Serving: PredictResponse as JSON

[mm-manu]

Hello folks,
does anybody know how to turn a PredictResponse to JSON/Dictionary?

The response looks like this:

outputs { key: "classes" value { dtype: DT_STRING tensor_shape { dim { size: 1 } dim { size: 5 } } string_val: "giant panda, panda, panda bear, coon bear, Ailuropoda melanoleuca" string_val: "indri, indris, Indri indri, Indri brevicaudatus" string_val: "lesser panda, red panda, panda, bear cat, cat bear, Ailurus fulgens" string_val: "gibbon, Hylobates lar" string_val: "sloth bear, Melursus ursinus, Ursus ursinus" } } outputs { key: "scores" value { dtype: DT_FLOAT tensor_shape { dim { size: 1 } dim { size: 5 } } float_val: 8.986181259155273 float_val: 5.365548610687256 float_val: 4.926231861114502 float_val: 2.988001823425293 float_val: 2.765580892562866 } }

[mountaintom]

In C++, Protocol Buffers has built-in support for Json serialization.

Here are a few lines of code from a project I wrote.

………..
………..
Status status = stub_->Predict(&context, predictRequest, &response);

if (status.ok()) {
………….
………….
google::protobuf::util::Status response_serialize_status =
google::protobuf::util::MessageToJsonString(
response, &serialized_result_object, jprint_options);

if (!response_serialize_status.ok()) {

………..
………..

The nice thing is it works directly on the response object returned by gRPC. You have to like the format it gives you, but for many things it is perfectly OK. Strings will be base64 encoded, so you will need to encode/decode them.

It looks like there may be something similar for Python.
https://developers.google.com/protocol-buffers/docs/reference/python/google.protobuf.json_format-module

[mm-manu]

Great, thanks for the hint!