feat: add processor_result to reject flow

src/f5_ai_gateway_sdk/result.py

@@ -156,6 +156,7 @@ class Reject(BaseModel):
     # be added to the metadata field
     metadata: Metadata = Field(default=Metadata(), exclude=True)
     tags: Tags = Field(default=Tags(), exclude=True)
+    processor_result: Metadata | None = None
 
     def is_empty(self) -> bool:
         """Compatability with Result(), always false due to required fields"""
@@ -169,6 +170,8 @@ def to_response(self) -> Response:
         """Return Reject as Response object"""
         if self.tags:
             self.metadata["tags"] = self.tags.to_response()
+        if self.processor_result:
+            self.metadata["processor_result"] = self.processor_result
         return MultipartResponse(
             fields=[
                 self.to_multipart_field(),

tests/test_processor.py

@@ -446,6 +446,54 @@ async def test_handle_rejected_prompt(self):
 
         self.assertDictEqual(expected_response_metadata, response_metadata)
 
+    async def test_handle_rejected_prompt_with_result(self):
+        prompt = TEST_REQ_INPUT.model_dump_json()
+        metadata = {"key": "value", "step_id": "12345", "request_id": "09876"}
+        request = fake_multipart_request(
+            prompt=prompt,
+            metadata=metadata,
+            parameters={"reject": True, "annotate": True},
+        )
+        result = Reject(
+            metadata=metadata,
+            code=RejectCode.POLICY_VIOLATION,
+            detail="dangerous question asked",
+            tags=FAKE_TAGS,
+            processor_result={"confidence": 0.99},
+        )
+        processor = fake_processor(result=result)
+
+        response = await processor.handle_request(request)
+
+        self.assertStatusCodeEqual(response, HTTP_200_OK)
+
+        content = await self.buffer_response(response)
+        multipart = MultipartDecoderHelper(
+            content=content, content_type=response.headers["Content-Type"]
+        )
+
+        self.assertFalse(
+            multipart.has_prompt(), "the rejected prompt should not be in the response"
+        )
+
+        multipart_metadata = multipart.metadata
+        self.assertEqual(
+            MultipartResponse.JSON_CONTENT_TYPE, multipart_metadata.content_type()
+        )
+        response_metadata = multipart_metadata.as_json()
+
+        expected_response_metadata = dict(
+            app_details=APP_DETAILS,
+            processor_id=processor.id(),
+            processor_result={"confidence": 0.99},
+            processor_version=processor.version,
+            tags={"test1": ["a", "b"]},
+        )
+        for k, v in metadata.items():
+            expected_response_metadata[k] = v
+
+        self.assertDictEqual(expected_response_metadata, response_metadata)
+
     async def test_handle_modified_prompt(self):
         prompt = TEST_REQ_INPUT.model_dump_json()
         metadata = {"key": "value"}