# Computes Dictation NPM

Performs massively parallel transcription of audio files using Watson Speech-to-Text and Computes.io.

## Usage

Computes.io kernel would look something like this:

````
// Watson Speech to Text
var computes = require("computes");
var operation = "npm://computes-dictation@latest";

var options = {
  domain: "domain-key-provided-by-computesio",
  priority: "normal",  //'low', 'normal' & 'high'
  ttl: 60000,  // milliseconds
  delay: 0 //milliseconds
};

var job = computes.connect(options.domain);

job.on("ready", function (){

  var params = {
    audioUrl: "https://www.dropbox.com/s/eflf4ola9c7kkxi/0.wav?dl=1",
    contentType: "audio/wav; rate=48000",
    username: "watson-username",
    password: "watson-password"
  };

  job.compute(operation, params, options);

});

job.on("result", function (results){
  console.log(results.result);
  job.disconnect();
});

````
````
{ result: 'the house has a thatched roof ' }
````

## Reference documentation
- [Computes.io](https://computes.io)

- [Watson Speech-To-Text API Docs](https://watson-developer-cloud.github.io/node-sdk/conversation-v1/)

- [Watson Audio formats](https://www.ibm.com/watson/developercloud/doc/speech-to-text/input.shtml)