns:
+ *
+ *   1. **Bold-in-numbered-list** (breaking briefs):
+ *      `1. **Digital Markets Act Enforcement** (TA-10-2026-0160, 2026-04-30)`
+ *      `   Parliament adopted a resolution …`
+ *   2. **Numbered subheading** (committee briefs):
+ *      `### 1. Clean Industrial Deal Implementation (ITRE/ENVI)`
+ *      `The Clean Industrial Deal framework …`
+ *   3. **Bold-leading paragraph** (synthesis variants):
+ *      `**Trigger 1: DMA Enforcement Resolution** (TA-10-2026-0160)`
+ *      `- Significance: 🟢 HIGH IMPACT …`
+ *
+ * Trailing parenthesised metadata (`(TA-10-2026-0160, 2026-04-30)`,
+ * `(ITRE/ENVI)`) is stripped from the headline so it stays headline-shaped
+ * (`Digital Markets Act Enforcement`) rather than boilerplate
+ * (`Digital Markets Act Enforcement (TA-10-2026-0160, 2026-04-30)`).
+ *
+ * @param body - Editorial artefact body
+ * @returns `{headline, summary}` when a priority finding was identified;
+ *   `null` when the body has no priority section or no usable item inside
+ */
+export declare function extractPriorityFindingHighlight(body: string): {
+    readonly headline: string;
+    readonly summary: string;
+} | null;
 /**
  * Build the per-language `{title, description}` pair using the
  * article-type–specific `*_TITLES` generator from