🐛 Modify the knowledge base front-end rendering #1458#1456

🐛 Modify the knowledge base front-end rendering #1458#1456

Author: liutao12138

backend/utils/document_vector_utils.py

@@ -237,10 +237,18 @@ def kmeans_cluster_documents(doc_embeddings: Dict[str, np.ndarray], k: Optional[
         doc_ids = list(doc_embeddings.keys())
         embeddings_array = np.array([doc_embeddings[doc_id] for doc_id in doc_ids])
 
+        # Handle single document case
+        if len(doc_ids) == 1:
+            logger.info("Only one document found, skipping clustering")
+            return {0: doc_ids}
+        
         # Determine K value
         if k is None:
             k = auto_determine_k(embeddings_array)
 
+        # Ensure k is not greater than number of documents
+        k = min(k, len(doc_ids))
+        
         logger.info(f"Clustering {len(doc_ids)} documents into {k} clusters")
 
         # Perform K-means clustering

frontend/components/ui/markdownRenderer.tsx

@@ -437,6 +437,15 @@ export const MarkdownRenderer: React.FC<MarkdownRendererProps> = ({
             const code = mmd[1];
             return <Diagram key={`mmd-${index}`} code={code} className="my-4" />;
           }
+          // Handle line breaks in text content
+          if (part.includes('\n')) {
+            return part.split('\n').map((line, lineIndex) => (
+              <React.Fragment key={`${index}-${lineIndex}`}>
+                {line}
+                {lineIndex < part.split('\n').length - 1 && <br />}
+              </React.Fragment>
+            ));
+          }
           return part;
         })}
       </>

test/backend/test_document_vector_utils.py

@@ -145,6 +145,19 @@ def test_kmeans_cluster_documents_empty(self):
         clusters = kmeans_cluster_documents(doc_embeddings)
 
         assert clusters == {}
+    
+    def test_kmeans_cluster_documents_single(self):
+        """Test handling of single document"""
+        doc_embeddings = {
+            'doc1': np.array([1.0, 1.0, 1.0])
+        }
+        clusters = kmeans_cluster_documents(doc_embeddings)
+        
+        # Should return single cluster with one document
+        assert len(clusters) == 1
+        assert 0 in clusters
+        assert len(clusters[0]) == 1
+        assert clusters[0][0] == 'doc1'
 
 
 class TestExtractRepresentativeChunksSmart: