Google NLP API performance - REST API vs Node js client API

[Kapil Gahlot]

During my testing of Google NLP API I found REST API response time (1.4 secs) to be faster than Node js client API response time (2 secs) . I am making API calls from a Meteor server. However Google API docs suggest to prefer client API over REST without providing any reasons. Is it a good decision to use REST API rather than client API just based on the response times that I am getting or are there other considerations in favor of the client API?

[Kapil Gahlot]

console.log("api time : ", moment.now() - startTime);

However Google API docs suggest to prefer client API over REST without providing any reasons. Is it a good decision to use REST API rather than client API just based on the response times that I am getting or are there other considerations in favor of the client API?

During my testing of Google NLP API I found REST API response time (1.4 secs) to be faster than Node js client API response time (2 secs) . I am making API calls from a Meteor server.

Here is a snapshot of the code: Node js client: in an async function: set up auth, request, etc.

let startTime = moment.now(); response = await client.annotateText(request); console.log("api time : ", moment.now() - startTime); REST API: in an async function: set up data, etc. let startTime = moment.now();

response = await HTTP.call("POST",

"https://language.googleapis.com/v1beta2/documents:annotateText?key=KEY_VALUE",

{ data }

);

Thanks

[Yannick (Cloud Platform Support)]

Hello Kapil, the justification for that recommendation can be found here. In essence the client libraries are recommended for their ease of use and you are not at all discouraged from using the REST API directly.

[Kapil Gahlot]

Thanks a lot. It helps. I will go ahead and use REST as it consistently provides better response time to me and using node js is anyway easy to work with json response from REST API.

Regards