# Veg Sahayyak SDK

> Enterprise B2B Telematics, Background GPS Velocity Processing, and Jetpack Compose Speedometer SDK for Android.

Veg Sahayyak SDK (`com.vegsahayyak:speed-core` and `com.vegsahayyak:speed-ui`) is an Android library designed for commercial fleet management, logistics, delivery, and driver safety applications. It handles Fused Location Provider polling, Exponential Moving Average (EMA [REDACTED]) signal smoothing, Dual Schmitt Trigger Hysteresis (+[UPPER_BOUND] / -[LOWER_BOUND]) alert gating, and hardware-accelerated Compose gauges.

## Primary Documentation Links for LLMs

- [Getting Started & Installation](/getting-started/index.md): Gradle BOM dependency setup, Maven repository configuration, and quick start initialization.
- [Core Architecture](/architecture/index.md): Modular design, mathematical signal processing pipeline, and internal service locator.
- [Core Integration Guide](/guide/index.md): `SpeedAssistant` builder API, lifecycle management, Coroutines `StateFlow<SpeedData>`, and XML ComposeView interop.
- [Driving Profiles & Corridors](/guide/driving-profiles.md): Pre-configured vehicle limits (Highway, City, Trucks, Agriculture, Scooters, Bikes, Walking, School zones) and dynamic `CustomProfile`.
- [Voice Commands & AI NLU](/guide/voice-commands.md): Hands-free operation with `VoiceCommandButton`, offline `RegexVoiceCommandParser`, and Gemini / LLM NLU parser injection.
- [Floating PiP Overlay Widget](/guide/floating-overlay.md): `:speed-overlay` module with `PipWidget` and `PipWidgetConfig` hovering above Google Maps/Waze with `SYSTEM_ALERT_WINDOW`.
- [In-Vehicle Auto-Start](/guide/auto-start.md): `:speed-autostart` module with `AutoStartManager`, `AutoStartReceiver`, and zero-touch Activity Recognition (`IN_VEHICLE`).
- [Telematics Engine & Signal Processing](/guide/signal-processing.md): Mathematical deep-dive into discrete Exponential Moving Average (EMA) smoothing, GPS gating, and Dual Schmitt Trigger Hysteresis.
- [Permissions & Battery Optimization](/guide/permissions-and-battery.md): Runtime permissions (Android 10 - 15), `FOREGROUND_SERVICE_LOCATION`, and OEM battery killer guidelines.
- [Testing & QA Guide](/guide/testing.md): Synthetic fixtures (`FakeSpeedRepository`), Unit Testing with MockK & Turbine, Compose UI tests, and FusedLocation mock APIs.
- [Jetpack Compose UI & HUD Mode](/ui-components/index.md): `@Composable SpeedGauge` parameter reference, `HudDisplay` windshield reflection, `Modifier.hudMirror()`, `ThemeMode` (`SYSTEM`, `DARK_GRAY`, `OLED_BLACK`, `HIGH_CONTRAST`, `HUD_MIRRORED`), `SpeedUiColorScheme`, and TalkBack accessibility tokens.
- [Production Showcase](/showcase/index.md): Real-world reference implementation in the Speed Assistant consumer app (< 5 MB memory, 60fps Compose UI).
- [API Reference & KDoc](/api-reference/index.md): Complete specifications for `SpeedAssistant`, `SpeedColorScheme`, `SpeedData`, `SpeedUnit`, `TelematicsEngine`, `ISpeedRepository`, and `IPreferenceRepository`.
- [Troubleshooting & FAQ](/troubleshooting/index.md): Android 12 `ForegroundServiceStartNotAllowedException`, OEM aggressive background kill prevention, and telemetry simulation.

## Key Code Patterns for AI Assistants

### 1. SDK Initialization & Configuration
```kotlin
import com.vegsahayyak.speedcore.api.SpeedAssistant
import com.vegsahayyak.speedcore.api.DrivingProfile

class App : Application() {
    override fun onCreate() {
        super.onCreate()
        SpeedAssistant.with(this)
            .setDrivingProfile(DrivingProfile.CarHighway)
            .startIfEnabled()
    }
}
```

### 2. Observing Telematics Stream in Jetpack Compose
```kotlin
import androidx.compose.runtime.Composable
import androidx.compose.runtime.collectAsState
import androidx.compose.runtime.getValue
import com.vegsahayyak.speedcore.api.SpeedAssistant
import com.vegsahayyak.speedcore.data.location.SpeedData
import com.vegsahayyak.speedui.components.SpeedGauge
import com.vegsahayyak.speedui.components.ThemeMode

@Composable
fun DriverDashboardScreen() {
    val speedState by SpeedAssistant.observeFlow().collectAsState(initial = SpeedData.SpeedUnavailable)

    when (val state = speedState) {
        is SpeedData.SpeedChanged -> {
            SpeedGauge(
                speed = state.speedMps * 3.6f, // Convert m/s to km/h
                warningSpeed = 80f,
                minSpeedLimit = 40f,
                isTrackingActive = true,
                isOverLimit = state.isOverLimit,
                gpsAccuracyMeters = state.accuracyMeters,
                themeMode = ThemeMode.DARK_GRAY
            )
        }
        is SpeedData.LowAccuracy -> Text("GPS Signal Weak (±${state.reason})")
        is SpeedData.GPSDisabled -> Text("GPS Disabled")
        is SpeedData.SpeedUnavailable,
        is SpeedData.Disabled -> SpeedGauge(speed = 0f, warningSpeed = 80f, isTrackingActive = false)
    }
}
```

### 3. Unit Testing with FakeSpeedRepository
```kotlin
val fakeRepo = FakeSpeedRepository()
SpeedAssistant.with(context).useRepository(fakeRepo).start()

// Emit simulated speed
fakeRepo.emitSpeed(speedKmh = 95f, isOverLimit = true)
```

## Complete Unified Context
For the complete unformatted text reference containing all documentation in a single file, fetch:
- `/llms-full.txt`