Skip to content
This repository has been archived by the owner on Mar 15, 2022. It is now read-only.

Adding DeepSpeech w/ TFLite #9

Merged
merged 1 commit into from
Mar 18, 2019
Merged

Adding DeepSpeech w/ TFLite #9

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions app/build.gradle
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,9 @@ android {
versionCode 1
versionName "1.0"
testInstrumentationRunner "android.support.test.runner.AndroidJUnitRunner"
ndk {
abiFilters 'armeabi-v7a', 'arm64-v8a', 'x86_64'
}
}
buildTypes {
release {
Expand All @@ -20,6 +23,10 @@ android {
sourceCompatibility JavaVersion.VERSION_1_8
targetCompatibility JavaVersion.VERSION_1_8
}

lintOptions {
abortOnError false
}
}

dependencies {
Expand All @@ -28,6 +35,7 @@ dependencies {
implementation 'com.android.support:appcompat-v7:27.1.1'
implementation 'com.android.support.constraint:constraint-layout:1.1.2'
implementation 'com.jjoe64:graphview:4.2.2'
implementation 'net.lingala.zip4j:zip4j:1.3.2'
testImplementation 'junit:junit:4.12'
androidTestImplementation 'com.android.support.test:runner:1.0.2'
androidTestImplementation 'com.android.support.test.espresso:espresso-core:3.0.2'
Expand Down
119 changes: 117 additions & 2 deletions app/src/main/java/com/mozilla/speechapp/MainActivity.java
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,32 +1,58 @@
package com.mozilla.speechapp;

import android.Manifest;

import android.app.Activity;
import android.app.DownloadManager;

import android.content.pm.PackageManager;
import android.content.BroadcastReceiver;
import android.content.Context;
import android.content.Intent;
import android.content.IntentFilter;

import android.database.Cursor;

import android.support.annotation.NonNull;
import android.support.v4.app.ActivityCompat;
import android.support.v7.app.AppCompatActivity;

import android.net.Uri;

import android.os.Bundle;
import android.os.AsyncTask;

import android.util.Log;

import android.view.View;
import android.view.WindowManager;

import android.widget.Button;
import android.widget.CompoundButton;
import android.widget.EditText;
import android.widget.Switch;
import android.widget.Toast;

import com.jjoe64.graphview.GraphView;
import com.jjoe64.graphview.series.DataPoint;
import com.jjoe64.graphview.series.LineGraphSeries;

import com.mozilla.speechlibrary.ISpeechRecognitionListener;
import com.mozilla.speechlibrary.MozillaSpeechService;
import com.mozilla.speechlibrary.STTResult;
import com.mozilla.speechmodule.R;

import java.io.File;

import net.lingala.zip4j.core.ZipFile;

import static android.support.constraint.Constraints.TAG;

public class MainActivity extends AppCompatActivity implements ISpeechRecognitionListener, CompoundButton.OnCheckedChangeListener {

private static long sDownloadId;
private static DownloadManager sDownloadManager;

private MozillaSpeechService mMozillaSpeechService;
private GraphView mGraph;
private long mDtstart;
Expand All @@ -47,6 +73,7 @@ private void initialize() {
EditText txtProdutTag, txtLanguage;
Switch switchTranscriptions = findViewById(R.id.switchTranscriptions);
Switch switchSamples = findViewById(R.id.switchSamples);
Switch useDeepSpeech = findViewById(R.id.useDeepSpeech);

if (ActivityCompat.checkSelfPermission(this, Manifest.permission.RECORD_AUDIO)
!= PackageManager.PERMISSION_GRANTED) {
Expand Down Expand Up @@ -75,7 +102,12 @@ private void initialize() {
mSeries1.resetData(new DataPoint[0]);
mMozillaSpeechService.setLanguage(txtLanguage.getText().toString());
mMozillaSpeechService.setProductTag(txtProdutTag.getText().toString());
mMozillaSpeechService.start(getApplicationContext());
mMozillaSpeechService.setModelPath(getExternalFilesDir("models").getAbsolutePath());
if (mMozillaSpeechService.ensureModelInstalled()) {
mMozillaSpeechService.start(getApplicationContext());
} else {
maybeDownloadOrExtractModel(getExternalFilesDir("models").getAbsolutePath(), mMozillaSpeechService.getLanguageDir());
Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This logic is required on the app side because after checking with a few potential users they indeed want to be more in charge of securing / handling the selection and distribution of models.

}
} catch (Exception e) {
Log.d(TAG, e.getLocalizedMessage());
e.printStackTrace();
Expand All @@ -93,8 +125,10 @@ private void initialize() {

switchTranscriptions.setOnCheckedChangeListener(this);
switchSamples.setOnCheckedChangeListener(this);
useDeepSpeech.setOnCheckedChangeListener(this);
switchTranscriptions.toggle();
switchSamples.toggle();
useDeepSpeech.toggle();

mGraph = findViewById(R.id.graph);
mSeries1 = new LineGraphSeries<>(new DataPoint[0]);
Expand Down Expand Up @@ -154,8 +188,89 @@ public void removeListener() {
public void onCheckedChanged(CompoundButton buttonView, boolean isChecked) {
if (buttonView.equals(findViewById(R.id.switchTranscriptions))) {
mMozillaSpeechService.storeTranscriptions(isChecked);
} else {
} else if (buttonView.equals(findViewById(R.id.switchSamples))) {
mMozillaSpeechService.storeSamples(isChecked);
} else if (buttonView.equals(findViewById(R.id.useDeepSpeech))) {
mMozillaSpeechService.useDeepSpeech(isChecked);
}
}

private class AsyncUnzip extends AsyncTask<String, Void, Boolean> {

@Override
protected void onPreExecute() {
Toast noModel = Toast.makeText(getApplicationContext(), "Extracting downloaded model", Toast.LENGTH_LONG);
mPlain_text_input.append("Extracting downloaded model\n");
noModel.show();
}

@Override
protected Boolean doInBackground(String...params) {
String aZipFile = params[0], aRootModelsPath = params[1];
try {
ZipFile zf = new ZipFile(aZipFile);
zf.extractAll(aRootModelsPath);
} catch (Exception e) {
Log.d(TAG, e.getLocalizedMessage());
e.printStackTrace();
}

return (new File(aZipFile)).delete();
}

@Override
protected void onPostExecute(Boolean result) {
Button buttonStart = findViewById(R.id.button_start), buttonCancel = findViewById(R.id.button_cancel);
mMozillaSpeechService.start(getApplicationContext());
buttonStart.setEnabled(true);
buttonCancel.setEnabled(true);
}

}

public void maybeDownloadOrExtractModel(String aModelsPath, String aLang) {
String zipFile = aModelsPath + "/" + aLang + ".zip";
Uri modelZipURL = Uri.parse(mMozillaSpeechService.getModelDownloadURL());
Uri modelZipFile = Uri.parse("file://" + zipFile);

Button buttonStart = findViewById(R.id.button_start), buttonCancel = findViewById(R.id.button_cancel);
buttonStart.setEnabled(false);
buttonCancel.setEnabled(false);

BroadcastReceiver receiver = new BroadcastReceiver() {
@Override
public void onReceive(Context context, Intent intent) {
String action = intent.getAction();
if (DownloadManager.ACTION_DOWNLOAD_COMPLETE.equals(action)) {
long downloadId = intent.getLongExtra(DownloadManager.EXTRA_DOWNLOAD_ID, 0);
DownloadManager.Query query = new DownloadManager.Query();
query.setFilterById(downloadId);
Cursor c = sDownloadManager.query(query);
if (c.moveToFirst()) {
int columnIndex = c.getColumnIndex(DownloadManager.COLUMN_STATUS);
if (DownloadManager.STATUS_SUCCESSFUL == c.getInt(columnIndex)) {
Log.d(TAG, "Download successfull");

new AsyncUnzip().execute(zipFile, aModelsPath);
}
}
}
}
};

Toast noModel = Toast.makeText(getApplicationContext(), "No model has been found for language '" + aLang + "'. Triggering download ...", Toast.LENGTH_LONG);
mPlain_text_input.append("No model has been found for language '" + aLang + "'. Triggering download ...\n");
noModel.show();

sDownloadManager = (DownloadManager) getSystemService(Context.DOWNLOAD_SERVICE);
DownloadManager.Request request = new DownloadManager.Request(modelZipURL);
request.setTitle("DeepSpeech " + aLang);
request.setDescription("DeepSpeech Model");
request.setNotificationVisibility(DownloadManager.Request.VISIBILITY_VISIBLE_NOTIFY_COMPLETED);
request.setVisibleInDownloadsUi(false);
request.setDestinationUri(modelZipFile);
sDownloadId = sDownloadManager.enqueue(request);

getApplicationContext().registerReceiver(receiver, new IntentFilter(DownloadManager.ACTION_DOWNLOAD_COMPLETE));
}
}
18 changes: 14 additions & 4 deletions app/src/main/res/layout/activity_main.xml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,12 +49,21 @@
android:inputType="textMultiLine"
android:singleLine="false" />

<Switch
android:id="@+id/useDeepSpeech"
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:layout_alignParentEnd="true"
android:layout_below="@+id/switchTranscriptions"
android:text="Use DeepSpeech" />

<Switch
android:id="@+id/switchTranscriptions"
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:layout_alignParentEnd="true"
android:layout_centerVertical="true"
android:layout_below="@+id/switchSamples"
android:text="Store Transcriptions" />

<Switch
Expand All @@ -71,7 +80,7 @@
android:layout_height="wrap_content"
android:layout_above="@+id/plain_text_input"
android:layout_alignStart="@+id/graph"
android:layout_marginBottom="-71dp"
android:layout_marginBottom="-50dp"
android:ems="10"
android:inputType="textPersonName"
android:text="ProductTag" />
Expand All @@ -82,9 +91,10 @@
android:layout_height="wrap_content"
android:layout_above="@+id/plain_text_input"
android:layout_alignStart="@+id/graph"
android:layout_marginBottom="-120dp"
android:layout_marginBottom="-80dp"
android:ems="10"
android:inputType="textPersonName"
android:text="Language" />
android:text="eng" />

</RelativeLayout>
</android.support.constraint.ConstraintLayout>
</android.support.constraint.ConstraintLayout>
14 changes: 12 additions & 2 deletions mozillaspeechlibrary/build.gradle
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,9 +1,9 @@
apply plugin: 'com.android.library'
apply from: 'maven-push.gradle'

def versionMajor = 1
def versionMajor = 2
def versionMinor = 0
def versionPatch = 4
def versionPatch = 0

android {
compileSdkVersion 25
Expand All @@ -13,6 +13,10 @@ android {
versionCode versionMajor * 10000 + versionMinor * 100 + versionPatch
versionName "${versionMajor}.${versionMinor}.${versionPatch}"
testInstrumentationRunner "android.support.test.runner.AndroidJUnitRunner"

ndk {
abiFilters 'armeabi-v7a', 'arm64-v8a', 'x86_64'
}
}

buildTypes {
Expand All @@ -27,15 +31,21 @@ android {
path 'src/main/cpp/Android.mk'
}
}

compileOptions {
sourceCompatibility = 1.7
targetCompatibility = 1.7
}

lintOptions {
abortOnError false
}
}

dependencies {
implementation fileTree(include: ['*.jar'], dir: 'libs')
implementation 'com.loopj.android:android-async-http:1.4.9'
implementation 'org.mozilla.deepspeech:libdeepspeech:0.5.0-alpha.1@aar'
testImplementation 'junit:junit:4.12'
androidTestImplementation 'com.android.support.test:runner:1.0.2'
androidTestImplementation 'com.android.support.test.espresso:espresso-core:3.0.2'
Expand Down
Loading