Patch/fix ingestion

py/core/main/api/data/ingestion_router_openapi.yml

@@ -67,7 +67,7 @@ ingest_chunks:
         source: |
           from r2r import R2RClient
 
-          client = R2RClient("http://localhost:7272") 
+          client = R2RClient("http://localhost:7272")
           # when using auth, do client.login(...)
 
           result = client.ingest_chunks(
@@ -76,12 +76,13 @@ ingest_chunks:
                     "text": "Another chunk of text",
                 },
                 {
-                    "text": "Yet another chunk of text", 
+                    "text": "Yet another chunk of text",
                 },
                 {
                     "text": "A chunk of text",
                 },
-            ],          )
+            ],
+          )
       - lang: Shell
         source: |
           curl -X POST "https://api.example.com/ingest_chunks" \
@@ -96,7 +97,7 @@ ingest_chunks:
                   "text": "Yet another chunk of text"
                 },
                 {
-                  "text": "A chunk of text" 
+                  "text": "A chunk of text"
                 }
               ],
               "document_id": "b4ac4dd6-5f27-596e-a55b-7cf242ca30aa",
@@ -106,4 +107,4 @@ ingest_chunks:
   input_descriptions:
     chunks: "A list of text chunks to ingest into the system."
     document_id: "An optional document id to associate the chunks with. If not provided, a unique document id will be generated."
-    metadata: "Optional JSON metadata to associate with the ingested chunks."
+    metadata: "Optional JSON metadata to associate with the ingested chunks."

py/core/main/api/retrieval_router.py

@@ -283,12 +283,16 @@ async def stream_generator():
             except Exception as e:
                 raise R2RException(str(e), 500)
 
-
         @self.router.post("/completion")
         @self.base_endpoint
         async def completion(
-            messages: list[Message] = Body(..., description="The messages to complete"),
-            generation_config: GenerationConfig = Body(default_factory=GenerationConfig, description="The generation config"),
+            messages: list[Message] = Body(
+                ..., description="The messages to complete"
+            ),
+            generation_config: GenerationConfig = Body(
+                default_factory=GenerationConfig,
+                description="The generation config",
+            ),
             auth_user=Depends(self.service.providers.auth.auth_wrapper),
             response_model=WrappedCompletionResponse,
         ):

py/core/main/orchestration/simple/ingestion_workflow.py

@@ -157,11 +157,7 @@ async def update_files(input_data):
                 "metadata": updated_metadata,
                 "document_id": str(doc_id),
                 "version": new_version,
-                "ingestion_config": (
-                    ingestion_config.model_dump_json()
-                    if ingestion_config
-                    else None
-                ),
+                "ingestion_config": ingestion_config,
                 "size_in_bytes": file_size_in_bytes,
                 "is_update": True,
             }

py/core/main/services/retrieval_service.py

@@ -118,7 +118,7 @@ async def search(
             )
 
             return results.as_dict()
-        
+
     @telemetry_event("Completion")
     async def completion(
         self,

py/pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "poetry.core.masonry.api"
 [tool.poetry]
 name = "r2r"
 readme = "README.md"
-version = "3.2.3"
+version = "3.2.4"
 
 description = "SciPhi R2R"
 authors = ["Owen Colegrove <[email protected]>"]

py/sdk/retrieval.py

@@ -45,7 +45,6 @@ async def search(
         }
         return await client._make_request("POST", "search", json=data)
 
-
     @staticmethod
     async def completion(
         client,
@@ -64,7 +63,7 @@ async def completion(
             "messages": [msg.model_dump() for msg in cast_messages],
             "generation_config": generation_config,
         }
-        
+
         return await client._make_request("POST", "completion", json=data)
 
     @staticmethod

py/shared/api/models/retrieval/responses.py

@@ -6,6 +6,7 @@
 from shared.abstractions.llm import LLMChatCompletion
 from shared.api.models.base import ResultsWrapper
 
+
 class SearchResponse(BaseModel):
     vector_search_results: list[VectorSearchResult] = Field(
         ...,
@@ -138,8 +139,9 @@ class Config:
             ]
         }
 
+
 WrappedCompletionResponse = ResultsWrapper[LLMChatCompletion]
 # Create wrapped versions of the responses
 WrappedSearchResponse = ResultsWrapper[SearchResponse]
 WrappedRAGResponse = ResultsWrapper[RAGResponse]
-WrappedRAGAgentResponse = ResultsWrapper[RAGAgentResponse]
+WrappedRAGAgentResponse = ResultsWrapper[RAGAgentResponse]