Implementing Feedback Feature in SUSI Android App

Recently, on SUSI Server, a new servlet was added which is used to rate SUSI Skills either positive or negative. The server stores the rating of a particular skill in a JSON file. These ratings help in improving answers provided by SUSI. So, the server part is done and it was required to implement this in the SUSI Android App. In this blog, I will cover the topic of implementation of the Rating or Feedback feature in SUSI Android App. This will including all the cases when feedback should be sent, when it should not be sent, when to send positive feedback, when to send negative feedback, etc.

API Information

For rating a SUSI Skill, we have to call on  /cms/rateSkill.json providing 5 parameters which are:

  1. model: The model of SUSI Skill. (String)
  2. group: The Group under the model in which that particular skill resides. (String)
  3. language: The language of skill. (String)
  4. skill: This is skill name. (String)
  5. rating: This can be two strings, either “positive” or “negative”. String)

Basically, in the SUSI Skill Data repo (in which all the skills are stored), models, groups language etc are part of folder structure.

So, if a skill is located here

This would mean

model = general

group = Knowledge

language = en

skill = news

rating = positive/negative

Implementation in SUSI Android App


So, when the like button on a particular skill is clicked, a positive call is made and when the dislike button is clicked, a negative call is made.

Let’s see example when the thumbs up button or like button is clicked.

There can be three cases possible:

  1. None of Like button or dislike button is clicked already: In this case, initially, both like and dislike button will be transparent/hollow. So, when like button is clicked, the like button will be colored blue and a call will be made with positive feedback.
  2. Like button is already clicked: In this case, like button is already clicked. So, it will already be blue. So, when user clicks again on positive button, it should get back to normal/hollow indicating rating which was sent is cancelled and a a call will be made with negative feedback thus cancelling or neutralizing the earlier, positive feedback.
  3. Dislike button is already clicked: In this case, the dislike button is already blue, indicating a negative call is already made. So, now when the like button is clicked, we need to cancel the earlier negative feedback call and sending another negative feedback call. Thus, sending two negative feedback calls. And after that coloring dislike button as blue.

Look at the code below. It is self explanatory. There are three if-else conditions covering all the above mentioned three cases.

thumbsUp.setOnClickListener(new View.OnClickListener() {
   public void onClick(View view) {
       if(!model.isPositiveRated() && !model.isNegativeRated()) {
           rateSusiSkill(Constant.POSITIVE, model.getSkillLocation(), context);
           setRating(true, true);
       } else if(!model.isPositiveRated() && model.isNegativeRated()) {
           setRating(false, false);
           rateSusiSkill(Constant.POSITIVE, model.getSkillLocation(), context);
           rateSusiSkill(Constant.POSITIVE, model.getSkillLocation(), context);
           setRating(true, true);
       } else if (model.isPositiveRated() && !model.isNegativeRated()) {
           rateSusiSkill(Constant.NEGATIVE, model.getSkillLocation(), context);
           setRating(false, true);

Similarly for when dislike button is clicked, the above three mentioned cases still hold resulting in this code snippet.

thumbsDown.setOnClickListener(new View.OnClickListener() {
   public void onClick(View view) {
       if(!model.isPositiveRated() && !model.isNegativeRated()) {
           rateSusiSkill(Constant.NEGATIVE, model.getSkillLocation(), context);
           setRating(true, false);
       } else if(model.isPositiveRated() && !model.isNegativeRated()) {
           setRating(false, true);
           rateSusiSkill(Constant.NEGATIVE, model.getSkillLocation(), context);
           rateSusiSkill(Constant.NEGATIVE, model.getSkillLocation(), context);
           setRating(true, false);
       } else if (!model.isPositiveRated() && model.isNegativeRated()) {
           rateSusiSkill(Constant.POSITIVE, model.getSkillLocation(), context);
           setRating(false, false);


So, this blog talked about how the Feedback feature in SUSI Android App is implemented. This included how a network call is made, logic for sending positive/negative feedback, logic to withdraw feedback etc. So, If you are looking forward to contribute to SUSI Android App, this can help you a little. But if not so, this may also help you in understanding and how rating mechanism in social media websites like Facebook, Twitter, Quora, Reddit, etc work.


  1. To know about servlets
  2. To see how to implement one
  3. To see how to make network calls in android using Retrofit
  4. To see how to implement click listeners on button

Introduction To Kotlin in SUSI Android App

Lately, we wrote some of the code of SUSI Android App in Kotlin. Kotlin is a very similar language to Java but with much more advantages than Java. It is easy to adapt and learn. There is no doubt that Kotlin is better than Java but with the announcement of Kotlin Support in Google IO’17 for Android development, Kotlin seems a decent way to write code for an Android App.

Advantages of Kotlin over Java

    1. Reduce Boilerplate Code: It helps making development of app faster as it reduces more than 20 percent of boilerplate code. Writing long statements again and again is a headache for developers. Kotlin comes to rescue in that situation.
    2. Removes Null Pointer Exception: Once a large company faced millions of dollars of loss due to null pointer exception. It causes crashes of apps more often than anything else. Thus Kotlin helps in Null checks and makes app free from Null pointer Exceptions.
    3. Interoperable with Java: Kotlin code and Java code are interoperable. Which means you can write half your code in kotlin and half in Java and it will work like a charm. You can call java methods from Kotlin code and vice versa. So, you can simply move your existing Java based app to Kotlin slowly making your app always running.
    4. Lambda and Inline functions: Yes, Kotlin also has functionalities from functional programming languages. Mainly and most widely used feature of those languages is Lambda functions.
    5. Direct Reference of Views by Id: You do not need to write findViewById( or use any other library like Butterknife for view binding. You can simply use the view by its id.
    6. No semicolon:  Last but not the least, you do not need to add a semicolon after each statement. In fact, you do not need to add semicolon at all.

Setting up Android Studio to work with Kotlin

If you have latest Android Studio Canary Version, there is already a build support for Kotlin in it. You need not do anything in that case. But if you don’t have the Canary version, you can add Kotlin Plugin in your Android Studio. Follow the below steps to do that.

  1. Install the Kotlin Plugin:

Android Studio → Preferences… →Plugins → Browse Repository → type “Kotlin” in search box → install

  1. Restart your Android Studio and Rebuild the project. Everything else is already set up in SUSI Android App but if you want to do it for your other apps, follow this link.

Implementation in SUSI Android App

So, I am not going to give unnecessary code but will point out specific things where Kotlin helped a lot to reduce unnecessary code and made the code compact.

1. Listeners:

Earlier with Java

Button signup = (Button) findViewById(;

signup.setOnClickListener(new View.OnClickListener() {
            public void onClick(View v) {
               startActivity(new Intent(LoginActivity.this, SignUpActivity.class));

Now, with Kotlin

fun signUp() {
   sign_up.setOnClickListener { startActivity(Intent(this@LoginActivity, }

2. Models

With Java

public class MapData {

    private double latitude;
    private double longitude;
    private double zoom;

    public MapData(double latitude, double longitude, double zoom) {
        this.latitude = latitude;
        this.longitude = longitude;
        this.zoom = zoom;

    public double getLatitude() {
        return latitude;

    public void setLatitude(double latitude) {
        this.latitude = latitude;

    public double getLongitude() {
        return longitude;

    public void setLongitude(double longitude) {
        this.longitude = longitude;

    public double getZoom() {
        return zoom;

    public void setZoom(double zoom) {
        this.zoom = zoom;

With Kotlin

class MapData (var latitude: Double, var longitude: Double, var zoom: Double) 

3. Constructor

With Java

public class LoginPresenter {
    private LoginActivity loginActivity;
    public LoginPresenter(loginActivity: LoginActivity){
        this.loginActivity = loginActivity;

With Kotlin

class LoginPresenter(loginActivity: LoginActivity) {


So, this blog was to give you an idea about Kotlin programming language, it’s advantages over java and information how you can set it up on your Android Studio so as to help you a little in understanding the codebase of SUSI Android App a little more.


  1. Official Kotlin Guide for Syntax Reference and further learning
  2. Blog by Elye on Setting up Kotlin on Android Studio
  3. Youtube Video tutorial by Derek Banas on Kotlin

Auto Deployment of SUSI Server using Kubernetes on Google Cloud Platform

Recently, we auto deployed SUSI Server on Google Cloud Platform using Kubernetes and Docker Images after each commit in the GitHub repo with the help of Travis Continuous Integration. So, basically, whenever a new commit is added to the repo, during the Travis build, we build the docker image of the server and then use it to deploy the server on Google Cloud Platform. We use Kubernetes for deployment since it is very easy to scale up the Project when traffic on the server is increased and Docker because using it we can easily build docker images which then can be used to update the deployment. This schematic will make things more clear what exactly is the procedure.


  1. You must be signed in to your Google Cloud Console and have enabled billing and must have credits left in your account.
  2. You must have a docker account and a repo in it. If you don’t have one, make it now.
  3. You should have enabled Travis on your repo and have a Travis.yml file in your repo.
  4. You must already have a project in Google Cloud Console. Make a new one if you don’t have.

Pre Deployment Steps

You will be needed to do some work on Google Cloud Platform before actually starting the auto deployment process. Those are:

  1. Creating a new Cluster.
  2. Adding and Formatting Persistence Disk
  3. Adding a Persistent Volume CLaim (PVC)
  4. Labeling a node as primary.

Check out this documentation on how to do that. It may help.


Img src:

1. The first step is simply to add this line in Travis.yml file and create an empty, file mentioned below.

- bash kubernetes/travis/

Now we’ll be moving line by line and adding commands in the empty file that we created in the previous step.

2. Next step is to remove obsolete Google Cloud files and install Google Cloud SDK and kubectl command. Use following lines to do that.

echo ">>> Removing obsolete gcoud files"
sudo rm -f /usr/bin/
sudo rm -f /usr/bin/bq
sudo rm -f /usr/bin/gsutil
sudo rm -f /usr/bin/gcloud

echo ">>> Installing new files"
curl | bash;
source ~/.bashrc
gcloud components install kubectl

3. In this step you will be needed to download a JSON file which contains your Google Cloud Credentials, then copy that file to your repo and encrypt it using Travis encryption keys. Follow this video to see how to do that.

4. So, now you have added your encrypted credentials.json files in your repo and now you need to use those credentials to login into your google cloud account. So, use below lines to do that.

echo ">>> Decrypting credentials and authenticating gcloud account"
# Decrypt the credentials we added to the repo using the key we added with the Travis command line tool
openssl aes-256-cbc -K $encrypted_YOUR_key -iv $encrypted_YOUR_iv -in ./kubernetes/travis/Credentials.json.enc -out Credentials.json -d
gcloud auth activate-service-account --key-file Credentials.json
export GOOGLE_APPLICATION_CREDENTIALS=$(pwd)/Credentials.json
#add gcoud project id
gcloud config set project YOUR_PROJECT_ID
gcloud container clusters get-credentials YOUR_CONTAINER

The above lines of code first decrypt your credentials, then login into your account and set the project you already created earlier.

5. Now, we have logged into Google Cloud, we need to build docker image from a dockerfile. Follow official docker docs to see how to write a dockerfile. Here is an example of dockerfile. You will need to add “$DOCKER_USERNAME” and “$DOCKER_PASSWORD” as environment variables in Travis settings of your repo.

echo ">>> Building Docker image"
cd kubernetes/images

docker login -u="$DOCKER_USERNAME" -p="$DOCKER_PASSWORD"

6. Now, just push the docker image created in previous step and update the deployment.

echo ">>> Pushing docker image"

echo ">>> Updating deployment"


This blog was about how we have configured travis build and auto deployed SUSI Server on Google Cloud Platform using Kubernetes and Docker. You can do the same with your server too or if you are looking to contribute to SUSI Server, this may help you a little in understanding the code of the repo.


  1. The documentation for setting up your project on Google CLoud Console before starting auto deployment
  2. The documentation for encrypting your google cloud credentials and adding them to your repo
  3. Docs for Docker to get you started with Docker
  4. Travis Documentation on how to secure your credentials
  5. Travis Documentation on how to add environment variables in your repo settings

Encoding and Saving Images as Strings in Preferences in SUSI Android App

In this blog post, I’ll be telling about how to store images in preferences by encoding them into Strings and also how to retrieve them back. Many a times, you need to store an image in preferences for various purposes and then need to retrieve it back when required. In SUSI Android App, we need to store an image in preference to set the chat background. We just simply select image from gallery, convert image to a byte array, then do a Base 64 encoding to string, store it in preferences and later decode it and set the chat background.

Base64 Encoding-Decoding in Java

You may already know what Base 64 is but still here is a link to Wikipedia article explaining it. So, how to do a Base64 encoding-decoding in java? For that java has a class with all such methods already present.

According to the docs:

This class consists exclusively of static methods for obtaining encoders and decoders for the Base64 encoding scheme. The implementation of this class supports the following types of Base64 as specified in RFC 4648 and RFC 2045.

  • Basic
  • URL and Filename safe
  • MIME

So, you may just use Base64.encode to encode a byte array and Base64.decode to decode a byte array.


1. Selecting image from gallery


Start Image Picker Intent and pick an image from gallery. After selecting you may also start a Crop Intent to crop image also. After selecting and cropping image, you will get a URI of the image.

override fun openImagePickerActivity() {
   val i = Intent(Intent.ACTION_PICK, android.provider.MediaStore.Images.Media.EXTERNAL_CONTENT_URI)
   startActivityForResult(i, SELECT_PICTURE)
val thePic = data.extras.getParcelable<Bitmap>("data")
val encodedImage = ImageUtils.Companion.cropImage(thePic)

2. Getting image from the URI using inputstream

Next step is to get the image from file using URI from the previous step and InputStream class and store it in a BitMap variable.

val imageStream: InputStream = context.contentResolver.openInputStream(selectedImageUri)
   val selectedImage: Bitmap
   val filePathColumn = arrayOf(MediaStore.Images.Media.DATA)
   val cursor = context.contentResolver.query(getImageUrl(context.applicationContext, selectedImageUri), filePathColumn, null, null, null)
   selectedImage = BitmapFactory.decodeStream(imageStream)

3. Converting the bitmap to ByteArray

Now, just convert the Bitmap thus obtained to a ByteArray using below code.

val baos = ByteArrayOutputStream()
   selectedImage.compress(Bitmap.CompressFormat.JPEG, 100, baos)
   val b = baos.toByteArray()

4. Base64 encode the ByteArray and store in preference

Encode the the byte array obtained in last step to a String and store it in preferences.

 val encodedImage = Base64.encodeToString(b, Base64.DEFAULT)
//now you have a string. You can store it in preferences

5. Decoding the String to image

Now whenever you want, you can just decode the stored Base64 encoded string to a byte array and then from byte array to a bitmap and use wherever you want.

fun decodeImage(context: Context, previouslyChatImage: String): Drawable {
   val b = Base64.decode(previouslyChatImage, Base64.DEFAULT)
   val bitmap = BitmapFactory.decodeByteArray(b, 0, b.size)
   return BitmapDrawable(context.resources, bitmap)


So, the main aim of this blog was to give an idea about how can you store images in preferences. There is no way to store them directly. So, you have to convert them to String by encoding them in Base64 format and then decoding it to use it. You also have other ways to store images like storing it in database etc but this one is simpler and fast.


  1. Stackoverflow answer to “How to save image as String”
  2. Other Stackoverflow answer about “Saving Images in preferences”
  3. Official docs of Base64 class
  4. Wikipedia link for learning about Base64
  5. Stackoverflow answer for “What is Base64 encoding used for?”

Implementation of Text to Speech alongside Hotword Detection in SUSI Android App

In this blog post, we’ll be learning about how to implement Text to speech. Now you may be wondering that what is so difficult in implementing text to speech. One can easily find many tutorials on that and can easily look at the official documentation of TTS but there’s a catch here. In this blog post I’ll be telling about how to implement Text to Speech alongside Hotword Detection.

Let me give you a rough idea about how hotword detection works in SUSI Android App. For more details, read my other blog here on Hotword Detection. So, there is a constantly running background recording thread which detects when hotword is detected. Now, you may be thinking why do we need to stop that thread for text to speech. Well there are 2 reasons to do that:

  1. Recording while playing causing problems with mic and may crash the app.
  2. Suppose we even implement that but what will happen if the answer contains word “susi” in it. Now, the hotword will be detected because the speech output contained word “susi” in it (which is our hotword).

So, to avoid these problems we had to come up a way to stop hotword detection only for that particular time when SUSI is giving speech output and resume it back immediately when speech output is finished.

Let’s see how we did that.


Check out this video to see how this work in the app

Initiating the TTS engine

The first task is to initiate the Text to speech engine. This process takes some time. So, it is done in the starting of app in a new handler.

new Handler().post(new Runnable() {
   public void run() {
       textToSpeech = new TextToSpeech(getApplicationContext(), new TextToSpeech.OnInitListener() {
           public void onInit(int status) {
               if (status != TextToSpeech.ERROR) {
                   Locale locale = textToSpeech.getLanguage();

Check Audio Focus

The next step is to check whether audio focus is granted. Suppose there is some music playing in the background, in that case we won’t be able to give voice output. So, we check audio focus using below code.

final AudioManager audiofocus = (AudioManager) getSystemService(Context.AUDIO_SERVICE);
 int result = audiofocus.requestAudioFocus(afChangeListener, AudioManager.STREAM_MUSIC, AudioManager.AUDIOFOCUS_GAIN);
if (result == AudioManager.AUDIOFOCUS_REQUEST_GRANTED) {

Using OnAudioFocusChangeListener, we keep a track of when we have access to give speech output and when we don’t.

private AudioManager.OnAudioFocusChangeListener afChangeListener =
       new AudioManager.OnAudioFocusChangeListener() {
           public void onAudioFocusChange(int focusChange) {
               if (focusChange == AUDIOFOCUS_LOSS_TRANSIENT) {
               } else if (focusChange == AudioManager.AUDIOFOCUS_GAIN) {
                   // Resume playback
               } else if (focusChange == AudioManager.AUDIOFOCUS_LOSS) {

Converting the given text to speech

Now we have audio focus, we just have to convert given text to speech. Use method textToSpeech.speak().

private void voiceReply(final String reply) {
       Handler handler = new Handler(); Runnable() {
           public void run() {
                   textToSpeech.speak(spokenReply, TextToSpeech.QUEUE_FLUSH, ttsParams);                  

Abandon Audio Focus

Now we are done with speech output, it’s time we abandon audio focus.


TTS alongside Hotword Detection

Okay so now the major part. How do we check when to stop hotword detection thread and when to resume it? How do we check if Speech output is finished?

Answer to these questions is textToSpeech.setOnUtteranceProgressListener. The UtteranceProgressListener overrides 3 methods:

  1. onStart: Indicates starting of text to speech conversion. Which means it’s time to stop hotword detection thread.
  2. onDone: Called when every word of the provided text is converted to speech. So, simply resume hotword detection
  3. onError: Called when there is an error and text is not converted to speech. Anyway, we need to resume hotword detection here too.
textToSpeech.setOnUtteranceProgressListener(new UtteranceProgressListener() {
                       public void onStart(String s) {
                           if(recordingThread !=null && isDetectionOn){
                               isDetectionOn = false;

                       public void onDone(String s) {
                           if(recordingThread != null && !isDetectionOn && checkHotwordPref()) {
                               isDetectionOn = true;

                       public void onError(String s) {
                           if(recordingThread != null && !isDetectionOn && checkHotwordPref()) {
                               isDetectionOn = true;

                   HashMap<String,String> ttsParams = new HashMap<String, String>();


So, the main thing required for implementation of Text to Speech alongside Hotword detection is a way to control stopping and resuming hotword detection when Text to speech is in process. For that we used UtteranceProgressListener of TextToSpeech class which makes it so easier to do the task we required. You may follow this same approach as well or if you have a better approach, open an issue here.


  1. Official Documentation of TextToSpeech
  2. Documentation of UtteranceProgressListener
  3. Blog link to Hotword Detection

Search Functionalities in SUSI Android App Using Android SearchView Widget

Searching is a common feature that is required in most applications. But the problem in implementing searching functionality is that there is no common way to do that. People fight over whose way is best to implement search functionality. In this blog post we’ll be looking at how search functionality works in SUSI Android App and how is it implemented. We have used Android’s SearchView widget to do that. There are many other ways to do so but this one is best suited for our requirements. Let’s see how it works.

UI Components used for Searching

1. Search icon (magnifying glass icon)

In the action bar, you can see a small icon. Clicking on the icon initiates search.

2. Edit text

An Obvious requirement is an edit test to enter search query.

3. Up and Down arrow keys

Required to search through the whole app. Simply use the up and down arrow keys to navigate through the app and find out each occurrence of the word you want to search.








4. Cross Button

Last but not the least, a close or cross button to close the search action.


We have used Android’s inbuilt Widget SearchView. According to official android documentation

A widget that provides a user interface for the user to enter a search query and submit a request to a search provider. Shows a list of query suggestions or results, if available, and allows the user to pick a suggestion or result to launch into.

This widget makes searching a lot easier. It provides all methods and listeners which are actually required for searching. Let’s cover them one by one.

  1. Starting the search: searchView.setOnSearchClickListener Listener simply activates when a user clicks on search icon in the toolbar. Do all your work which needs to be done at the starting of the search like, hiding some other UI elements of doing an animation inside the listener
  1. Stop the Search: searchView.setOnCloseListener Listener gets activated when a user clicks on the cross icon to close the search. Add all the code snippet you want which is needed to be executed when the search is closed inside this like maybe notify the adapter about data set changes or closing the database etc.
  1.  Searching a query:  searchView.setOnQueryTextListener Listener overrides 2 methods:

3.1 onQueryTextSubmit: As the name suggests, this method is called when the query to be searched is submitted.

3.2 onQueryTextChange: This method is called when query you are writing changes.

We, basically wanted same thing to happen if user has submitted the query or if he is still typing and that is to take the query at that particular moment, find it in database and highlight it. So, chatPresenter.onSearchQuerySearched(query) this method is called in both onQueryTextSubmit and onQueryTextSubmit  to do that.

 searchView.setOnQueryTextListener(object : SearchView.OnQueryTextListener {
      override fun onQueryTextSubmit(query: String): Boolean {
           //Handle Search Query
           recyclerAdapter.query = query
           return false

       override fun onQueryTextChange(newText: String): Boolean {
           if (TextUtils.isEmpty(newText)) {
               recyclerAdapter.highlightMessagePosition = -1
               if (!editText.isFocused) {
           } else {
               recyclerAdapter.query = newText
           return false
   return true
  1. Finding query in database: Now we have a query to be searched, we can just use a database operation to do that. The below code snippet finds all the messages which has the query present in it and work on it. If the query is not found, it simply displays a toast saying “Not found”
override fun onSearchQuerySearched(query: String) {
   results = databaseRepository.getSearchResults(query)
   offset = 1
   if (results.size > 0) {
       chatView?.searchMovement(results[results.size - offset].id.toInt())
   } else {

This is the database operation.

override fun getSearchResults(query: String): RealmResults<ChatMessage> {
   return realm.where(,
           query, Case.INSENSITIVE).findAll()

  1. Highlighting the part of message: Now, we know which message has the query, we just want to highlight it with a bright color to display the result. For that, we used SpannableString to highlight a part of complete string.
String text = chatTextView.getText().toString();
SpannableString modify = new SpannableString(text);
Pattern pattern = Pattern.compile(query, Pattern.CASE_INSENSITIVE);
Matcher matcher = pattern.matcher(modify);
while (matcher.find()) {
   int startIndex = matcher.start();
   int endIndex = matcher.end();
   modify.setSpan(new BackgroundColorSpan(Color.parseColor("#ffff00")), startIndex, endIndex, Spannable.SPAN_EXCLUSIVE_EXCLUSIVE);


The whole point of this blog post was to educate about SearchView widget of android and how it makes it easy to search queries. All the methods you need are already implemented. You just need to call them and add database operation.


  1. The link to official android documentation explaining about different methods in SearchView Class
  2. Another tutorial about SearchView

Hotword Detection in SUSI Android App using Snowboy

Hotword Detection is as cool as it sounds. What exactly is hotword detection? Hotword detection is a feature in which a device gets activated when it listens to a specific word or a phrase. You must have said “OK Google” or “Hey Cortana” or “Siri” or “Alexa” at least once in your lifetime. These all are hotwords which trigger the specific action attached to them. That specific action can be anything.

Implementing hotword detection from scratch in SUSI Android is not an easy task. You have to define language model, train the model and do various other processes before implementing it in Android. In short, not feasible to implement that along with the code of our Android app. There are many open source projects on hotword detection and speech recognition. They already have done what we need and we can make use of it. One such project is Snowboy. According to Snowboy GitHub repo “Snowboy is a DNN based hotword and wake word detection toolkit.”

Img src:

In SUSI Android App, we have used Snowboy for hotword detection with hotword as “susi” (pronounced as ‘suzi’). In this blog, I will tell you how Hotword detection is implemented in SUSI Android app. So, you can just follow the steps and you will be able to implement it in your application too or if you want to contribute in SUSI android app, it may help you a little in knowing the codebase better.

Pre Processing before Implementation

1. Generating Hotword Model

The start of implementation of hotword detection begins with creating a hotword model from snowboy website . Just log in and search for susi and then train it by saying “susi” thrice and download the susi.pmdl file.

There are two types of models:

  1. .pmdl : Personal Model
  2. .umdl : Universal Model

The personal model is specifically trained for you and is instantly available for you to download once you train the hotword by your voice. On the other hand, the Universal model is trained by minimum 500 hundred people and is only available once it is trained. So, we are going to use personal model for now since training of universal model is not yet completed.

Img src:

2. Adding some predefined native binary files in your app.

Once you have downloaded the susi.pmdl file and you need to copy some already written native binary file in your app.

    1. In your assets folder, make a directory named snowboy and add your downloaded susi.pmdl file along with this file in it.
    2. Copy this folder and add it in your  /app/src/main/java folder as it is. These are autogenerated swig files. So, don’t change it unless you know what you are doing.
    3. Also, create a new folder in your /app/src/main folder called jniLibs and add these files to it.

Implementation in SUSI Android App

Check out the implementation of Hotword detection in SUSI Android App here

You now have everything ready. Now you just need to implement some code in your MainActivity and you are good to go. Initiate Hotword detection in the app by using below written code snippet. The call initHotword() method from your onCreate method in MainActivity. Make sure you have given permission to RECORD_AUDIO and WRITE_EXTERNAL_STORAGE.

private void initHotword() {
   if (ActivityCompat.checkSelfPermission(this,
           Manifest.permission.WRITE_EXTERNAL_STORAGE) == PackageManager.PERMISSION_GRANTED &&
           Manifest.permission.RECORD_AUDIO) == PackageManager.PERMISSION_GRANTED) {

       recordingThread = new RecordingThread(new Handler() {
           public void handleMessage(Message msg) {
               MsgEnum message = MsgEnum.getMsgEnum(msg.what);
               switch(message) {
                   case MSG_ACTIVE:
                   case MSG_INFO:
                   case MSG_VAD_SPEECH:
                   case MSG_VAD_NOSPEECH:
                   case MSG_ERROR:
       }, new AudioDataSaver());

In the above code under case, MSG_ACTIVE add your code which needs to be executed once hotword is detected. Now, hotword detection is initiated and you just have to start recording and stop recording whenever you want.

To start recording : (can be written in onStart() method)

if(recordingThread !=null && !isDetectionOn) {
    isDetectionOn = true;

To stop recording : (can be written in onStop() method)

if(recordingThread !=null && isDetectionOn){
   isDetectionOn = false;

So, this this is the simple process to add hotword detection feature in your app.

Though this is great and will word wonderfully for you but it has a little problem and a solution which is discussed below.

Problems for now and future steps

As mentioned above, the susi.pmdl model is a personal model and only trained for your voice. So, it will work wonderfully for you and for people with a similar voice like you but not for everyone. There are two ways to make this solve this problem:

  1. Replace susi.pmdl with susi.umdl: This is very easy way to fix this problem. But to get the susi.umdl file at least 500 people have to train susi hotword on snowboy website. Once that target is achieved, you can download susi.umdl file and replace it with susi.pmdl file and you are good to go.
  2. Train and obtain susi.pmdl file for each user: Right now you trained and used your own personal model for everyone but there is a way you can use every user’s own personal model in the app. Now hotword will be trained by the user himself in the app. Snowboy provides a training API to train the hotword and getting personal model. Quoting snowboy docs “You can define truly customized hotword for each of your end customer. Just ask them to say the hotword 3 times and a model will be trained on the fly!”


  1. Official docs of Snowboy hotword detection
  2. Code for snowboy hotword detection. It contains readme files which give very nice description about the project.
  3. This readme file of snowboy which guides on how to set up snowboy in an android app

Getting user Location in SUSI Android App and using it for various SUSI Skills

Using user location in skills is a very common phenomenon among various personal assistant like Google Assistant, Siri, Cortana etc. SUSI is no different. SUSI has various skills which uses user’s current location to implement skills. Though skills like “restaurant nearby” or “hotels nearby” are still under process but skills like “Where am I” works perfectly indicating SUSI has all basic requirements to create more advance skills in near future.

So let’s learn about how the SUSI Android App gets location of a user and sends it to SUSI Server where it is used to implement various location based skills.

Sources to find user location in an Android app

There are three sources from which android app gets users location :

  1. GPS
  2. Network
  3. Public IP Address

All three of these have various advantages and disadvantages. The SUSI Android app uses cleverly each of them to always get user location so that it can be used anytime and anywhere.

Some factors for comparison of these three sources :

Factors GPS Network IP Address
Source Satellites Wifi/Cell Tower Public IP address of user’s mobile
Accuracy Most Accurate (20ft) Moderately Accurate (200ft) Least Accurate (5000+ ft)
Requirements GPS in mobile Wifi or sim card Internet connection
Time taken to give location Takes long time to get location Fastest way to get location Fast enough (depends on internet speed)
Battery Consumption High Medium Low
Permission Required User permission required User permission required No permission required
Location Factor Works in outdoors. Does not work near tall buildings Works everywhere Works everywhere

Implementation of location finding feature in SUSI Android App

SUSI Android app very cleverly uses all the advantages of each location finding source to get most accurate location, consume less power and find location in any scenario.

The /susi/chat.json endpoint of SUSI API requires following 7 parameters :

Sno. Parameter Type Requirement
1 q String Compulsory
2 timezoneOffset int Optional
3 longitude double Optional
4 latitude double Optional
5 geosource String Optional
6 language Language  code Optional
7 access_token String Optional

In this blog we will be talking about latitude , longitude and geosource. So, we need these three things to pass as parameters for location related skills. Let’s see how we do that.

Finding location using IP Address: At the starting of app, user location is found by making an API call to . This results in following JSON response having a field “loc” giving location of user (latitude and longitude.

  "ip": "YOUR_IP_ADDRESS",
  "city": "YOUR_CITY",
  "region": "YOUR_REGION",
  "country": "YOUR_COUNTRY_CODE",
  "org": "YOUR_ISP"

By this way we got latitude, longitude and geosource will be “ip” . We find location using IP address only once the app is started because there is no need of finding it again and again as making network calls takes time and drains battery.

So, now we have user’s location but this is not accurate. So, we will now proceed to check if we can find location using network is more accurate than location using IP address.

Finding location using Network Service Provider : To actually use the network provider and find out location requires ACCESS_COARSE_LOCATION permission from user which can be asked during the run time. Also, the location can only be found out using this if user has his location setting is enabled. So, we check following condition.

if (ActivityCompat.checkSelfPermission(mContext, Manifest.permission.ACCESS_COARSE_LOCATION) == PackageManager.PERMISSION_GRANTED) {

If permission is granted by user to find location using network provider, we use following code snippet to find location. It updates location of user after every 5 minutes or 10 meters (whichever is achieved first).

locationManager.requestLocationUpdates(LocationManager.NETWORK_PROVIDER, 5 * 60 * 1000, 10, this);
location = locationManager.getLastKnownLocation(LocationManager.NETWORK_PROVIDER);
if (location != null) {
   source = "network";
   canGetLocation = true;
   latitude = location.getLatitude();
   longitude = location.getLongitude();

So, whenever we are about to send query to SUSI Server, we take location from Network services, thus updating previous values of latitude, longitude and geosource (found using IP address) with the new values (found using Network Provider), provided the user has granted permission. So, we now have location is from Network Provider which is more accurate than location from IP address. Now we will check if we can find location from GPS or not.

Finding location using GPS Service Provider : Finding location from GPS Provider is almost same as Network Provider. To find location using GPS Provider user must give  ACCESS_FINE_LOCATION permission. We just check if GPS of user is enabled and user has given permission to use GPS and also if GPS can actually provide location. Sometimes, GPS can not provide location because user is indoor. In that cases we leave location from GPS.

So, now if we update previous values of latitude, longitude and geosource (found using Network Provider) with the new values (found using GPS Provider) and send query to SUSI Server.


To send location to server for location skills, we need latitude, longitude and geosource. We first find these 3 things using IP address (no that accurate). So, geosource will be “ip” for now. Then check if we can find values using network provider. If yes, we update those 3 values with the ones got from network Provider (more accurate). Geosource will change to “network”. Finally, we check if we can find values using GPS provider. If yes, we update those 3 values with the ones got from GPS Provider (most accurate). Geosource will change to “gps”. So, by this way we can find location of user in any circumstance possible. If you want to use location in your app too. Just follow the above steps and you are good to go.



Implementing Real Time Speech to Text in SUSI Android App

Android has a very interesting feature built in called Speech to Text. As the name suggests, it converts user’s speech into text. So, android provides this feature in form of a recognizer intent. Just the below 4 line code is needed to start recognizer intent to capture speech and convert into text. This was implemented earlier in SUSI Android App too.  Simple, isn’t it?

Intent intent = new Intent(RecognizerIntent.ACTION_RECOGNIZE_SPEECH);
startActivityForResult(intent, REQ_CODE_SPEECH_INPUT);

But there is a problem here. It popups an alert dialog box to take a speech input like this :

Although this is great and working but there were many disadvantages of using this :

  1. It breaks connection between user and app.
  2. When user speaks, it first listens to complete sentence and then convert it into text. So, user was not able to see what the first word of sentence was until he says the complete sentence.
  3. There is no way for user to check if the sentence he is speaking is actually converting correctly to speech or not. He only gets to know about it when the sentence is completed.
  4. User can not cancel the conversion of Speech to text if some word is converted wrongly because the user doesn’t even get to know if the word is converted wrongly. It is kind of suspense that a whether the conversion is right or not.

So, now the main challenge was how can we improve the Speech to Text and implement it something like Google now?  So, what google now does is, it converts speech to text alongside taking input making it a real time speech to text convertor. So, basically if you speak second word of sentence, the first word is already converted and displayed. You can also cancel the conversion of rest of the sentence if the first word of sentence is converted wrongly.

I searched this problem a lot on internet, read various blogs and stackoverflow answers, read Documentation of Speech to Text until I finally came across SpeechRecognizer class and RecognitionListener. The plan was to use object of SpeechRecognizer class with RecognizerIntent and RecognitionListener to generate partial results.

So, I just added this line to the the above code snippet of recognizer intent.


And introduced SpeechRecognizer class and RecognitionListener like this

 SpeechRecognizer recognizer = SpeechRecognizer

RecognitionListener listener = new RecognitionListener() {
   public void onResults(Bundle results) {
   public void onError(int error) {
   public void onPartialResults(Bundle partialResults) {
    //The main method required. Just keep printing results in this.

Now combine SpeechRecognizer class with RecognizerIntent and RecognitionListener with these two lines.


Everything is now set up. You can now use result in onPartialResults method and keep displaying partial results. So, if you want to speak “My name is…” , “My” will be displayed on screen by the time you speak “name” and “My name” will be displayed on screen by the time you speak “is” . Isn’t it cool?

This is how Realtime Speech to Text is implemented in SUSI Android App. If you want to implement this in your app too, just follow the above steps and you are good to go. So, the results look like these.