Add conversation_history object to query api

.semversioner/next-release/minor-20250114014943106649.json

@@ -0,0 +1,4 @@
+{
+  "type": "minor",
+  "description": "add conversation_history to api"
+}

graphrag/api/query.py

@@ -37,6 +37,9 @@
     get_global_search_engine,
     get_local_search_engine,
 )
+from graphrag.query.context_builder.conversation_history import (
+    ConversationHistory,
+)
 from graphrag.query.indexer_adapters import (
     read_indexer_communities,
     read_indexer_covariates,
@@ -68,6 +71,7 @@ async def global_search(
     dynamic_community_selection: bool,
     response_type: str,
     query: str,
+    conversation_history: ConversationHistory | None = None,
 ) -> tuple[
     str | dict[str, Any] | list[dict[str, Any]],
     str | list[pd.DataFrame] | dict[str, pd.DataFrame],
@@ -121,7 +125,7 @@ async def global_search(
         reduce_system_prompt=reduce_prompt,
         general_knowledge_inclusion_prompt=knowledge_prompt,
     )
-    result: SearchResult = await search_engine.asearch(query=query)
+    result: SearchResult = await search_engine.asearch(query=query, conversation_history=conversation_history)
     response = result.response
     context_data = _reformat_context_data(result.context_data)  # type: ignore
     return response, context_data
@@ -138,6 +142,7 @@ async def global_search_streaming(
     dynamic_community_selection: bool,
     response_type: str,
     query: str,
+    conversation_history: ConversationHistory | None = None,
 ) -> AsyncGenerator:
     """Perform a global search and return the context data and response via a generator.
 
@@ -190,7 +195,7 @@ async def global_search_streaming(
         reduce_system_prompt=reduce_prompt,
         general_knowledge_inclusion_prompt=knowledge_prompt,
     )
-    search_result = search_engine.astream_search(query=query)
+    search_result = search_engine.astream_search(query=query, conversation_history=conversation_history)
 
     # when streaming results, a context data object is returned as the first result
     # and the query response in subsequent tokens
@@ -217,6 +222,7 @@ async def local_search(
     community_level: int,
     response_type: str,
     query: str,
+    conversation_history: ConversationHistory | None = None,
 ) -> tuple[
     str | dict[str, Any] | list[dict[str, Any]],
     str | list[pd.DataFrame] | dict[str, pd.DataFrame],
@@ -268,7 +274,7 @@ async def local_search(
         system_prompt=prompt,
     )
 
-    result: SearchResult = await search_engine.asearch(query=query)
+    result: SearchResult = await search_engine.asearch(query=query, conversation_history=conversation_history)
     response = result.response
     context_data = _reformat_context_data(result.context_data)  # type: ignore
     return response, context_data
@@ -286,6 +292,7 @@ async def local_search_streaming(
     community_level: int,
     response_type: str,
     query: str,
+    conversation_history: ConversationHistory | None = None,
 ) -> AsyncGenerator:
     """Perform a local search and return the context data and response via a generator.
 
@@ -333,7 +340,7 @@ async def local_search_streaming(
         response_type=response_type,
         system_prompt=prompt,
     )
-    search_result = search_engine.astream_search(query=query)
+    search_result = search_engine.astream_search(query=query, conversation_history=conversation_history)
 
     # when streaming results, a context data object is returned as the first result
     # and the query response in subsequent tokens
@@ -358,6 +365,7 @@ async def drift_search(
     relationships: pd.DataFrame,
     community_level: int,
     query: str,
+    conversation_history: ConversationHistory | None = None,
 ) -> tuple[
     str | dict[str, Any] | list[dict[str, Any]],
     str | list[pd.DataFrame] | dict[str, pd.DataFrame],
@@ -410,7 +418,7 @@ async def drift_search(
         local_system_prompt=prompt,
     )
 
-    result: SearchResult = await search_engine.asearch(query=query)
+    result: SearchResult = await search_engine.asearch(query=query, conversation_history=conversation_history)
     response = result.response
     context_data = _reformat_context_data(result.context_data)  # type: ignore
 
@@ -430,6 +438,7 @@ async def basic_search(
     config: GraphRagConfig,
     text_units: pd.DataFrame,
     query: str,
+    conversation_history: ConversationHistory | None = None,
 ) -> tuple[
     str | dict[str, Any] | list[dict[str, Any]],
     str | list[pd.DataFrame] | dict[str, pd.DataFrame],
@@ -468,7 +477,7 @@ async def basic_search(
         system_prompt=prompt,
     )
 
-    result: SearchResult = await search_engine.asearch(query=query)
+    result: SearchResult = await search_engine.asearch(query=query, conversation_history=conversation_history)
     response = result.response
     context_data = _reformat_context_data(result.context_data)  # type: ignore
     return response, context_data
@@ -479,6 +488,7 @@ async def basic_search_streaming(
     config: GraphRagConfig,
     text_units: pd.DataFrame,
     query: str,
+    conversation_history: ConversationHistory | None = None,
 ) -> AsyncGenerator:
     """Perform a local search and return the context data and response via a generator.
 
@@ -513,7 +523,7 @@ async def basic_search_streaming(
         system_prompt=prompt,
     )
 
-    search_result = search_engine.astream_search(query=query)
+    search_result = search_engine.astream_search(query=query, conversation_history=conversation_history)
 
     # when streaming results, a context data object is returned as the first result
     # and the query response in subsequent tokens