fix(api): handle blank search text

Signed-off-by: phernandez <paul@basicmachines.co>

Author: phernandez

src/basic_memory/api/v2/routers/search_router.py

@@ -4,6 +4,8 @@
 V1 uses string-based project names which are less efficient and less stable.
 """
 
+import asyncio
+
 from fastapi import APIRouter, HTTPException, Path
 
 import logfire
@@ -74,8 +76,10 @@ async def search(
                 page=page,
                 page_size=page_size,
             ):
-                results = await search_service.search(query, limit=page_size, offset=offset)
-                total = await search_service.count(query)
+                results, total = await asyncio.gather(
+                    search_service.search(query, limit=page_size, offset=offset),
+                    search_service.count(query),
+                )
         except SemanticSearchDisabledError as exc:
             raise HTTPException(status_code=400, detail=str(exc)) from exc
         except SemanticDependenciesMissingError as exc:

src/basic_memory/repository/postgres_search_repository.py

@@ -1012,12 +1012,7 @@ async def count(
         min_similarity: Optional[float] = None,
     ) -> int:
         """Count indexed content matching the Postgres FTS query."""
-        mode = (
-            retrieval_mode.value
-            if isinstance(retrieval_mode, SearchRetrievalMode)
-            else str(retrieval_mode)
-        )
-        if mode != SearchRetrievalMode.FTS.value:
+        if retrieval_mode != SearchRetrievalMode.FTS:
             return await super().count(
                 search_text=search_text,
                 permalink=permalink,

src/basic_memory/repository/search_repository_base.py

@@ -261,6 +261,9 @@ async def count(
         min_similarity: Optional[float] = None,
     ) -> int:
         """Count results for retrieval modes that cannot use a backend COUNT query."""
+        # Trigger: vector and hybrid modes rank after embedding lookup, filtering, and fusion.
+        # Why: that scoring pipeline is not expressible as a portable database COUNT query.
+        # Outcome: fetch the bounded candidate set and count the final in-memory results.
         results = await self.search(
             search_text=search_text,
             permalink=permalink,

src/basic_memory/repository/sqlite_search_repository.py

@@ -1031,12 +1031,7 @@ async def count(
         min_similarity: Optional[float] = None,
     ) -> int:
         """Count indexed content matching the SQLite FTS query."""
-        mode = (
-            retrieval_mode.value
-            if isinstance(retrieval_mode, SearchRetrievalMode)
-            else str(retrieval_mode)
-        )
-        if mode != SearchRetrievalMode.FTS.value:
+        if retrieval_mode != SearchRetrievalMode.FTS:
             return await super().count(
                 search_text=search_text,
                 permalink=permalink,

src/basic_memory/services/search_service.py

@@ -155,10 +155,10 @@ def _prepare_query(self, query: SearchQuery) -> _PreparedSearchQuery | None:
         tags = query.tags
 
         # Support tag:<tag> shorthand by mapping to tags filter.
-        if search_text:
-            text = search_text.strip()
-            if text.lower().startswith("tag:"):
-                tag_values = re.split(r"[,\s]+", text[4:].strip())
+        if search_text is not None:
+            search_text = search_text.strip() or None
+            if search_text and search_text.lower().startswith("tag:"):
+                tag_values = re.split(r"[,\s]+", search_text[4:].strip())
                 parsed_tags = [t for t in tag_values if t]
                 if parsed_tags:
                     tags = parsed_tags

tests/api/v2/test_search_router.py

@@ -57,7 +57,7 @@ async def test_search_entities(
     )
 
     # Search for the entity
-    response = await client.post(f"{v2_project_url}/search/", json={"search_text": "Searchable"})
+    response = await client.post(f"{v2_project_url}/search/", json={"text": "Searchable"})
 
     assert response.status_code == 200
     data = response.json()
@@ -243,7 +243,7 @@ async def test_search_with_type_filter(
 
     # Search with type filter
     response = await client.post(
-        f"{v2_project_url}/search/", json={"search_text": "Type", "note_types": ["note"]}
+        f"{v2_project_url}/search/", json={"text": "Type", "note_types": ["note"]}
     )
 
     assert response.status_code == 200
@@ -276,7 +276,7 @@ async def test_search_with_date_filter(
     # Search with date filter
     response = await client.post(
         f"{v2_project_url}/search/",
-        json={"search_text": "Date Filtered", "after_date": "2024-01-01T00:00:00Z"},
+        json={"text": "Date Filtered", "after_date": "2024-01-01T00:00:00Z"},
     )
 
     assert response.status_code == 200
@@ -297,12 +297,42 @@ async def test_search_empty_query(
     assert response.status_code in [200, 422]
 
 
+@pytest.mark.asyncio
+async def test_search_whitespace_text_is_treated_as_empty(
+    client: AsyncClient,
+    test_project: Project,
+    v2_project_url: str,
+    entity_repository,
+    search_service,
+    file_service,
+):
+    """Whitespace-only text should not become an unfiltered project-wide search."""
+    entity_data = {
+        "title": "Whitespace Regression Entity",
+        "note_type": "note",
+        "content_type": "text/markdown",
+        "file_path": "whitespace_regression.md",
+        "checksum": "whitespace123",
+    }
+    await create_test_entity(
+        test_project, entity_data, entity_repository, search_service, file_service
+    )
+
+    response = await client.post(f"{v2_project_url}/search/", json={"text": "   "})
+
+    assert response.status_code == 200
+    data = response.json()
+    assert data["total"] == 0
+    assert data["has_more"] is False
+    assert data["results"] == []
+
+
 @pytest.mark.asyncio
 async def test_search_invalid_project_id(
     client: AsyncClient,
 ):
     """Test searching with invalid project ID returns 404."""
-    response = await client.post("/v2/projects/999999/search/", json={"search_text": "test"})
+    response = await client.post("/v2/projects/999999/search/", json={"text": "test"})
 
     assert response.status_code == 404
 
@@ -342,7 +372,7 @@ async def test_v2_search_endpoints_use_project_id_not_name(
 ):
     """Test that v2 search endpoints reject string project names."""
     # Try to use project name instead of ID - should fail
-    response = await client.post(f"/v2/{test_project.name}/search/", json={"search_text": "test"})
+    response = await client.post(f"/v2/{test_project.name}/search/", json={"text": "test"})
 
     # FastAPI path validation should reject non-integer project_id
     assert response.status_code in [404, 422]
@@ -358,11 +388,14 @@ class RaisingSearchService:
         async def search(self, *args, **kwargs):
             raise SemanticSearchDisabledError("Semantic search is disabled for this project.")
 
+        async def count(self, *args, **kwargs):
+            raise SemanticSearchDisabledError("Semantic search is disabled for this project.")
+
     app.dependency_overrides[get_search_service_v2_external] = lambda: RaisingSearchService()
     try:
         response = await client.post(
             f"{v2_project_url}/search/",
-            json={"search_text": "semantic query", "retrieval_mode": "vector"},
+            json={"text": "semantic query", "retrieval_mode": "vector"},
         )
     finally:
         app.dependency_overrides.pop(get_search_service_v2_external, None)
@@ -381,11 +414,14 @@ class RaisingSearchService:
         async def search(self, *args, **kwargs):
             raise SemanticDependenciesMissingError("Semantic dependencies are missing.")
 
+        async def count(self, *args, **kwargs):
+            raise SemanticDependenciesMissingError("Semantic dependencies are missing.")
+
     app.dependency_overrides[get_search_service_v2_external] = lambda: RaisingSearchService()
     try:
         response = await client.post(
             f"{v2_project_url}/search/",
-            json={"search_text": "semantic query", "retrieval_mode": "hybrid"},
+            json={"text": "semantic query", "retrieval_mode": "hybrid"},
         )
     finally:
         app.dependency_overrides.pop(get_search_service_v2_external, None)
@@ -404,6 +440,9 @@ class RaisingSearchService:
         async def search(self, *args, **kwargs):
             raise ValueError("Vector retrieval requires a text query.")
 
+        async def count(self, *args, **kwargs):
+            raise ValueError("Vector retrieval requires a text query.")
+
     app.dependency_overrides[get_search_service_v2_external] = lambda: RaisingSearchService()
     try:
         response = await client.post(
@@ -517,7 +556,7 @@ async def count(self, *args, **kwargs):
     try:
         response = await client.post(
             f"{v2_project_url}/search/",
-            json={"search_text": "pricing"},
+            json={"text": "pricing"},
         )
     finally:
         app.dependency_overrides.pop(get_search_service_v2_external, None)
@@ -561,7 +600,7 @@ async def count(self, *args, **kwargs):
     try:
         response = await client.post(
             f"{v2_project_url}/search/",
-            json={"search_text": "general"},
+            json={"text": "general"},
         )
     finally:
         app.dependency_overrides.pop(get_search_service_v2_external, None)