small cleanups in req logger

components/seldon-request-logger/app/default_logger.py

@@ -86,7 +86,7 @@ def metadata():
         print('in metadata endpoint - this is for debugging')
         print(request.args)
         serving_engine = request.args.get('serving_engine','SeldonDeployment')
-        if serving_engine is None or not serving_engine:
+        if not serving_engine:
             serving_engine = 'SeldonDeployment'
 
         namespace = request.args.get('namespace','seldon')
@@ -145,7 +145,7 @@ def process_and_update_elastic_doc(
 
                 item_body[message_type]["instance"] = item
 
-                if type(elements) == type([]) and len(elements) >= num:
+                if type(elements) == type([]) and len(elements) > num:
                     item_body[message_type]["elements"] = elements[num]
 
                 item_request_id = build_request_id_batched(
@@ -477,7 +477,7 @@ def createElelmentsArray(X: np.ndarray, names: list, namespace_name, serving_eng
         sys.stdout.flush()
 
     results = None
-    if not metadata_schema or metadata_schema is None:
+    if not metadata_schema:
         results = createElementsNoMetadata(X, names, results)
     else:
         results = createElementsWithMetadata(X, names, results, metadata_schema, message_type)
@@ -597,7 +597,7 @@ def mergeLinkedColumns(raw_list, metadata_dict):
 def lookupValueWithMetadata(name, metadata_dict, raw_value):
     metadata_elem = metadata_dict[name]
 
-    if metadata_elem is None:
+    if not metadata_elem:
         return raw_value
 
     #categorical currently only case where we replace value

components/seldon-request-logger/app/log_helper.py

@@ -29,7 +29,7 @@ def get_max_payload_bytes(default_value):
 
 def extract_request_id(headers):
     request_id = headers.get(REQUEST_ID_HEADER_NAME)
-    if request_id is None:
+    if not request_id:
         # TODO: need to fix this upstream - https://github.com/kubeflow/kfserving/pull/699/files#diff-de6e9737c409666fc6c48dbcb50363faR18
         request_id = headers.get(CLOUD_EVENT_ID)
     return request_id
@@ -51,22 +51,22 @@ def build_index_name(headers):
     # otherwise create an index per deployment
     # index_name = "inference-log-" + serving_engine(headers)
     namespace = clean_header(NAMESPACE_HEADER_NAME, headers)
-    if namespace is None:
+    if not namespace:
         index_name = index_name + "-unknown-namespace"
     else:
         index_name = index_name + "-" + namespace
     inference_service_name = clean_header(INFERENCESERVICE_HEADER_NAME, headers)
     # won't get inference service name for older kfserving versions i.e. prior to https://github.com/kubeflow/kfserving/pull/699/
-    if inference_service_name is None or not inference_service_name:
+    if not inference_service_name:
         inference_service_name = clean_header(MODELID_HEADER_NAME, headers)
 
-    if inference_service_name is None:
+    if not inference_service_name:
         index_name = index_name + "-unknown-inferenceservice"
     else:
         index_name = index_name + "-" + inference_service_name
 
     endpoint_name = clean_header(ENDPOINT_HEADER_NAME, headers)
-    if endpoint_name is None:
+    if not endpoint_name:
         index_name = index_name + "-unknown-endpoint"
     else:
         index_name = index_name + "-" + endpoint_name
@@ -111,14 +111,14 @@ def set_metadata(content, headers, message_type, request_id):
 
     inference_service_name = content.get(INFERENCESERVICE_HEADER_NAME)
     # kfserving won't set inferenceservice header
-    if inference_service_name is None or not inference_service_name:
+    if not inference_service_name:
         content[INFERENCESERVICE_HEADER_NAME] = clean_header(
             MODELID_HEADER_NAME, headers
         )
 
     if message_type == "request" or not "@timestamp" in content:
         timestamp = headers.get(TIMESTAMP_HEADER_NAME)
-        if timestamp is None:
+        if not timestamp:
             timestamp = datetime.datetime.now(datetime.timezone.utc).isoformat()
         content["@timestamp"] = timestamp
 
@@ -134,17 +134,17 @@ def serving_engine(headers):
         return "inferenceservice"
 
 def get_header(header_name, headers):
-    if not headers.get(header_name) is None:
+    if headers.get(header_name):
         return clean_header(header_name, headers)
 
 def field_from_header(content, header_name, headers):
-    if not headers.get(header_name) is None:
+    if headers.get(header_name):
         content[header_name] = clean_header(header_name, headers)
 
 
 def clean_header(header_name, headers):
     header_val = headers.get(header_name)
-    if not header_val is None:
+    if header_val:
         header_val = header_val.translate({ord(c): None for c in '!@#$"<>/?'})
     return header_val
 

components/seldon-request-logger/app/log_mapping.py

@@ -135,14 +135,14 @@ def init_api():
         config.verify_ssl = False
         os.environ["CURL_CA_BUNDLE"] = ""
 
-    if not config.auth_method or config.auth_method is None:
+    if not config.auth_method:
         config.auth_method = 'password_grant'
 
-    if not config.host or config.host is None:
+    if not config.host:
         print('No DEPLOY_API_HOST - will not look up metadata from Deploy')
         return
 
-    if not config.oidc_server or config.oidc_server is None:
+    if not config.oidc_server:
         print('No OIDC_PROVIDER - auth will not be used in connecting to metadata')
         return
 
@@ -176,7 +176,6 @@ def fetch_metadata(namespace, serving_engine, inferenceservice_name, predictor_n
     if not deployment_type:
         print('unknown deployment type for '+namespace+' / '+inferenceservice_name)
         print(deployment_type)
-        #TODO: should be sending deployment_type in call below but currently sdk says unexpected keyword argument
 
     if metadata_api is None:
         print('metadata service not configured')
@@ -185,7 +184,8 @@ def fetch_metadata(namespace, serving_engine, inferenceservice_name, predictor_n
     # TODO: in next iteration will only need one lookup straight to model metadata
     # was expcting to set deployment_type=serving_engine but deployment_type does not seem to be a param
     runtime_metadata = metadata_api.model_metadata_service_list_runtime_metadata_for_model(
-        deployment_name=inferenceservice_name,deployment_namespace=namespace,predictor_name=predictor_name)
+        deployment_name=inferenceservice_name,deployment_namespace=namespace,
+        predictor_name=predictor_name,deployment_type=deployment_type)
 
     if runtime_metadata is not None and runtime_metadata and \
             runtime_metadata.runtime_metadata is not None and runtime_metadata.runtime_metadata:

components/seldon-request-logger/requirements.txt

@@ -4,4 +4,4 @@ dict_digger==0.2.1
 seldon_core
 elasticsearch==7.12.1
 click==8.0.0a1
-seldon-deploy-sdk==1.3.0.dev2
+seldon-deploy-sdk==1.3.0.dev3