Interpretation

Each interpretation represent how the NLU engine understand user's query then performing the corresponding fulfillment.

Sample interpretation

{
   "domain" : "entertainment",
   "intent" : "music",
   "score" : 4.15378032316986,
   "displayText" : "Playing coldplay",
   "speechText" : "Playing coldplay",
   "hints" : [],
   "debug" : {},
   "status" : 200,
   "reason" : "",
   "fulfillment" : {...}
}

Descriptions

domain

 Identify the specific domain for this particular interpretation. Could be "entertainment",
 "navigation", "communication", "generic" and several others.

intent

 The intent represent the interpretation between what user says and what actions should be taken
 by the software or product.

score

 A number to indicate the relative confidence score when compared to other interpretations.

speechText

 Text string that client can use to call the text to speech service.

displayText

 The text string that client can use to display on the GUI.

debug

 TBD, the engine debug information stays here.

hints

 TBD, possible reprompt to tell users what they can say to finish this task flow.

status

 Indicate the status of the fulfilling this specific interpretation. Check the reason field
 if status is above 200.

reason

 The detailed description about why this interpretation is failed to fulfill.

fulfillment

This is the one of the final results after parsing NLU's response, calling service providers and aggregating their results.

The client can directly use the results for actions, the action could be rendering the results on the GUI, generating the speech response to the users, or using the weblink and deeplink to start streaming the third party contents.

-Documentation for fulfillment

Interpretation errors

400:

We didn't try to fulfill the interpretation because we knew it would fail

{
  "status": 400,
  "reason": "invalid latitude and longitude"
}

503:

We tried to fulfill the interpretation, but it failed.

{
  "status": 503,
  "reason": error.message
}