Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix tags propogation with seldon client #3374

Merged
merged 6 commits into from
Jul 8, 2021
Merged

Fix tags propogation with seldon client #3374

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
b42a51c
client
majolo Jul 6, 2021
6c24d32
tags
majolo Jul 6, 2021
4402fb7
try catch
majolo Jul 6, 2021
792b9a6
fixes
majolo Jul 6, 2021
11fce5d
comment
majolo Jul 6, 2021
b1804a9
lint
majolo Jul 8, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion python/licenses/license_info.no_versions.csv
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@
"cffi","MIT License"
"chardet","GNU Library or Lesser General Public License (LGPL)"
"click","BSD License"
"cryptography","Apache Software License
"cryptography","Apache Software License; BSD License"
"flatbuffers","Apache Software License"
"gast","BSD License"
"google-auth","Apache Software License"
Expand Down
74 changes: 39 additions & 35 deletions python/seldon_core/batch_processor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -267,7 +267,7 @@ def _send_batch_predict_multi_request(
retries: int,
batch_id: str,
payload_type: str,
) -> str:
) -> [str]:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this coming from lint?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No it was a warning from my IDE

"""
Send an request using the Seldon Client with batch context including the
unique ID of the batch and the Batch enumerated index as metadata. This
Expand Down Expand Up @@ -297,13 +297,11 @@ def _send_batch_predict_multi_request(
instance_ids = [x[1] for x in input_data]

predict_kwargs = {}
meta = {
"tags": {
"batch_id": batch_id,
}
tags = {
"batch_id": batch_id,
}

predict_kwargs["meta"] = meta
predict_kwargs["meta"] = tags
predict_kwargs["headers"] = {"Seldon-Puid": instance_ids[0]}

try:
Expand All @@ -327,7 +325,7 @@ def _send_batch_predict_multi_request(
except Exception as e:
error_resp = {
"status": {"info": "FAILURE", "reason": str(e), "status": 1},
"meta": meta,
"meta": tags,
}
print("Exception: %s" % e)
str_output = json.dumps(error_resp)
Expand All @@ -343,26 +341,34 @@ def _send_batch_predict_multi_request(
tensor_ndarray = tensor.reshape(shape)

for i in range(len(input_data)):
new_response = copy.deepcopy(response)
if payload_type == "ndarray":
# Format new responses for each original prediction request
new_response["data"]["ndarray"] = [response["data"]["ndarray"][i]]
new_response["meta"]["tags"]["tags"]["batch_index"] = indexes[i]
new_response["meta"]["tags"]["tags"]["batch_instance_id"] = instance_ids[i]
responses.append(json.dumps(new_response))
elif payload_type == "tensor":
# Format new responses for each original prediction request
new_response["data"]["tensor"]["shape"][0] = 1
new_response["data"]["tensor"]["values"] = np.ndarray.tolist(
tensor_ndarray[i]
)
new_response["meta"]["tags"]["tags"]["batch_index"] = indexes[i]
new_response["meta"]["tags"]["tags"]["batch_instance_id"] = instance_ids[i]
responses.append(json.dumps(new_response))
else:
raise RuntimeError(
"Only `ndarray` and `tensor` input are currently supported for batch size greater than 1."
)
try:
new_response = copy.deepcopy(response)
if payload_type == "ndarray":
# Format new responses for each original prediction request
new_response["data"]["ndarray"] = [response["data"]["ndarray"][i]]
new_response["meta"]["tags"]["batch_index"] = indexes[i]
new_response["meta"]["tags"]["batch_instance_id"] = instance_ids[i]
responses.append(json.dumps(new_response))
elif payload_type == "tensor":
# Format new responses for each original prediction request
new_response["data"]["tensor"]["shape"][0] = 1
new_response["data"]["tensor"]["values"] = np.ndarray.tolist(
tensor_ndarray[i]
)
new_response["meta"]["tags"]["batch_index"] = indexes[i]
new_response["meta"]["tags"]["batch_instance_id"] = instance_ids[i]
responses.append(json.dumps(new_response))
else:
raise RuntimeError(
"Only `ndarray` and `tensor` input are currently supported for batch size greater than 1."
)
except Exception as e:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not clear what exception are we trying to caught here

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

But looking at other parts of code they follow same pattern so I guess it is okay for now.

error_resp = {
"status": {"info": "FAILURE", "reason": str(e), "status": 1},
"meta": tags,
}
print("Exception: %s" % e)
responses.append(json.dumps(error_resp))

return responses

Expand Down Expand Up @@ -405,14 +411,12 @@ def _send_batch_predict(
"""

predict_kwargs = {}
meta = {
"tags": {
"batch_id": batch_id,
"batch_instance_id": batch_instance_id,
"batch_index": batch_idx,
}
tags = {
"batch_id": batch_id,
"batch_instance_id": batch_instance_id,
"batch_index": batch_idx,
}
predict_kwargs["meta"] = meta
predict_kwargs["meta"] = tags
predict_kwargs["headers"] = {"Seldon-Puid": batch_instance_id}
try:
data = json.loads(input_raw)
Expand All @@ -439,7 +443,7 @@ def _send_batch_predict(
except Exception as e:
error_resp = {
"status": {"info": "FAILURE", "reason": str(e), "status": 1},
"meta": meta,
"meta": tags,
}
print("Exception: %s" % e)
str_output = json.dumps(error_resp)
Expand Down
16 changes: 11 additions & 5 deletions python/seldon_core/seldon_client.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -327,7 +327,7 @@ def predict(
http_path:
Custom http path for predict call to use
meta:
Custom meta map
Custom meta map, supplied as tags
client_return_type
the return type of all functions can be either dict or proto
raw_data
Expand Down Expand Up @@ -1208,6 +1208,7 @@ def rest_predict_seldon(
names: Iterable[str] = None,
client_return_type: str = "proto",
raw_data: Dict = None,
meta: Dict = {},
**kwargs,
) -> SeldonClientPrediction:
"""
Expand Down Expand Up @@ -1235,28 +1236,33 @@ def rest_predict_seldon(
the return type of all functions can be either dict or proto
raw_data
Raw payload (dictionary) given by the user
meta
Custom meta data map, supplied as tags
kwargs

Returns
-------
Seldon Client Prediction

"""
metaKV = prediction_pb2.Meta()
metaJson = {"tags": meta}
json_format.ParseDict(metaJson, metaKV)
if raw_data:
request = json_to_seldon_message(raw_data)
payload = raw_data
else:
if bin_data is not None:
request = prediction_pb2.SeldonMessage(binData=bin_data)
request = prediction_pb2.SeldonMessage(binData=bin_data, meta=metaKV)
elif str_data is not None:
request = prediction_pb2.SeldonMessage(strData=str_data)
request = prediction_pb2.SeldonMessage(strData=str_data, meta=metaKV)
elif json_data is not None:
request = json_to_seldon_message({"jsonData": json_data})
else:
if data is None:
data = np.random.rand(*shape)
datadef = array_to_grpc_datadef(payload_type, data, names=names)
request = prediction_pb2.SeldonMessage(data=datadef)
request = prediction_pb2.SeldonMessage(data=datadef, meta=metaKV)
payload = seldon_message_to_json(request)

response_raw = requests.post(
Expand Down Expand Up @@ -1808,7 +1814,7 @@ def grpc_predict_gateway(
channel_credentials
Channel credentials - see SeldonChannelCredentials
meta
Custom meta data map
Custom meta data map, supplied as tags
client_return_type
the return type of all functions can be either dict or proto
raw_data
Expand Down