From be6e8d1dddb386b2d34cd2566400672817c3e415 Mon Sep 17 00:00:00 2001
From: Mat Manna <git@matmanna.dev>
Date: Fri, 8 May 2026 21:32:55 -0400
Subject: [PATCH] fix: excerpt block extraction (#44)

---
 src/lib/content.ts    | 17 ++++++++++++++++-
 src/pages/index.astro | 18 ++++++++++++++++--
 src/styles/main.scss  | 11 ++++-------
 3 files changed, 36 insertions(+), 10 deletions(-)

diff --git a/src/lib/content.ts b/src/lib/content.ts
index a1c591d..dd1d025 100644
--- a/src/lib/content.ts
+++ b/src/lib/content.ts
@@ -170,14 +170,29 @@ function extractTextBlocks(body: string): string[] {
 
     return blocks
         .slice(leadingImage ? 1 : 0)
+        .filter((block) => {
+            if (/^#{1,6}\s+/.test(block.trim())) return false;
+            if (/^[=-]{3,}\s*$/.test(block.trim())) return false;
+            return true;
+        })
         .map((block) => stripMarkdown(block))
         .filter(Boolean);
 }
 
 function toExcerpt(entry: { body: string; data: { excerpt?: string } }): string {
     const explicitExcerpt = entry.data.excerpt;
-    const source = explicitExcerpt ?? extractTextBlocks(entry.body)[0] ?? "";
+    if (explicitExcerpt) {
+        return stripMarkdown(explicitExcerpt);
+    }
+
+    const blocks = extractTextBlocks(entry.body);
+    const substantial = blocks.find((b) => {
+        const cleaned = stripMarkdown(b);
+        const wordCount = cleaned.split(/\s+/).filter(Boolean).length;
+        return wordCount >= 8 && cleaned.length >= 60;
+    });
 
+    const source = substantial ?? blocks[0] ?? "";
     return stripMarkdown(source);
 }
 
diff --git a/src/pages/index.astro b/src/pages/index.astro
index a9d7c2c..3e569e0 100644
--- a/src/pages/index.astro
+++ b/src/pages/index.astro
@@ -17,7 +17,21 @@ const overrideHeadline = site.headlineOverride
   ? posts.find((post) => post.slug === site.headlineOverride)
   : undefined;
 const headline = overrideHeadline ?? allNews[0];
-const headlineExcerpts = headline?.paragraphs ?? [];
+const headlineExcerpts = (() => {
+  const paras = headline?.paragraphs ?? [];
+  const filtered = paras.filter((p) => {
+    const s = p.replace(/\s+/g, " ").trim();
+    if (!s) return false;
+    if (/^#{1,6}\s+/.test(s)) return false;
+    if (/^[=-]{3,}\s*$/.test(s)) return false;
+    const words = s.split(/\s+/).length;
+    if (words < 6) return false;
+    if (s.length < 40) return false;
+    return true;
+  });
+
+  return filtered.slice(0, headline?.leadingImage ? 1 : 2);
+})();
 
 const displayOpinion = posts.filter((post) => post.category === "opinion").slice(0, 3);
 
@@ -51,7 +65,7 @@ const acknowledgements = await getAcknowledgements();
             <hr class="block-rule" />
             <div class="block-excerpts">
               {headlineExcerpts.map((paragraph) => (
-                <p class="block-excerpt">{headline.leadingImage ? paragraph : truncateWords(paragraph, 28)}</p>
+                <p class="block-excerpt">{headline.leadingImage ? paragraph : truncateWords(paragraph, 120)}</p>
               ))}
             </div>
           </div>
diff --git a/src/styles/main.scss b/src/styles/main.scss
index 1a041ce..f98e6f3 100644
--- a/src/styles/main.scss
+++ b/src/styles/main.scss
@@ -139,6 +139,7 @@ a {
   display: block;
   color: #000;
   transition: background 0.15s;
+  overflow: hidden;
 
   &:hover {
     background: #f0f0eb;
@@ -188,11 +189,10 @@ a {
   .block-excerpts {
     flex: 1;
     min-height: 0;
-    max-height: calc(1.6em * 10);
     display: -webkit-box;
     -webkit-box-orient: vertical;
-    -webkit-line-clamp: 10;
-    line-clamp: 10;
+    -webkit-line-clamp: 6;
+    line-clamp: 6;
     overflow: hidden;
     text-overflow: ellipsis;
   }
@@ -200,10 +200,7 @@ a {
     font-size: 1rem;
     line-height: 1.6;
     color: #222;
-    margin: 0;
-  }
-  .block-excerpt + .block-excerpt {
-    margin-top: 0.8em;
+    margin: 0 0 1.2em 0;
   }
 }
 
-- 
2.51.2