From 6ee61a7c67f7c408888078f65340f1836367ca20 Mon Sep 17 00:00:00 2001
From: "brobert (aider)" <borja@brobert.net>
Date: Tue, 1 Apr 2025 15:12:14 +0200
Subject: [PATCH] fix: preserve paragraph structure in epub generation

---
 generate-epub.ts | 21 ++++++++++++++-------
 1 file changed, 14 insertions(+), 7 deletions(-)

diff --git a/generate-epub.ts b/generate-epub.ts
index 1390860..617953c 100644
--- a/generate-epub.ts
+++ b/generate-epub.ts
@@ -48,14 +48,21 @@ async function generateEpub() {
 
   // Convert each post to ePub chapter format
   for (const post of posts) {
-    // Parse the content HTML and extract text nodes
+    // Parse the content HTML and preserve paragraph structure
     const { document } = parseHTML(`<div>${post.content}</div>`);
     
-    // Get all text content from paragraphs and other text containers
-    const paragraphs = [...document.querySelectorAll('p, div, span')]
-      .map(el => el.textContent?.trim())
-      .filter(Boolean)
-      .join('\n\n');
+    // Clean up content while preserving structure
+    const paragraphs = [...document.querySelectorAll('p')]
+      .map(p => {
+        // Remove any images but keep their alt text
+        const images = p.querySelectorAll('img');
+        images.forEach(img => {
+          const alt = img.getAttribute('alt') || '';
+          img.replaceWith(alt ? `[Imagen: ${alt}]` : '');
+        });
+        return p.outerHTML;
+      })
+      .join('');
 
     options.content.push({
       title: post.title,
@@ -63,7 +70,7 @@ async function generateEpub() {
       data: `
         <h1>${post.title}</h1>
         <p><em>Por ${post.author}</em></p>
-        ${paragraphs || 'No content available'}
+        ${paragraphs || '<p>No content available</p>'}
         <p><a href="${post.url}">Publicación original</a></p>
       `
     });