feat(chat): support non-Google speech-to-text providers (fixes #1038) (#1042)

app/src/main/AndroidManifest.xml

@@ -13,6 +13,12 @@
     <uses-permission android:name="android.permission.RECORD_AUDIO" />
     <uses-permission android:name="android.permission.REQUEST_INSTALL_PACKAGES" />
 
+    <queries>
+        <intent>
+            <action android:name="android.speech.action.RECOGNIZE_SPEECH" />
+        </intent>
+    </queries>
+
     <application
         android:name=".HermesControlApp"
         android:allowBackup="false"

app/src/main/java/com/m57/hermescontrol/ui/chat/ChatScreen.kt

@@ -2,12 +2,12 @@ package com.m57.hermescontrol.ui.chat
 
 import android.Manifest
 import android.app.Activity
+import android.content.ActivityNotFoundException
 import android.content.Intent
 import android.content.pm.PackageManager
 import android.net.Uri
 import android.provider.OpenableColumns
 import android.speech.RecognizerIntent
-import android.speech.SpeechRecognizer
 import android.util.Log
 import androidx.activity.compose.rememberLauncherForActivityResult
 import androidx.activity.result.contract.ActivityResultContracts
@@ -343,21 +343,18 @@ fun ChatScreen(
         ) { granted ->
             ExternalActivityLifecycleGuard.externalActivityReturned()
             if (granted) {
-                if (SpeechRecognizer.isRecognitionAvailable(context)) {
-                    val intent =
-                        Intent(RecognizerIntent.ACTION_RECOGNIZE_SPEECH).apply {
-                            putExtra(
-                                RecognizerIntent.EXTRA_LANGUAGE_MODEL,
-                                RecognizerIntent.LANGUAGE_MODEL_FREE_FORM,
-                            )
-                            putExtra(
-                                RecognizerIntent.EXTRA_PROMPT,
-                                micListeningPrompt,
-                            )
-                        }
+                if (SpeechInputHelper.isSpeechInputAvailable(context)) {
+                    val intent = SpeechInputHelper.createSpeechIntent(micListeningPrompt)
                     isListening = true
                     launchExternalActivity {
-                        speechLauncher.launch(intent)
+                        try {
+                            speechLauncher.launch(intent)
+                        } catch (_: ActivityNotFoundException) {
+                            isListening = false
+                            scrollScope.launch {
+                                snackbarHostState.showSnackbar(sttNotAvailableMsg)
+                            }
+                        }
                     }
                 } else {
                     scrollScope.launch {
@@ -781,21 +778,18 @@ fun ChatScreen(
                             Manifest.permission.RECORD_AUDIO,
                         ) == PackageManager.PERMISSION_GRANTED
                     ) {
-                        if (SpeechRecognizer.isRecognitionAvailable(context)) {
-                            val intent =
-                                Intent(RecognizerIntent.ACTION_RECOGNIZE_SPEECH).apply {
-                                    putExtra(
-                                        RecognizerIntent.EXTRA_LANGUAGE_MODEL,
-                                        RecognizerIntent.LANGUAGE_MODEL_FREE_FORM,
-                                    )
-                                    putExtra(
-                                        RecognizerIntent.EXTRA_PROMPT,
-                                        micListeningPrompt,
-                                    )
-                                }
+                        if (SpeechInputHelper.isSpeechInputAvailable(context)) {
+                            val intent = SpeechInputHelper.createSpeechIntent(micListeningPrompt)
                             isListening = true
                             launchExternalActivity {
-                                speechLauncher.launch(intent)
+                                try {
+                                    speechLauncher.launch(intent)
+                                } catch (_: ActivityNotFoundException) {
+                                    isListening = false
+                                    scrollScope.launch {
+                                        snackbarHostState.showSnackbar(sttNotAvailableMsg)
+                                    }
+                                }
                             }
                         } else {
                             scrollScope.launch {

app/src/main/java/com/m57/hermescontrol/ui/chat/SpeechInputHelper.kt

@@ -0,0 +1,54 @@
+package com.m57.hermescontrol.ui.chat
+
+import android.content.Context
+import android.content.Intent
+import android.content.pm.PackageManager
+import android.os.Build
+import android.speech.RecognizerIntent
+import android.speech.SpeechRecognizer
+
+object SpeechInputHelper {
+    /**
+     * Checks if speech recognition is available on this device through any registered provider,
+     * including standard Android [SpeechRecognizer] services or offline/third-party apps
+     * that handle [RecognizerIntent.ACTION_RECOGNIZE_SPEECH] (e.g. FUTO Voice Input).
+     */
+    fun isSpeechInputAvailable(
+        context: Context,
+        isRecognizerServiceAvailable: (Context) -> Boolean = { ctx ->
+            runCatching { SpeechRecognizer.isRecognitionAvailable(ctx) }.getOrDefault(false)
+        },
+    ): Boolean {
+        val intent = Intent(RecognizerIntent.ACTION_RECOGNIZE_SPEECH)
+        val activities =
+            if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.TIRAMISU) {
+                runCatching {
+                    context.packageManager.queryIntentActivities(
+                        intent,
+                        PackageManager.ResolveInfoFlags.of(0L),
+                    )
+                }.getOrDefault(emptyList())
+            } else {
+                @Suppress("DEPRECATION")
+                runCatching {
+                    context.packageManager.queryIntentActivities(intent, 0)
+                }.getOrDefault(emptyList())
+            }
+        return activities.isNotEmpty() || isRecognizerServiceAvailable(context)
+    }
+
+    /**
+     * Builds the standard speech recognition intent with language model and prompt.
+     */
+    fun createSpeechIntent(prompt: String): Intent =
+        Intent(RecognizerIntent.ACTION_RECOGNIZE_SPEECH).apply {
+            putExtra(
+                RecognizerIntent.EXTRA_LANGUAGE_MODEL,
+                RecognizerIntent.LANGUAGE_MODEL_FREE_FORM,
+            )
+            putExtra(
+                RecognizerIntent.EXTRA_PROMPT,
+                prompt,
+            )
+        }
+}

app/src/test/java/com/m57/hermescontrol/ui/chat/SpeechInputHelperTest.kt

@@ -0,0 +1,92 @@
+package com.m57.hermescontrol.ui.chat
+
+import android.content.Context
+import android.content.Intent
+import android.content.pm.PackageManager
+import android.content.pm.ResolveInfo
+import io.mockk.every
+import io.mockk.mockk
+import io.mockk.mockkConstructor
+import io.mockk.unmockkConstructor
+import org.junit.After
+import org.junit.Assert.assertFalse
+import org.junit.Assert.assertNotNull
+import org.junit.Assert.assertTrue
+import org.junit.Before
+import org.junit.Test
+
+class SpeechInputHelperTest {
+    @Before
+    fun setUp() {
+        mockkConstructor(Intent::class)
+        every { anyConstructed<Intent>().setAction(any()) } answers { self as Intent }
+        every { anyConstructed<Intent>().putExtra(any<String>(), any<String>()) } answers { self as Intent }
+    }
+
+    @After
+    fun tearDown() {
+        unmockkConstructor(Intent::class)
+    }
+
+    @Test
+    fun isSpeechInputAvailable_returnsTrue_whenPackageManagerFindsActivities() {
+        val context = mockk<Context>()
+        val packageManager = mockk<PackageManager>()
+        val resolveInfo = mockk<ResolveInfo>()
+
+        every { context.packageManager } returns packageManager
+        every {
+            packageManager.queryIntentActivities(any<Intent>(), any<Int>())
+        } returns listOf(resolveInfo)
+
+        val result =
+            SpeechInputHelper.isSpeechInputAvailable(
+                context = context,
+                isRecognizerServiceAvailable = { false },
+            )
+        assertTrue(result)
+    }
+
+    @Test
+    fun isSpeechInputAvailable_returnsTrue_whenRecognitionServiceAvailableEvenIfNoActivities() {
+        val context = mockk<Context>()
+        val packageManager = mockk<PackageManager>()
+
+        every { context.packageManager } returns packageManager
+        every {
+            packageManager.queryIntentActivities(any<Intent>(), any<Int>())
+        } returns emptyList()
+
+        val result =
+            SpeechInputHelper.isSpeechInputAvailable(
+                context = context,
+                isRecognizerServiceAvailable = { true },
+            )
+        assertTrue(result)
+    }
+
+    @Test
+    fun isSpeechInputAvailable_returnsFalse_whenNoActivitiesAndNoRecognitionService() {
+        val context = mockk<Context>()
+        val packageManager = mockk<PackageManager>()
+
+        every { context.packageManager } returns packageManager
+        every {
+            packageManager.queryIntentActivities(any<Intent>(), any<Int>())
+        } returns emptyList()
+
+        val result =
+            SpeechInputHelper.isSpeechInputAvailable(
+                context = context,
+                isRecognizerServiceAvailable = { false },
+            )
+        assertFalse(result)
+    }
+
+    @Test
+    fun createSpeechIntent_constructsIntentWithActionAndPrompt() {
+        val prompt = "Listening..."
+        val intent = SpeechInputHelper.createSpeechIntent(prompt)
+        assertNotNull(intent)
+    }
+}