Update app.py

app.py

@@ -13,15 +13,31 @@ with open("Untitled document.txt", "r", encoding="utf-8") as f:
 
 # Step 2: Preprocess text into sentence chunks
 def preprocess_text(text):
-    
     cleaned_text = text.strip()
-    chunks = cleaned_text.split(".")
-    combined_chunks = []
+    sentences = [s.strip() for s in cleaned_text.split('.') if s.strip()]
+    sentence_chunks = [s.strip() for s in sentences if len(s.strip()) > 10]
     
+    combined_chunks = []
     for i in range(0, len(sentences), 2):
         chunk = '. '.join(sentences[i:i+3]).strip()
-        
+        if len(chunk) > 20:
+            combined_chunks.append(chunk)
+    
     paragraphs = [p.strip() for p in cleaned_text.split('\n\n') if p.strip()]
+    paragraph_chunks = [p for p in paragraphs if len(p) > 30]
+    
+    all_chunks = sentence_chunks + combined_chunks + paragraph_chunks
+    
+    seen = set()
+    final_chunks = []
+    for chunk in all_chunks:
+        if chunk not in seen and len(chunk) > 15:
+            seen.add(chunk)
+            final_chunks.append(chunk)
+    
+    print(f"Created {len(final_chunks)} chunks using advanced strategy")
+    print(f"Sample chunks: {final_chunks[:3]}")
+    return final_chunks
 
 cleaned_chunks = preprocess_text(skincare_text)