refactor: centralize endpoint path strings into constants

Move hardcoded endpoint_path string literals from individual endpoint handlers into shared Final[str] constants in constants.py. This makes metric labeling paths consistent and easier to maintain across infer, query, streaming_query, and responses endpoints.

Signed-off-by: Major Hayden <major@redhat.com>

Author: major

src/app/endpoints/query.py

@@ -22,6 +22,7 @@
 from authorization.middleware import authorize
 from client import AsyncLlamaStackClientHolder
 from configuration import configuration
+from constants import ENDPOINT_PATH_QUERY
 from log import get_logger
 from models.common.responses.responses_api_params import ResponsesApiParams
 from models.config import Action
@@ -170,7 +171,7 @@ async def query_endpoint_handler(
 
     # Moderation input is the raw user content (query + attachments) without injected RAG
     # context, to avoid false positives from retrieved document content.
-    endpoint_path = "/v1/query"
+    endpoint_path = ENDPOINT_PATH_QUERY
     moderation_input = prepare_input(query_request)
     moderation_result = await run_shield_moderation(
         client, moderation_input, endpoint_path, query_request.shield_ids

src/app/endpoints/responses.py

@@ -36,7 +36,7 @@
 from authorization.middleware import authorize
 from client import AsyncLlamaStackClientHolder
 from configuration import configuration
-from constants import SUBSTITUTED_INSTRUCTIONS_PLACEHOLDER
+from constants import ENDPOINT_PATH_RESPONSES, SUBSTITUTED_INSTRUCTIONS_PLACEHOLDER
 from log import get_logger
 from models.common.responses.responses_api_params import ResponsesApiParams
 from models.common.responses.responses_context import ResponsesContext
@@ -329,7 +329,7 @@ async def responses_endpoint_handler(
     )
     attachments_text = extract_attachments_text(original_request.input)
 
-    endpoint_path = "/v1/responses"
+    endpoint_path = ENDPOINT_PATH_RESPONSES
     moderation_result = await run_shield_moderation(
         client,
         input_text + "\n\n" + attachments_text,

src/app/endpoints/rlsapi_v1.py

@@ -22,6 +22,7 @@
 from authorization.middleware import authorize
 from client import AsyncLlamaStackClientHolder
 from configuration import configuration
+from constants import ENDPOINT_PATH_INFER
 from log import get_logger
 from metrics import recording
 from models.config import Action
@@ -241,7 +242,7 @@ async def retrieve_simple_response(
     instructions: str,
     tools: Optional[list[Any]] = None,
     model_id: Optional[str] = None,
-    endpoint_path: str = "/v1/infer",
+    endpoint_path: str = ENDPOINT_PATH_INFER,
 ) -> str:
     """Retrieve a simple response from the LLM for a stateless query.
 
@@ -678,7 +679,7 @@ async def infer_endpoint(  # pylint: disable=R0914
     if quota_id is not None:
         check_tokens_available(configuration.quota_limiters, quota_id)
 
-    endpoint_path = "/v1/infer"
+    endpoint_path = ENDPOINT_PATH_INFER
 
     request_id = get_suid()
 

src/app/endpoints/streaming_query.py

@@ -47,6 +47,7 @@
 from client import AsyncLlamaStackClientHolder
 from configuration import configuration
 from constants import (
+    ENDPOINT_PATH_STREAMING_QUERY,
     INTERRUPTED_RESPONSE_MESSAGE,
     LLM_TOKEN_EVENT,
     LLM_TOOL_CALL_EVENT,
@@ -227,7 +228,7 @@ async def streaming_query_endpoint_handler(  # pylint: disable=too-many-locals
     # Moderation input is the raw user content (query + attachments) without injected RAG
     # context, to avoid false positives from retrieved document content.
     moderation_input = prepare_input(query_request)
-    endpoint_path = "/v1/streaming_query"
+    endpoint_path = ENDPOINT_PATH_STREAMING_QUERY
     moderation_result = await run_shield_moderation(
         client, moderation_input, endpoint_path, query_request.shield_ids
     )

src/constants.py

@@ -244,6 +244,13 @@
 # system prompt for the client's instructions.  Avoids leaking the actual
 # server prompt back to the client.
 SUBSTITUTED_INSTRUCTIONS_PLACEHOLDER: Final[str] = "<server prompt applied>"
+
+# API endpoint path constants used for metric labeling across endpoint handlers.
+ENDPOINT_PATH_INFER: Final[str] = "/v1/infer"
+ENDPOINT_PATH_QUERY: Final[str] = "/v1/query"
+ENDPOINT_PATH_STREAMING_QUERY: Final[str] = "/v1/streaming_query"
+ENDPOINT_PATH_RESPONSES: Final[str] = "/v1/responses"
+
 # Input size limits for API request validation
 # Maximum character length for the question field in /v1/infer requests (32 KiB)
 RLSAPI_V1_QUESTION_MAX_LENGTH: Final[int] = 32_768