Android M1: meeting recording via foreground service

- core/recording: MeetingRecorder (AudioRecord at the device's native rate, WAV on disk + 60 s rolling window resampled to 16 kHz), streaming WavWriter, thread-safe RollingWindow, linear resampler; WavReader moved here from the app - app: RecorderService (foreground, type microphone) with ongoing notification and StateFlow state; Record/Stop UI with timer and level meter; runtime permission flow; finished recordings auto-load for transcription - launcher icon (mic + waveform, matching the desktop brand) and notification glyph - fix real M0 bug caught by the new unit tests: WavReader parsed 16-bit fmt fields (audioFormat/channels/bitsPerSample) with 32-bit reads - JVM tests: WAV round trip, native-rate header, resampler, rolling window — 5/5 green; assembleDebug and aapt2 APK checks pass - validated on-device on a Fairphone 6 (Android 16): model download, engine load, recording and transcription all working
2026-09-07 11:31:29 +02:00
parent e5ec95a5fb
commit 8e079467fa
19 changed files with 803 additions and 45 deletions
@@ -32,6 +32,7 @@ android {
 }

 dependencies {
+    implementation(project(":core:recording"))
    implementation(project(":core:whisper"))
    implementation(libs.androidx.core.ktx)
    implementation(libs.androidx.activity.compose)
@@ -1,10 +1,16 @@
 <?xml version="1.0" encoding="utf-8"?>
 <manifest xmlns:android="http://schemas.android.com/apk/res/android">

-    <!-- Only used to download Whisper GGML models from Hugging Face. -->
+    <!-- Model downloads from Hugging Face. -->
    <uses-permission android:name="android.permission.INTERNET" />
+    <!-- Meeting recording (foreground service, type microphone). -->
+    <uses-permission android:name="android.permission.RECORD_AUDIO" />
+    <uses-permission android:name="android.permission.FOREGROUND_SERVICE" />
+    <uses-permission android:name="android.permission.FOREGROUND_SERVICE_MICROPHONE" />
+    <uses-permission android:name="android.permission.POST_NOTIFICATIONS" />

    <application
+        android:icon="@mipmap/ic_launcher"
        android:label="MeetRec"
        android:theme="@android:style/Theme.Material.Light.NoActionBar"
        android:allowBackup="true">
@@ -17,5 +23,10 @@
                <category android:name="android.intent.category.LAUNCHER" />
            </intent-filter>
        </activity>
+
+        <service
+            android:name=".RecorderService"
+            android:exported="false"
+            android:foregroundServiceType="microphone" />
    </application>
 </manifest>
@@ -1,6 +1,9 @@
 package com.meetrec.android

+import android.Manifest
 import android.content.Intent
+import android.content.pm.PackageManager
+import android.os.Build
 import android.os.Bundle
 import android.provider.OpenableColumns
 import androidx.activity.ComponentActivity
@@ -16,25 +19,29 @@ import androidx.compose.foundation.layout.padding
 import androidx.compose.foundation.lazy.LazyColumn
 import androidx.compose.foundation.lazy.items
 import androidx.compose.material3.Button
+import androidx.compose.material3.ButtonDefaults
 import androidx.compose.material3.CircularProgressIndicator
 import androidx.compose.material3.DropdownMenuItem
 import androidx.compose.material3.ExperimentalMaterial3Api
 import androidx.compose.material3.ExposedDropdownMenuBox
 import androidx.compose.material3.ExposedDropdownMenuDefaults
+import androidx.compose.material3.LinearProgressIndicator
 import androidx.compose.material3.MaterialTheme
 import androidx.compose.material3.OutlinedButton
 import androidx.compose.material3.OutlinedTextField
 import androidx.compose.material3.Text
 import androidx.compose.runtime.Composable
+import androidx.compose.runtime.LaunchedEffect
+import androidx.compose.runtime.collectAsState
 import androidx.compose.runtime.getValue
 import androidx.compose.runtime.mutableStateOf
 import androidx.compose.runtime.remember
 import androidx.compose.runtime.rememberCoroutineScope
 import androidx.compose.runtime.setValue
-import androidx.compose.ui.Alignment
 import androidx.compose.ui.Modifier
 import androidx.compose.ui.platform.LocalContext
 import androidx.compose.ui.unit.dp
+import com.meetrec.core.recording.WavReader
 import com.meetrec.core.whisper.WhisperEngine
 import kotlinx.coroutines.Dispatchers
 import kotlinx.coroutines.launch
@@ -42,9 +49,8 @@ import kotlinx.coroutines.withContext
 import java.io.File

 /**
- * M0 screen: proves the whisper.cpp engine on device.
- * Pick a model (downloads from Hugging Face), load it, pick a WAV file,
- * transcribe, share the text. Recording arrives in M1.
+ * M1 screen: record a meeting (foreground service), then transcribe the
+ * recording on device. Live transcript arrives with M2.
 */
 class MainActivity : ComponentActivity() {
    override fun onCreate(savedInstanceState: Bundle?) {
@@ -72,6 +78,8 @@ fun MeetRecScreen() {
    val modelsDir = remember { File(context.filesDir, "models") }
    val modelFile = ModelDownloader.modelFile(model, modelsDir)

+    val recState by RecorderService.state.collectAsState()
+
    fun <T> runBusy(block: suspend () -> T) {
        if (busy) return
        busy = true
@@ -86,8 +94,61 @@ fun MeetRecScreen() {
        }
    }

+    // A finished recording is auto-loaded for transcription.
+    LaunchedEffect(recState) {
+        val fin = recState as? RecordingState.Finished ?: return@LaunchedEffect
+        runBusy {
+            val samples = withContext(Dispatchers.IO) {
+                fin.file.inputStream().use { WavReader.read(it) }
+            }
+            wavName = fin.file.name
+            wavSamples = samples
+            segments = emptyList()
+            status = "Recorded ${fmtMs(fin.durationMs)} at ${fin.sampleRate} Hz" +
+                " — ready to transcribe"
+        }
+    }
+    (recState as? RecordingState.Failed)?.let {
+        LaunchedEffect(it) { status = "Recording failed: ${it.message}" }
+    }
+
+    fun startRecording() = context.startForegroundService(
+        Intent(context, RecorderService::class.java)
+            .setAction(RecorderService.ACTION_START),
+    )
+
+    fun stopRecording() = context.startService(
+        Intent(context, RecorderService::class.java)
+            .setAction(RecorderService.ACTION_STOP),
+    )
+
+    val permLauncher = rememberLauncherForActivityResult(
+        ActivityResultContracts.RequestMultiplePermissions(),
+    ) { grants ->
+        if (grants[Manifest.permission.RECORD_AUDIO] == true) {
+            startRecording()
+        } else {
+            status = "Microphone permission is required to record"
+        }
+    }
+
+    fun toggleRecording() {
+        if (recState is RecordingState.Recording) {
+            stopRecording()
+            return
+        }
+        val wanted = buildList {
+            add(Manifest.permission.RECORD_AUDIO)
+            if (Build.VERSION.SDK_INT >= 33) add(Manifest.permission.POST_NOTIFICATIONS)
+        }
+        val missing = wanted.filter {
+            context.checkSelfPermission(it) != PackageManager.PERMISSION_GRANTED
+        }
+        if (missing.isEmpty()) startRecording() else permLauncher.launch(missing.toTypedArray())
+    }
+
    val pickWav = rememberLauncherForActivityResult(
-        ActivityResultContracts.OpenDocument()
+        ActivityResultContracts.OpenDocument(),
    ) { uri ->
        if (uri == null) return@rememberLauncherForActivityResult
        runBusy {
@@ -101,7 +162,7 @@ fun MeetRecScreen() {
            wavName = name
            wavSamples = samples
            segments = emptyList()
-            status = "$name: ${samples.size / WavReader.TARGET_RATE} s @ 16 kHz"
+            status = "$name: ${samples.size / 16000} s @ 16 kHz"
        }
    }

@@ -172,12 +233,41 @@ fun MeetRecScreen() {
            ) { Text("Load engine") }
        }

-        // wav + transcribe actions
+        // recording
+        val recording = recState as? RecordingState.Recording
+        Button(
+            onClick = { toggleRecording() },
+            enabled = !busy || recording != null,
+            colors = if (recording == null) {
+                ButtonDefaults.buttonColors()
+            } else {
+                ButtonDefaults.buttonColors(
+                    containerColor = MaterialTheme.colorScheme.error,
+                )
+            },
+            modifier = Modifier.fillMaxWidth(),
+        ) {
+            Text(
+                if (recording == null) {
+                    "\u25CF  Record"
+                } else {
+                    "\u25A0  Stop  (${fmtMs(recording.elapsedMs)})"
+                },
+            )
+        }
+        if (recording != null) {
+            LinearProgressIndicator(
+                progress = { recording.level.coerceIn(0f, 1f) },
+                modifier = Modifier.fillMaxWidth(),
+            )
+        }
+
+        // transcribe actions
        Row(horizontalArrangement = Arrangement.spacedBy(12.dp)) {
            OutlinedButton(
                onClick = { pickWav.launch(arrayOf("audio/*", "application/octet-stream")) },
                enabled = !busy,
-            ) { Text(if (wavName == null) "Pick WAV file" else "Pick another WAV") }
+            ) { Text(if (wavName == null) "Pick WAV file" else wavName ?: "WAV") }
            Button(
                onClick = {
                    val samples = wavSamples
@@ -194,8 +284,8 @@ fun MeetRecScreen() {
                            }
                            segments = segs
                            val secs = (System.currentTimeMillis() - start) / 1000.0
-                            status = "${segs.size} segments in $secs s (audio " +
-                                "${samples.size / WavReader.TARGET_RATE} s)"
+                            status = "${segs.size} segments in $secs s " +
+                                "(audio ${samples.size / 16000} s)"
                        }
                    }
                },
@@ -214,7 +304,7 @@ fun MeetRecScreen() {
                            .setType("text/plain")
                            .putExtra(Intent.EXTRA_TEXT, text),
                        "Share transcript",
-                    )
+                    ),
                )
            }) { Text("Share transcript") }
        }
@@ -238,5 +328,8 @@ fun MeetRecScreen() {

 private fun fmtMs(ms: Long): String {
    val total = ms / 1000
-    return "%02d:%02d".format(total / 60, total % 60)
+    val h = total / 3600
+    val m = (total % 3600) / 60
+    val s = total % 60
+    return if (h > 0) "%d:%02d:%02d".format(h, m, s) else "%02d:%02d".format(m, s)
 }
@@ -0,0 +1,147 @@
+package com.meetrec.android
+
+import android.Manifest
+import android.app.Notification
+import android.app.NotificationChannel
+import android.app.NotificationManager
+import android.app.PendingIntent
+import android.app.Service
+import android.content.Intent
+import android.content.pm.PackageManager
+import android.content.pm.ServiceInfo
+import android.os.IBinder
+import androidx.core.app.NotificationCompat
+import com.meetrec.core.recording.MeetingRecorder
+import java.io.File
+import java.text.SimpleDateFormat
+import java.util.Date
+import java.util.Locale
+import kotlinx.coroutines.flow.MutableStateFlow
+import kotlinx.coroutines.flow.StateFlow
+
+/** Recording state published to the UI (collect RecorderService.state). */
+sealed interface RecordingState {
+    data object Idle : RecordingState
+    data class Recording(val file: File, val elapsedMs: Long, val level: Float) :
+        RecordingState
+
+    data class Finished(
+        val file: File,
+        val durationMs: Long,
+        val sampleRate: Int,
+    ) : RecordingState
+
+    data class Failed(val message: String) : RecordingState
+}
+
+/**
+ * Foreground service (type: microphone) that owns the MeetingRecorder,
+ * mirroring the desktop recorder lifecycle. Keeps recording alive while
+ * the app is backgrounded and posts an ongoing notification.
+ */
+class RecorderService : Service() {
+
+    companion object {
+        const val ACTION_START = "com.meetrec.android.action.START"
+        const val ACTION_STOP = "com.meetrec.android.action.STOP"
+        private const val CHANNEL_ID = "meetrec_recording"
+        private const val NOTIFICATION_ID = 1
+
+        val state: StateFlow<RecordingState> = MutableStateFlow(RecordingState.Idle)
+        private val _state = state as MutableStateFlow<RecordingState>
+    }
+
+    private var recorder: MeetingRecorder? = null
+    private var currentFile: File? = null
+
+    override fun onBind(intent: Intent?): IBinder? = null
+
+    override fun onStartCommand(intent: Intent?, flags: Int, startId: Int): Int {
+        when (intent?.action) {
+            ACTION_STOP -> recorder?.stop()
+            else -> startRecording()
+        }
+        return START_NOT_STICKY
+    }
+
+    private fun startRecording() {
+        if (recorder != null) return
+        if (checkSelfPermission(Manifest.permission.RECORD_AUDIO)
+            != PackageManager.PERMISSION_GRANTED
+        ) {
+            _state.value = RecordingState.Failed("Microphone permission is missing")
+            stopSelf()
+            return
+        }
+
+        createChannel()
+        startForeground(
+            NOTIFICATION_ID,
+            notification("Recording meeting…"),
+            ServiceInfo.FOREGROUND_SERVICE_TYPE_MICROPHONE,
+        )
+
+        val dir = File(filesDir, "recordings").apply { mkdirs() }
+        val stamp = SimpleDateFormat("yyyyMMdd-HHmmss", Locale.US).format(Date())
+        currentFile = File(dir, "meeting-$stamp.wav")
+
+        val rec = MeetingRecorder(currentFile!!, listener)
+        recorder = rec
+        _state.value = RecordingState.Recording(currentFile!!, 0, 0f)
+        rec.start()
+    }
+
+    private val listener = object : MeetingRecorder.Listener {
+        override fun onTick(elapsedMs: Long, level: Float) {
+            currentFile?.let { _state.value = RecordingState.Recording(it, elapsedMs, level) }
+        }
+
+        override fun onFinish(file: File, durationMs: Long, sampleRate: Int) {
+            _state.value = RecordingState.Finished(file, durationMs, sampleRate)
+            cleanup()
+        }
+
+        override fun onError(message: String) {
+            _state.value = RecordingState.Failed(message)
+            cleanup()
+        }
+    }
+
+    private fun cleanup() {
+        stopForeground(STOP_FOREGROUND_REMOVE)
+        stopSelf()
+        recorder = null
+    }
+
+    override fun onDestroy() {
+        recorder?.stop()
+        recorder = null
+        super.onDestroy()
+    }
+
+    private fun createChannel() {
+        val nm = getSystemService(NotificationManager::class.java)
+        nm.createNotificationChannel(
+            NotificationChannel(
+                CHANNEL_ID,
+                "Meeting recording",
+                NotificationManager.IMPORTANCE_LOW,
+            ),
+        )
+    }
+
+    private fun notification(text: String): Notification {
+        val pi = PendingIntent.getActivity(
+            this, 0,
+            Intent(this, MainActivity::class.java),
+            PendingIntent.FLAG_IMMUTABLE,
+        )
+        return NotificationCompat.Builder(this, CHANNEL_ID)
+            .setSmallIcon(R.drawable.ic_mic)
+            .setContentTitle("MeetRec")
+            .setContentText(text)
+            .setOngoing(true)
+            .setContentIntent(pi)
+            .build()
+    }
+}
@@ -1,107 +0,0 @@
-package com.meetrec.android
-
-import java.io.InputStream
-
-/**
- * Reads a PCM WAV file into 16 kHz mono float32 — the format Whisper wants.
- *
- * Supports 16-bit PCM and 32-bit IEEE-float WAVs of any channel count and
- * sample rate (multi-channel is averaged, rates are linearly resampled).
- * Pure Kotlin so it is unit-testable on the JVM.
- */
-object WavReader {
-
-    const val TARGET_RATE = 16000
-
-    fun read(input: InputStream): FloatArray {
-        val data = input.readBytes()
-        require(data.size > 44) { "file too small to be a WAV" }
-        require(String(data, 0, 4) == "RIFF" && String(data, 8, 4) == "WAVE") {
-            "not a RIFF/WAVE file"
-        }
-
-        var channels = 0
-        var sampleRate = 0
-        var audioFormat = 0
-        var bitsPerSample = 0
-        var samples: FloatArray? = null
-
-        var pos = 12
-        while (pos + 8 <= data.size) {
-            val id = String(data, pos, 4)
-            val size = leInt(data, pos + 4)
-            val body = pos + 8
-            when (id) {
-                "fmt " -> {
-                    audioFormat = leInt(data, body)
-                    channels = leInt(data, body + 2)
-                    sampleRate = leInt(data, body + 4)
-                    bitsPerSample = leInt(data, body + 14)
-                }
-                "data" -> {
-                    val fmt = audioFormat
-                    require(fmt == 1 || fmt == 3) {
-                        "unsupported WAV audio format $fmt (need PCM or IEEE float)"
-                    }
-                    require(bitsPerSample == 16 || bitsPerSample == 32) {
-                        "unsupported bit depth $bitsPerSample (need 16 or 32)"
-                    }
-                    require(channels > 0 && sampleRate > 0) { "malformed fmt chunk" }
-                    val len = minOf(size, data.size - body)
-                    samples = decode(data, body, len, fmt, bitsPerSample, channels)
-                }
-            }
-            pos = body + size + (size and 1)   // chunks are word-aligned
-        }
-
-        val mono = samples ?: error("no data chunk found")
-        return if (sampleRate == TARGET_RATE) mono else resample(mono, sampleRate, TARGET_RATE)
-    }
-
-    private fun decode(
-        data: ByteArray, body: Int, len: Int,
-        audioFormat: Int, bits: Int, channels: Int,
-    ): FloatArray {
-        val bytesPerSample = bits / 8
-        val frames = len / (bytesPerSample * channels)
-        val out = FloatArray(frames)
-        for (f in 0 until frames) {
-            var acc = 0f
-            for (c in 0 until channels) {
-                val o = body + (f * channels + c) * bytesPerSample
-                acc += when {
-                    audioFormat == 1 && bits == 16 ->
-                        ((data[o].toInt() and 0xFF) or (data[o + 1].toInt() shl 8)) / 32768f
-                    audioFormat == 1 && bits == 32 ->
-                        leInt(data, o) / 2147483648f
-                    audioFormat == 3 && bits == 32 ->
-                        Float.fromBits(leInt(data, o))
-                    else -> 0f
-                }
-            }
-            out[f] = acc / channels
-        }
-        return out
-    }
-
-    private fun resample(x: FloatArray, from: Int, to: Int): FloatArray {
-        if (x.isEmpty()) return x
-        val n = (x.size.toDouble() * to / from).toInt().coerceAtLeast(1)
-        val out = FloatArray(n)
-        val step = from.toDouble() / to
-        for (i in 0 until n) {
-            val t = i * step
-            val i0 = t.toInt().coerceAtMost(x.size - 1)
-            val i1 = (i0 + 1).coerceAtMost(x.size - 1)
-            val frac = (t - i0).toFloat()
-            out[i] = x[i0] * (1f - frac) + x[i1] * frac
-        }
-        return out
-    }
-
-    private fun leInt(b: ByteArray, off: Int): Int =
-        (b[off].toInt() and 0xFF) or
-            ((b[off + 1].toInt() and 0xFF) shl 8) or
-            ((b[off + 2].toInt() and 0xFF) shl 16) or
-            ((b[off + 3].toInt() and 0xFF) shl 24)
-}
@@ -0,0 +1,31 @@
+<?xml version="1.0" encoding="utf-8"?>
+<vector xmlns:android="http://schemas.android.com/apk/res/android"
+    android:width="108dp"
+    android:height="108dp"
+    android:viewportWidth="108"
+    android:viewportHeight="108">
+    <!-- MeetRec glyph: mic + waveform (voice becoming transcript). -->
+    <path
+        android:fillColor="#FFFFFF"
+        android:pathData="M45.6,24.6h16.8a8.4,8.4 0 0 1 8.4,8.4v8.4a8.4,8.4 0 0 1 -8.4,8.4h-16.8a8.4,8.4 0 0 1 -8.4,-8.4v-8.4a8.4,8.4 0 0 1 8.4,-8.4z" />
+    <path
+        android:strokeColor="#FFFFFF"
+        android:strokeWidth="5.4"
+        android:strokeLineCap="round"
+        android:pathData="M37.2,42.6 v4.8 a16.8,16.8 0 0 0 33.6,0 v-4.8" />
+    <path
+        android:fillColor="#FFFFFF"
+        android:pathData="M34.8,71.4h4.8a2.4,2.4 0 0 1 2.4,2.4v3.6a2.4,2.4 0 0 1 -2.4,2.4h-4.8a2.4,2.4 0 0 1 -2.4,-2.4v-3.6a2.4,2.4 0 0 1 2.4,-2.4z" />
+    <path
+        android:fillColor="#FFFFFF"
+        android:pathData="M43.2,68.4h4.8a2.4,2.4 0 0 1 2.4,2.4v9.6a2.4,2.4 0 0 1 -2.4,2.4h-4.8a2.4,2.4 0 0 1 -2.4,-2.4v-9.6a2.4,2.4 0 0 1 2.4,-2.4z" />
+    <path
+        android:fillColor="#FFFFFF"
+        android:pathData="M51.6,70.8h4.8a2.4,2.4 0 0 1 2.4,2.4v4.8a2.4,2.4 0 0 1 -2.4,2.4h-4.8a2.4,2.4 0 0 1 -2.4,-2.4v-4.8a2.4,2.4 0 0 1 2.4,-2.4z" />
+    <path
+        android:fillColor="#FFFFFF"
+        android:pathData="M59.9,68.4h4.8a2.4,2.4 0 0 1 2.4,2.4v9.6a2.4,2.4 0 0 1 -2.4,2.4h-4.8a2.4,2.4 0 0 1 -2.4,-2.4v-9.6a2.4,2.4 0 0 1 2.4,-2.4z" />
+    <path
+        android:fillColor="#FFFFFF"
+        android:pathData="M68.4,71.4h4.8a2.4,2.4 0 0 1 2.4,2.4v3.6a2.4,2.4 0 0 1 -2.4,2.4h-4.8a2.4,2.4 0 0 1 -2.4,-2.4v-3.6a2.4,2.4 0 0 1 2.4,-2.4z" />
+</vector>
@@ -0,0 +1,10 @@
+<?xml version="1.0" encoding="utf-8"?>
+<vector xmlns:android="http://schemas.android.com/apk/res/android"
+    android:width="24dp"
+    android:height="24dp"
+    android:viewportWidth="24"
+    android:viewportHeight="24">
+    <path
+        android:fillColor="#FFFFFF"
+        android:pathData="M12,14c1.66,0 3,-1.34 3,-3V5c0,-1.66 -1.34,-3 -3,-3S9,3.34 9,5v6C9,12.66 10.34,14 12,14zM17,11c0,2.76 -2.24,5 -5,5s-5,-2.24 -5,-5H5c0,3.53 2.61,6.43 6,6.92V21h2v-3.08c3.39,-0.49 6,-3.39 6,-6.92H17z" />
+</vector>
@@ -0,0 +1,5 @@
+<?xml version="1.0" encoding="utf-8"?>
+<adaptive-icon xmlns:android="http://schemas.android.com/apk/res/android">
+    <background android:drawable="@color/launcher_bg" />
+    <foreground android:drawable="@drawable/ic_launcher_foreground" />
+</adaptive-icon>
@@ -0,0 +1,5 @@
+<?xml version="1.0" encoding="utf-8"?>
+<adaptive-icon xmlns:android="http://schemas.android.com/apk/res/android">
+    <background android:drawable="@color/launcher_bg" />
+    <foreground android:drawable="@drawable/ic_launcher_foreground" />
+</adaptive-icon>
@@ -0,0 +1,4 @@
+<?xml version="1.0" encoding="utf-8"?>
+<resources>
+    <color name="launcher_bg">#C0392B</color>
+</resources>