Fix: Update UI, Shards (16), and Router Logic

app.py

@@ -10,7 +10,7 @@ from src.data_pipeline import get_embedding
 # --- Configuration ---
 COLLECTION_NAME = "dashVector_v1"
 VECTOR_SIZE = 384 # MiniLM-L6-v2
-NUM_CLUSTERS = 32
+NUM_CLUSTERS = 16
 EMBEDDING_MODEL = "sentence-transformers/all-MiniLM-L6-v2"
 
 # --- Initialize Backend ---
@@ -90,14 +90,13 @@ NAVBAR_HTML = """
 <header class="bg-white border-b border-slate-200 sticky top-0 z-40 shadow-sm w-full">
     <div class="max-w-7xl mx-auto px-4 sm:px-6 lg:px-8 h-16 flex items-center justify-between">
         <div class="flex items-center gap-2">
-            <!-- User Logo -->
-            <img src="file/logo.png" alt="dashVector Logo" class="h-8 w-auto" />
+            <!-- User Logo Removed -->
             <h1 class="text-xl font-bold tracking-tight text-slate-900">dashVector</h1>
         </div>
         <div class="flex items-center gap-4">
             <div class="hidden md:flex items-center gap-1.5 px-3 py-1 bg-slate-100 rounded-full border border-slate-200">
                 <span class="material-symbols-outlined text-slate-500 text-sm">database</span>
-                <span class="text-xs font-medium text-slate-600">Dataset: <span class="font-bold text-slate-800">MS Marco</span></span>
+                <span class="text-xs font-medium text-slate-600">Dataset: <span class="font-bold text-slate-800">MS Marco (25k)</span></span>
             </div>
         </div>
     </div>
@@ -257,16 +256,18 @@ def run_benchmark(query):
 
         # Router Prediction
         if router:
-            print("DEBUG: Predicting cluster...")
-            target_cluster, confidence = router.predict(query_vec)
-            print(f"DEBUG: Predicted cluster {target_cluster} with confidence {confidence}")
+            print("DEBUG: Predicting clusters...")
+            # Now returns list of clusters and cumulative confidence
+            target_clusters, confidence = router.predict(query_vec)
+            print(f"DEBUG: Predicted clusters {target_clusters} with cumulative confidence {confidence}")
         else:
             print("DEBUG: No router loaded, using mock.")
-            target_cluster, confidence = 0, 0.95 # Mock
+            target_clusters, confidence = [0], 0.95 # Mock
             
         # Search
         print("DEBUG: Searching Qdrant...")
-        results, mode = vector_db.search_hybrid(query_vec, target_cluster, confidence)
+        # Now accepts list of clusters
+        results, mode = vector_db.search_hybrid(query_vec, target_clusters, confidence)
         print(f"DEBUG: Search complete. Found {len(results)} results.")
         
         end_total = time.time()
@@ -274,15 +275,17 @@ def run_benchmark(query):
         
         # Construct Data Rows
         
-        # Live Row (MiniLM + LightGBM)
-        # Mocking shards searched based on confidence for demo visual
-        shards_searched = 2 if confidence > 0.8 else 33
-        total_shards = 33
-        direct_time = latency_ms * (total_shards / shards_searched) * 1.2 # Estimate baseline
+        # Live Row (MiniLM + Logistic Regression)
+        shards_searched = len(target_clusters)
+        total_shards = 16 # Updated to 16
+        
+        # Estimate baseline time (mock calculation for demo if we don't run full scan)
+        # Or we could actually run full scan if we wanted true comparison, but for speed we estimate
+        direct_time = latency_ms * (total_shards / max(shards_searched, 1)) * 1.1 
         
         live_row = {
             "embedding": "MiniLM-L6-v2 (Active)",
-            "router": "LightGBM",
+            "router": "Logistic Regression", # Updated label
             "optimizedTime": f"{latency_ms:.1f} ms",
             "shardsSearched": f"{shards_searched} / {total_shards}",
             "totalShards": total_shards,
@@ -292,7 +295,7 @@ def run_benchmark(query):
             "efficiency": f"+{((1 - latency_ms/direct_time)*100):.1f}%"
         }
         
-        # Reference Rows (Static)
+        # Reference Rows (Static - Updated)
         ref_rows = [
             {
                 "embedding": "Gemma 300M",
@@ -307,7 +310,7 @@ def run_benchmark(query):
             },
             {
                 "embedding": "Qwen 600M",
-                "router": "XGBoost",
+                "router": "Tiny MLP",
                 "optimizedTime": "109 ms",
                 "shardsSearched": "7 / 16",
                 "totalShards": 16,