fix(query): preserve CamelCase in code term extraction

Pass original query to _extract_code_terms instead of lowercased version

Author: DevanshuNEU

backend/routes/playground.py

@@ -421,8 +421,9 @@ async def playground_search(
     try:
         sanitized_query = InputValidator.sanitize_string(request.query, max_length=200)
 
-        # Check cache
-        cached_results = cache.get_search_results(sanitized_query, repo_id)
+        # Check cache (include flags in key to avoid returning wrong results)
+        cache_key = f"{sanitized_query}:v3={request.use_v3}:tests={request.include_tests}"
+        cached_results = cache.get_search_results(cache_key, repo_id)
         if cached_results:
             return {
                 "results": cached_results,
@@ -468,8 +469,8 @@ async def playground_search(
                 "is_test_file": r.get("is_test_file", False),  # V3 feature
             })
 
-        # Cache results
-        cache.set_search_results(sanitized_query, repo_id, results, ttl=3600)
+        # Cache results (using same key that includes flags)
+        cache.set_search_results(cache_key, repo_id, results, ttl=3600)
 
         search_time = int((time.time() - start_time) * 1000)
 

backend/scripts/benchmark_search_v3.py

@@ -232,8 +232,8 @@ async def run_benchmark(repo_id: str):
         else:
             print("⚠️ Voyage AI not enabled - using OpenAI embeddings")
             print("   Set VOYAGE_API_KEY for better code search accuracy!")
-    except:
-        pass
+    except Exception as e:
+        print(f"⚠️ Could not check Voyage status: {e}")
 
 
 if __name__ == "__main__":

backend/scripts/cross_repo_test.py

@@ -7,7 +7,11 @@
 import sys
 
 sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
-os.environ["VOYAGE_API_KEY"] = "pa-LPXoLbJ3W-S01F70zQsHcRFUTPXZ52dZ3d9PrXnxm7A"
+
+# Load from environment (set in .env or export manually)
+if not os.environ.get("VOYAGE_API_KEY"):
+    print("❌ VOYAGE_API_KEY not set. Export it or add to .env file.")
+    sys.exit(1)
 
 from services.indexer_optimized import OptimizedCodeIndexer
 

backend/scripts/edge_case_test.py

@@ -7,7 +7,11 @@
 import sys
 
 sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
-os.environ["VOYAGE_API_KEY"] = "pa-LPXoLbJ3W-S01F70zQsHcRFUTPXZ52dZ3d9PrXnxm7A"
+
+# Load from environment (set in .env or export manually)
+if not os.environ.get("VOYAGE_API_KEY"):
+    print("❌ VOYAGE_API_KEY not set. Export it or add to .env file.")
+    sys.exit(1)
 
 from services.indexer_optimized import OptimizedCodeIndexer
 

backend/scripts/extended_query_test.py

@@ -8,7 +8,11 @@
 import time
 
 sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
-os.environ["VOYAGE_API_KEY"] = "pa-LPXoLbJ3W-S01F70zQsHcRFUTPXZ52dZ3d9PrXnxm7A"
+
+# Load from environment (set in .env or export manually)
+if not os.environ.get("VOYAGE_API_KEY"):
+    print("❌ VOYAGE_API_KEY not set. Export it or add to .env file.")
+    sys.exit(1)
 
 from services.indexer_optimized import OptimizedCodeIndexer
 

backend/scripts/extended_v3_test.py

@@ -8,7 +8,11 @@
 import sys
 
 sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
-os.environ["VOYAGE_API_KEY"] = "pa-LPXoLbJ3W-S01F70zQsHcRFUTPXZ52dZ3d9PrXnxm7A"
+
+# Load from environment (set in .env or export manually)
+if not os.environ.get("VOYAGE_API_KEY"):
+    print("❌ VOYAGE_API_KEY not set. Export it or add to .env file.")
+    sys.exit(1)
 
 from services.indexer_optimized import OptimizedCodeIndexer
 
@@ -61,13 +65,15 @@ async def run_extended_tests():
         # V2
         try:
             v2_results = await indexer.search_v2(query, repo_id, top_k=3)
-        except:
+        except Exception as e:
+            print(f"  V2 error: {e}")
             v2_results = []
 
         # V3
         try:
             v3_results = await indexer.search_v3(query, repo_id, top_k=3, include_tests=False)
-        except:
+        except Exception as e:
+            print(f"  V3 error: {e}")
             v3_results = []
 
         # Check for test files in top 3

backend/scripts/final_v3_test.py

@@ -8,7 +8,11 @@
 import time
 
 sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
-os.environ["VOYAGE_API_KEY"] = "pa-LPXoLbJ3W-S01F70zQsHcRFUTPXZ52dZ3d9PrXnxm7A"
+
+# Load from environment (set in .env or export manually)
+if not os.environ.get("VOYAGE_API_KEY"):
+    print("❌ VOYAGE_API_KEY not set. Export it or add to .env file.")
+    sys.exit(1)
 
 from services.indexer_optimized import OptimizedCodeIndexer
 

backend/scripts/human_query_test.py

@@ -10,8 +10,10 @@
 
 sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
 
-# Set env vars before imports
-os.environ["VOYAGE_API_KEY"] = "pa-LPXoLbJ3W-S01F70zQsHcRFUTPXZ52dZ3d9PrXnxm7A"
+# Load from environment (set in .env or export manually)
+if not os.environ.get("VOYAGE_API_KEY"):
+    print("❌ VOYAGE_API_KEY not set. Export it or add to .env file.")
+    sys.exit(1)
 
 from services.indexer_optimized import OptimizedCodeIndexer
 

backend/services/indexer_optimized.py

@@ -547,7 +547,6 @@ async def embed_query(q: str) -> List[float]:
                 pinecone_index=self.index,
                 embedding_fn=embed_query,
             )
-            searcher.embed = embed_query
 
             results = await searcher.search(
                 query=query,

backend/services/search_v3/code_graph_ranker.py

@@ -91,8 +91,9 @@ def calculate_importance(
         # find max for normalization
         max_dependents = max(dependent_counts.values()) if dependent_counts else 1
 
-        # calculate importance for each file
-        for file_path in file_dependencies.keys():
+        # calculate importance for each file (include files that only appear as dependencies)
+        all_files = set(file_dependencies.keys()) | set(dependent_counts.keys())
+        for file_path in all_files:
             is_test = self._is_test_file(file_path)
             is_core = self._is_core_file(file_path)
             dep_count = dependent_counts.get(file_path, 0)