Search: index author names + friendly empty state (#3)

scripts/smoke.mjs

@@ -145,4 +145,13 @@ check('snippet falls back to the start of the text when nothing matches', () =>
   assert.ok(s.startsWith('Gratitude'));
 });
 
+check('search matches an author name', () => {
+  const idx = new SearchIndex().build([
+    { title: 'On stillness', themes: ['Silence'], author: 'Nirodbaran', summary: 'x', body: 'y' },
+  ]);
+  const hits = idx.search('Nirodbaran');
+  assert.equal(hits.length, 1);
+  assert.ok(hits[0].matched.includes('nirodbaran'));
+});
+
 console.log(`\n${passed} checks passed.`);

src/reader/reader.js

@@ -250,6 +250,10 @@ function renderList(items) {
 /** Render search hits with a highlighted contextual snippet (issue #3). */
 function renderSearchResults(hits) {
   els.results.replaceChildren();
+  if (!hits.length) {
+    els.results.append(tag('li', 'card', 'No articles match that search. Try a broader word or a theme.'));
+    return;
+  }
   for (const { article, matched } of hits) {
     const li = document.createElement('li');
     li.className = 'card';

src/search.js

@@ -2,7 +2,7 @@
  * search.js — small, dependency-free full-text search for an edition.
  *
  * Builds an in-memory inverted index over each article's title, themes,
- * summary, and body, then ranks matches with TF-IDF plus light field
+ * author, summary, and body, then ranks matches with TF-IDF plus light field
  * weighting (a hit in the title counts for more than one in the body). It is
  * meant for a single edition's worth of articles — hundreds, not millions —
  * which is exactly the scale of a monthly magazine, and it keeps the whole
@@ -15,6 +15,7 @@
 const FIELD_WEIGHTS = Object.freeze({
   title: 3,
   themes: 2,
+  author: 2,
   summary: 1.5,
   body: 1,
 });