Update Unstructured Provider, Fix Type Error in SDK (#2103)

* Fix SDK type error

* Update dashboard tag

Author: NolanTrem

docker/compose.full.swarm.yaml

@@ -386,7 +386,7 @@ services:
         tag: backend
 
   r2r-dashboard:
-    image: sciphiai/r2r-dashboard:1.0.2
+    image: sciphiai/r2r-dashboard:1.0.3
     environment:
       - NEXT_PUBLIC_R2R_DEPLOYMENT_URL=${R2R_DEPLOYMENT_URL:-http://localhost:7272}
       - NEXT_PUBLIC_HATCHET_DASHBOARD_URL=${HATCHET_DASHBOARD_URL:-http://localhost:${R2R_HATCHET_DASHBOARD_PORT:-7274}}

docker/compose.full.yaml

@@ -193,7 +193,7 @@ services:
         tag: backend
 
   r2r-dashboard:
-    image: sciphiai/r2r-dashboard:1.0.2
+    image: sciphiai/r2r-dashboard:1.0.3
     env_file:
       - ./env/r2r-dashboard.env
     ports:

docker/compose.yaml

@@ -50,7 +50,7 @@ services:
       - host.docker.internal:host-gateway
 
   r2r-dashboard:
-    image: sciphiai/r2r-dashboard:1.0.2
+    image: sciphiai/r2r-dashboard:1.0.3
     env_file:
       - ./env/r2r-dashboard.env
     ports:

py/core/providers/ingestion/r2r/base.py

@@ -259,15 +259,6 @@ async def parse(
                 logger.info(
                     f"Using parser_override for {document.document_type} with input value {parser_overrides[document.document_type.value]}"
                 )
-                # TODO - Cleanup this approach to be less hardcoded
-                # if (
-                #     document.document_type != DocumentType.PDF
-                #     or parser_overrides[DocumentType.PDF.value] != "zerox"
-                # ):
-                #     raise ValueError(
-                #         "Only Zerox PDF parser override is available."
-                #     )
-
                 if parser_overrides[DocumentType.PDF.value] == "zerox":
                     # Collect content from VLMPDFParser
                     async for chunk in self.parsers[

py/core/providers/ingestion/unstructured/base.py

@@ -290,12 +290,26 @@ async def parse(
             logger.info(
                 f"Using parser_override for {document.document_type} with input value {parser_overrides[document.document_type.value]}"
             )
-            async for element in self.parse_fallback(
-                file_content,
-                ingestion_config=ingestion_config,
-                parser_name=f"zerox_{DocumentType.PDF.value}",
-            ):
-                elements.append(element)
+            if parser_overrides[document.document_type.value] == "zerox":
+                async for element in self.parse_fallback(
+                    file_content,
+                    ingestion_config=ingestion_config,
+                    parser_name=f"zerox_{DocumentType.PDF.value}",
+                ):
+                    logger.warning(
+                        f"Using parser_override for {document.document_type}"
+                    )
+                    elements.append(element)
+            elif parser_overrides[document.document_type.value] == "ocr":
+                async for element in self.parse_fallback(
+                    file_content,
+                    ingestion_config=ingestion_config,
+                    parser_name=f"ocr_{DocumentType.PDF.value}",
+                ):
+                    logger.warning(
+                        f"Using OCR parser_override for {document.document_type}"
+                    )
+                    elements.append(element)
 
         elif document.document_type in self.R2R_FALLBACK_PARSERS.keys():
             logger.info(

py/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "r2r"
-version = "3.5.7"
+version = "3.5.8"
 description = "SciPhi R2R"
 readme = "README.md"
 license = {text = "MIT"}

py/sdk/sync_methods/documents.py

@@ -62,7 +62,7 @@ def create(
             dict | IngestionMode
         ] = None,  # Type hint clarifies IngestionMode enum usage
         run_with_orchestration: Optional[bool] = True,
-        ingestion_mode: IngestionMode = IngestionMode.custom,  # Added enum parameter as per router
+        ingestion_mode: IngestionMode | str = IngestionMode.custom,
     ) -> WrappedIngestionResponse:
         """Create a new document from either a file, raw text, or chunks.