Simplified the preprocessing interface.

llama_stack/apis/preprocessing/preprocessing.py

@@ -70,19 +70,8 @@ class PreprocessorStore(Protocol):
 class Preprocessing(Protocol):
     preprocessor_store: PreprocessorStore
 
-    input_types: List[PreprocessingDataType]
-    output_types: List[PreprocessingDataType]
-
     @webmethod(route="/preprocess", method="POST")
     async def preprocess(
-        self,
-        preprocessor_id: str,
-        preprocessor_inputs: List[PreprocessingDataElement],
-        options: Optional[PreprocessorOptions] = None,
-    ) -> PreprocessorResponse: ...
-
-    @webmethod(route="/chain_preprocess", method="POST")
-    async def chain_preprocess(
         self,
         preprocessors: PreprocessorChain,
         preprocessor_inputs: List[PreprocessingDataElement],

llama_stack/distribution/routers/routers.py

@@ -45,7 +45,6 @@ from llama_stack.apis.preprocessing import (
     Preprocessing,
     PreprocessingDataElement,
     PreprocessorChain,
-    PreprocessorOptions,
     PreprocessorResponse,
 )
 from llama_stack.apis.safety import RunShieldResponse, Safety
@@ -714,22 +713,6 @@ class PreprocessingRouter(Preprocessing):
         pass
 
     async def preprocess(
-        self,
-        preprocessor_id: str,
-        preprocessor_inputs: List[PreprocessingDataElement],
-        options: Optional[PreprocessorOptions] = None,
-    ) -> PreprocessorResponse:
-        logcat.debug(
-            "core",
-            f"PreprocessingRouter.preprocess: {preprocessor_id}, {len(preprocessor_inputs)} inputs, options={options}",
-        )
-        return await self.routing_table.get_provider_impl(preprocessor_id).preprocess(
-            preprocessor_id=preprocessor_id,
-            preprocessor_inputs=preprocessor_inputs,
-            options=options,
-        )
-
-    async def chain_preprocess(
         self,
         preprocessors: PreprocessorChain,
         preprocessor_inputs: List[PreprocessingDataElement],

llama_stack/distribution/utils/chain.py

@@ -8,16 +8,16 @@ from itertools import pairwise
 from typing import List
 
 from llama_stack.apis.preprocessing import (
-    Preprocessing,
     PreprocessingDataElement,
     PreprocessorChain,
     PreprocessorResponse,
 )
+from llama_stack.providers.datatypes import PreprocessorsProtocolPrivate
 
 log = logging.getLogger(__name__)
 
 
-def validate_chain(chain_impls: List[Preprocessing]) -> bool:
+def validate_chain(chain_impls: List[PreprocessorsProtocolPrivate]) -> bool:
     if len(chain_impls) == 0:
         log.error("Empty preprocessing chain was provided")
         return False
@@ -37,7 +37,7 @@ def validate_chain(chain_impls: List[Preprocessing]) -> bool:
 
 async def execute_preprocessor_chain(
     preprocessor_chain: PreprocessorChain,
-    preprocessor_chain_impls: List[Preprocessing],
+    preprocessor_chain_impls: List[PreprocessorsProtocolPrivate],
     preprocessor_inputs: List[PreprocessingDataElement],
 ) -> PreprocessorResponse:
     if not validate_chain(preprocessor_chain_impls):
@@ -50,7 +50,7 @@ async def execute_preprocessor_chain(
     # TODO: replace with a parallel implementation
     for i, current_params in enumerate(preprocessor_chain):
         current_impl = preprocessor_chain_impls[i]
-        response = await current_impl.preprocess(
+        response = await current_impl.do_preprocess(
             preprocessor_id=current_params.preprocessor_id,
             preprocessor_inputs=current_inputs,
             options=current_params.options,

llama_stack/providers/datatypes.py

@@ -13,7 +13,13 @@ from llama_stack.apis.benchmarks import Benchmark
 from llama_stack.apis.datasets import Dataset
 from llama_stack.apis.datatypes import Api
 from llama_stack.apis.models import Model
-from llama_stack.apis.preprocessing import Preprocessor
+from llama_stack.apis.preprocessing import (
+    PreprocessingDataElement,
+    PreprocessingDataType,
+    Preprocessor,
+    PreprocessorOptions,
+    PreprocessorResponse,
+)
 from llama_stack.apis.scoring_functions import ScoringFn
 from llama_stack.apis.shields import Shield
 from llama_stack.apis.tools import Tool
@@ -60,10 +66,20 @@ class ToolsProtocolPrivate(Protocol):
 
 
 class PreprocessorsProtocolPrivate(Protocol):
+    input_types: List[PreprocessingDataType]
+    output_types: List[PreprocessingDataType]
+
     async def register_preprocessor(self, preprocessor: Preprocessor) -> None: ...
 
     async def unregister_preprocessor(self, preprocessor_id: str) -> None: ...
 
+    async def do_preprocess(
+        self,
+        preprocessor_id: str,
+        preprocessor_inputs: List[PreprocessingDataElement],
+        options: Optional[PreprocessorOptions] = None,
+    ) -> PreprocessorResponse: ...
+
 
 @json_schema_type
 class ProviderSpec(BaseModel):

llama_stack/providers/inline/preprocessing/basic/basic.py

@@ -52,7 +52,7 @@ class InclineBasicPreprocessorImpl(Preprocessing, PreprocessorsProtocolPrivate):
 
     async def unregister_preprocessor(self, preprocessor_id: str) -> None: ...
 
-    async def preprocess(
+    async def do_preprocess(
         self,
         preprocessor_id: str,
         preprocessor_inputs: List[PreprocessingDataElement],
@@ -98,12 +98,12 @@ class InclineBasicPreprocessorImpl(Preprocessing, PreprocessorsProtocolPrivate):
             success=True, output_data_type=PreprocessingDataType.raw_text_document, results=results
         )
 
-    async def chain_preprocess(
+    async def preprocess(
         self,
         preprocessors: PreprocessorChain,
         preprocessor_inputs: List[PreprocessingDataElement],
     ) -> PreprocessorResponse:
-        return await self.preprocess(preprocessor_id="", preprocessor_inputs=preprocessor_inputs)
+        return await self.do_preprocess(preprocessor_id="", preprocessor_inputs=preprocessor_inputs)
 
     @staticmethod
     def _resolve_input_type(preprocessor_input: PreprocessingDataElement) -> PreprocessingDataType:

llama_stack/providers/inline/preprocessing/docling/docling.py

@@ -47,7 +47,7 @@ class InclineDoclingPreprocessorImpl(Preprocessing, PreprocessorsProtocolPrivate
 
     async def unregister_preprocessor(self, preprocessor_id: str) -> None: ...
 
-    async def preprocess(
+    async def do_preprocess(
         self,
         preprocessor_id: str,
         preprocessor_inputs: List[PreprocessingDataElement],
@@ -106,9 +106,9 @@ class InclineDoclingPreprocessorImpl(Preprocessing, PreprocessorsProtocolPrivate
         )
         return PreprocessorResponse(success=True, output_data_type=output_data_type, results=results)
 
-    async def chain_preprocess(
+    async def preprocess(
         self,
         preprocessors: PreprocessorChain,
         preprocessor_inputs: List[PreprocessingDataElement],
     ) -> PreprocessorResponse:
-        return await self.preprocess(preprocessor_id="", preprocessor_inputs=preprocessor_inputs)
+        return await self.do_preprocess(preprocessor_id="", preprocessor_inputs=preprocessor_inputs)

llama_stack/providers/inline/preprocessing/simple_chunking/simple_chunking.py

@@ -47,7 +47,7 @@ class InclineSimpleChunkingImpl(Preprocessing, PreprocessorsProtocolPrivate):
 
     async def unregister_preprocessor(self, preprocessor_id: str) -> None: ...
 
-    async def preprocess(
+    async def do_preprocess(
         self,
         preprocessor_id: str,
         preprocessor_inputs: List[PreprocessingDataElement],
@@ -72,12 +72,12 @@ class InclineSimpleChunkingImpl(Preprocessing, PreprocessorsProtocolPrivate):
 
         return PreprocessorResponse(success=True, output_data_type=PreprocessingDataType.chunks, results=chunks)
 
-    async def chain_preprocess(
+    async def preprocess(
         self,
         preprocessors: PreprocessorChain,
         preprocessor_inputs: List[PreprocessingDataElement],
     ) -> PreprocessorResponse:
-        return await self.preprocess(preprocessor_id="", preprocessor_inputs=preprocessor_inputs)
+        return await self.do_preprocess(preprocessor_id="", preprocessor_inputs=preprocessor_inputs)
 
     def _resolve_chunk_size_params(self, options: PreprocessorOptions) -> Tuple[int, int]:
         window_len = (options or {}).get(

llama_stack/providers/inline/tool_runtime/rag/memory.py

@@ -81,7 +81,7 @@ class MemoryToolRuntimeImpl(ToolsProtocolPrivate, ToolRuntime, RAGToolRuntime):
         preprocessor_chain: Optional[PreprocessorChain] = None,
     ) -> None:
         preprocessor_inputs = [self._rag_document_to_preprocessor_input(d) for d in documents]
-        preprocessor_response = await self.preprocessing_api.chain_preprocess(
+        preprocessor_response = await self.preprocessing_api.preprocess(
             preprocessors=preprocessor_chain or self.DEFAULT_PREPROCESSING_CHAIN,
             preprocessor_inputs=preprocessor_inputs,
         )