react-native-dialogflow-hermes (react-native-api-ai)

A React-Native Bridge for the Google Dialogflow AI SDK.

Support for iOS 10+ and Android!

Dialogflow is a powerful tool for building delightful and natural conversational experiences. You can build chat and speech bots and may intergrate it in a lot of platform like twitter, facebook, slack, or alexa.

Install

This package depends on react-native-voice, follow their readme to setup it.

Add react-native-dialogflow-hermes and link it:

npm install --save react-native-dialogflow-hermes react-native-voice

react-native link react-native-dialogflow-hermes
react-native link react-native-voice

iOS: IMPORTANT xCode plist settings

Also, you need open the React Native xCode project and add two new keys into Info.plist Just right click on Info.plist -> Open As -> Source Code and paste these strings somewhere into root <dict> tag

<key>NSSpeechRecognitionUsageDescription</key>
<string>Your usage description here</string>
<key>NSMicrophoneUsageDescription</key>
<string>Your usage description here</string>

Application will crash if you don't do this.

Usage

Import Dialogflow:

import Dialogflow from "react-native-dialogflow-hermes";

or for V2

import { Dialogflow_V2 } from "react-native-dialogflow-hermes"

Configuration

Set the accessToken and the language in your constructor:

 constructor(props) {
        super(props);

        Dialogflow.setConfiguration(
          "4xxxxxxxe90xxxxxxxxc372", Dialogflow.LANG_GERMAN
        );
    }

For V2 you can set the client_email and private_key of the credential json auth setup. In addition you have to set your projectId:

 constructor(props) {
        super(props);

        Dialogflow_V2.setConfiguration(
            "[email protected]",
            '-----BEGIN PRIVATE KEY-----\nMIIEvgIBADAN...1oqO\n-----END PRIVATE KEY-----\n',
            Dialogflow_V2.LANG_GERMAN,
            'testv2-3b5ca'
        );
    }

Listening

Start listening with integrated speech recognition:

   <Button onPress={() => {
            Dialogflow.startListening(result=>{
                console.log(result);
            }, error=>{
                console.log(error);
            });
        }}
   />

In iOS only you have to call finishListening(). Android detects the end of your speech automatically. That's the reason why we didn't implement the finish method in Android.

// only for iOS
Dialogflow.finishListening();
// after this call your callbacks from the startListening will be executed.

Text Request

For using your own speech recognition:

   <Button onPress={() => {
           Dialogflow.requestQuery("Some text for your Dialogflow agent", result=>console.log(result), error=>console.log(error));
        }}
   />

Request an Event

For sending an event to Dialogflow (Contexts and Entities have no effect!):

Dialogflow.requestEvent(
    "WELCOME",
    {param1: "yo mr. white!"},
    result=>{console.log(result);},
    error=>{console.log(error);}
);

Contexts

Set contexts (will take affect on next startListening or queryRequest):

const contexts = [{
  name: "deals",
  lifespan: 1,
  parameters: {
      Shop: "Rewe"
  }
}];

Dialogflow.setContexts(contexts);

Reset all (non-permantent) contexts for current session:

Dialogflow.resetContexts(result=>{
       console.log(result);
    }, error=>{
       console.log(error);
    });

Set permanent contexts, which will be set automatically before every request. This is useful for e.g. access tokens in webhooks:

const permanentContexts = [{
  name: "Auth",
  // lifespan 1 is set automatically, but it's overrideable
  parameters: {
      AccessToken: "1234yo1234"
  }
}];

Dialogflow.setPermanentContexts(permanentContexts);

Entities

Set UserEntities (will take affect on next startListening or queryRequest):

const entities = [{
  "name":"shop",
  "extend":true,
  "entries":[
      {
          "value":"Media Markt",
          "synonyms":["Media Markt"]
      }
  ]
 }];

 Dialogflow.setEntities(entities);

Listener for Android

Only in Android we have four additional methods: onListeningStarted, onListeningCanceled, onListeningFinished and onAudioLevel. In iOS they will be never called:

   <Button onPress={() => {

            Dialogflow.onListeningStarted(()=>{
                console.log("listening started");
            });

            Dialogflow.onListeningCanceled(()=>{
                console.log("listening canceled");
            });

            Dialogflow.onListeningFinished(()=>{
                console.log("listening finished");
            });

            Dialogflow.onAudioLevel(level=>{
                console.log(level);
            });


            Dialogflow.startListening(result=>{
                console.log(result);
            }, error=>{
                console.log(error);
            });
        }}
   />

Note: Make sure you are setting the callbacks before startListening every single time again. Don't set the callbacks in e.g. constructor or componentsDidMount if you are executing startListening more than one times.

Supported Languages

Set the language in your configuration:

Dialogflow.setConfiguration("4xxxxxxxe90xxxxxxxxc372", Dialogflow.LANG_GERMAN);

• LANG_CHINESE_CHINA
• LANG_CHINESE_HONGKONG
• LANG_CHINESE_TAIWAN
• LANG_DUTCH
• LANG_ENGLISH
• LANG_ENGLISH_GB
• LANG_ENGLISH_US
• LANG_FRENCH
• LANG_GERMAN
• LANG_ITALIAN
• LANG_JAPANESE
• LANG_KOREAN
• LANG_PORTUGUESE
• LANG_PORTUGUESE_BRAZIL
• LANG_RUSSIAN
• LANG_SPANISH
• LANG_UKRAINIAN

Methods

| name | platform | param1 | param2 | param3 | param4 | | --------------------- | -------- | --------- | --------- | --------- | --------- | | setConfiguration (V1) | both | accessToken: String | languageTag: String | | setConfiguration (V2) | both | client_email: String | private_key: String | languageTag: String | projectId: String | | startListening | both | resultCallback: (result: object)=>{} | errorCallback: (error: object)=>{} | | | finishListening | ios | | | | | requestQuery | both | query: String | resultCallback: (result: object)=>{} | errorCallback: (error: object)=>{} | | requestEvent | both | eventName: String | eventData: Object | resultCallback: (result: object)=>{} | errorCallback: (error: object)=>{} | | onListeningStarted | both | callback: ()=>{} | | | | onListeningCanceled | none | callback: ()=>{} || | | onListeningFinished | both | callback: ()=>{} | | | | onAudioLevel | android | callback: (level: number)=>{} || | | setContexts | both | array || | | resetContexts | both | resultCallback: (result: object)=>{} | errorCallback: (error: object)=>{} | | | setPermanentContexts| both | array || | | setEntities (V1 only)| both | array || |

Blogpost

Deutsch

Sprachsteuerung mit Api.ai in einer React-Native App

English

Contributors

• Anton Spöck
• Tobias Jonas
• Maximilian Grassl

Powered by innFactory