new hybrid document search

py/core/base/providers/database.py

@@ -521,11 +521,11 @@ async def full_text_search(
     ) -> list[VectorSearchResult]:
         pass
 
-    @abstractmethod
-    async def search_documents(
-        self, query_text: str, settings: DocumentSearchSettings
-    ) -> list[dict]:
-        pass
+    # @abstractmethod
+    # async def search_documents(
+    #     self, query_text: str, settings: DocumentSearchSettings
+    # ) -> list[dict]:
+    #     pass
 
     @abstractmethod
     async def hybrid_search(
@@ -1430,9 +1430,14 @@ async def hybrid_search(
         )
 
     async def search_documents(
-        self, query_text: str, settings: DocumentSearchSettings
+        self,
+        query_text: str,
+        settings: DocumentSearchSettings,
+        query_embedding: Optional[list[float]] = None,
     ) -> list[dict]:
-        return await self.vector_handler.search_documents(query_text, settings)
+        return await self.document_handler.search_documents(
+            query_text, query_embedding, settings
+        )
 
     async def delete(
         self, filters: dict[str, Any]

py/core/main/api/retrieval_router.py

@@ -127,8 +127,14 @@ async def search_documents(
             Allowed operators include `eq`, `neq`, `gt`, `gte`, `lt`, `lte`, `like`, `ilike`, `in`, and `nin`.
             """
 
+            query_embedding = (
+                await self.service.providers.embedding.async_get_embedding(
+                    query
+                )
+            )
             results = await self.service.search_documents(
                 query=query,
+                query_embedding=query_embedding,
                 settings=settings,
             )
             return results

py/core/main/services/ingestion_service.py

@@ -252,7 +252,7 @@ async def augment_document_info(
             document_info.summary = response.choices[0].message.content
 
             embedding = await self.providers.embedding.async_get_embedding(
-                text = document_info.summary,
+                text=document_info.summary,
             )
             document_info.summary_embedding = embedding
         return

py/core/main/services/retrieval_service.py

@@ -3,6 +3,7 @@
 import time
 from typing import Optional
 from uuid import UUID
+
 from fastapi import HTTPException
 
 from core import R2RStreamingRAGAgent
@@ -122,11 +123,13 @@ async def search_documents(
         self,
         query: str,
         settings: DocumentSearchSettings,
+        query_embedding: Optional[list[float]] = None,
     ) -> list[dict]:
 
         return await self.providers.database.search_documents(
             query_text=query,
             settings=settings,
+            query_embedding=query_embedding,
         )
 
     @telemetry_event("Completion")

py/core/pipes/retrieval/vector_search_pipe.py

@@ -59,7 +59,7 @@ async def search(  # type: ignore
             message,
             purpose=EmbeddingPurpose.QUERY,
         )
-        
+
         search_results = await (
             self.database_provider.hybrid_search(
                 query_vector=query_vector,