wiki-formant 0.47.0 → 0.47.1

package/dist/metadata.js

@@ -19,7 +19,6 @@
 // with no image while their og:image named a card, one listed citations, one
 // collection page had no ItemList at all.
 import { decodeEntities } from './markdown.js';
-import { stripHtml } from './text.js';
 export function pageMetadata(o) {
     const { title, description, url, type = 'website', image, imageSize = { width: 1200, height: 630 }, siteName, locale, handle, markdownTwin, article } = o;
     const images = image ? [{ url: image, ...imageSize, alt: o.imageAlt ?? title }] : undefined;
@@ -118,14 +117,20 @@ export function citationsFromReferences(items, max = 50) {
  * characters is a heading with no section, and is skipped.
  */
 export function questionPairs(html) {
+    // Plain text, not `stripHtml`, which keeps an href in parentheses for a model
+    // following citations: a heading's permalink anchor would end it in "(#id)"
+    // rather than "?". The anchor itself goes, glyph and all.
+    const plain = (s) => decodeEntities(s.replace(/<a\b[^>]*href="#[^"]*"[^>]*>[\s\S]*?<\/a>/gi, '').replace(/<[^>]+>/g, ' '))
+        .replace(/\s+/g, ' ')
+        .trim();
     const pairs = [];
     const sections = html.split(/<h[23][^>]*>/i);
     for (const section of sections.slice(1)) {
         const [heading = '', ...rest] = section.split(/<\/h[23]>/i);
-        const q = stripHtml(heading).replace(/\s+/g, ' ');
+        const q = plain(heading).replace(/\s+([?.,;:!])/g, '$1');
         if (!q.endsWith('?'))
             continue;
-        const a = stripHtml(rest.join('')).replace(/\s+/g, ' ');
+        const a = plain(rest.join('')).replace(/\s+([?.,;:!])/g, '$1');
         if (a.length >= 40)
             pairs.push({ q, a: a.slice(0, 1000) });
     }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "wiki-formant",
-  "version": "0.47.0",
+  "version": "0.47.1",
   "description": "The portable half of a wiki: derived taxonomy and facet controls, a version-negotiating MCP transport, markdown twins, block rendering, a rich-text editor engine, conditional-GET plumbing, and the headers and CSP every origin sends.",
   "type": "module",
   "license": "MIT",