refactor: improve paragraph chunking logic in content chunker

Author: DeepanshKhurana

build/utils/markdown-chunker.ts

@@ -350,23 +350,39 @@ export function chunkContent(
           currentLines = 0;
         }
       }
-    } else if (block.type === 'paragraph' && block.estimatedLines > linesPerPage) {
-      if (currentChunk.trim()) {
-        chunks.push(currentChunk.trim());
-        currentChunk = '';
-        currentLines = 0;
-      }
-
-      const sentences = block.content.match(/[^.!?]+[.!?]+/g) || [block.content];
-      for (const sentence of sentences) {
-        const sentenceLines = Math.ceil(sentence.length / charsPerLine);
-        if (currentLines + sentenceLines > linesPerPage && currentChunk.trim()) {
-          chunks.push(currentChunk.trim());
-          currentChunk = sentence;
-          currentLines = sentenceLines;
-        } else {
-          currentChunk += sentence;
-          currentLines += sentenceLines;
+    } else if (block.type === 'paragraph' || block.type === 'blockquote') {
+      const remainingSpace = linesPerPage - currentLines;
+      const quotePrefix = block.type === 'blockquote' ? '> ' : '';
+      const contentWithoutPrefix = block.type === 'blockquote' 
+        ? block.content.replace(/^>\s?/gm, '') 
+        : block.content;
+      
+      if (remainingSpace >= block.estimatedLines) {
+        currentChunk += (currentChunk ? '\n\n' : '') + block.content;
+        currentLines += block.estimatedLines;
+      } else {
+        const sentences = contentWithoutPrefix.match(/[^.!?]+[.!?]+|[^.!?]+$/g) || [contentWithoutPrefix];
+        let isFirstSentence = true;
+        for (const sentence of sentences) {
+          const fullSentence = isFirstSentence && block.type === 'blockquote' 
+            ? quotePrefix + sentence.trim() 
+            : sentence.trim();
+          const sentenceLines = Math.ceil(fullSentence.length / charsPerLine) + 1;
+          if (currentLines + sentenceLines > linesPerPage && currentChunk.trim()) {
+            chunks.push(currentChunk.trim());
+            currentChunk = fullSentence;
+            currentLines = sentenceLines;
+          } else {
+            if (currentChunk && !isFirstSentence) {
+              currentChunk += ' ' + fullSentence;
+            } else if (currentChunk) {
+              currentChunk += '\n\n' + fullSentence;
+            } else {
+              currentChunk = fullSentence;
+            }
+            currentLines += sentenceLines;
+          }
+          isFirstSentence = false;
         }
       }
     } else {