examples : add whisper.swiftui demo app (ggml-org#308)

* Add SwiftUI demo project.

* Add -DGGML_USE_ACCELERATE

Author: Digipom

examples/whisper.swiftui/README.md

@@ -0,0 +1,12 @@
+A sample SwiftUI app using [whisper.cpp](https://github.com/ggerganov/whisper.cpp/) to do voice-to-text transcriptions.
+See also: [whisper.objc](https://github.com/ggerganov/whisper.cpp/tree/master/examples/whisper.objc).
+
+To use:
+
+1. Select a model from the [whisper.cpp repository](https://github.com/ggerganov/whisper.cpp/tree/master/models).[^1]
+2. Add the model to "whisper.swiftui.demo/Resources/models" via Xcode.
+3. Select a sample audio file (for example, [jfk.wav](https://github.com/ggerganov/whisper.cpp/raw/master/samples/jfk.wav)).
+4. Add the model to "whisper.swiftui.demo/Resources/samples" via Xcode.
+5. Select the "release" build configuration under "Run", then deploy and run to your device.
+
+[^1]: I recommend the tiny, base or small models for running on an iOS device.

examples/whisper.swiftui/whisper.cpp.swift/LibWhisper.swift

@@ -0,0 +1,70 @@
+import Foundation
+
+enum WhisperError: Error {
+    case couldNotInitializeContext
+}
+
+// Meet Whisper C++ constraint: Don't access from more than one thread at a time.
+actor WhisperContext {
+    private var context: OpaquePointer
+    
+    init(context: OpaquePointer) {
+        self.context = context
+    }
+    
+    deinit {
+        whisper_free(context)
+    }
+    
+    func fullTranscribe(samples: [Float]) {
+        // Leave 2 processors free (i.e. the high-efficiency cores).
+        let maxThreads = max(1, min(8, cpuCount() - 2))
+        print("Selecting \(maxThreads) threads")
+        var params = whisper_full_default_params(WHISPER_SAMPLING_GREEDY)
+        "en".withCString { en in
+            // Adapted from whisper.objc
+            params.print_realtime = true
+            params.print_progress = false
+            params.print_timestamps = true
+            params.print_special = false
+            params.translate = false
+            params.language = en
+            params.n_threads = Int32(maxThreads)
+            params.offset_ms = 0
+            params.no_context = true
+            params.single_segment = false
+            
+            whisper_reset_timings(context)
+            print("About to run whisper_full")
+            samples.withUnsafeBufferPointer { samples in
+                if (whisper_full(context, params, samples.baseAddress, Int32(samples.count)) != 0) {
+                    print("Failed to run the model")
+                } else {
+                    whisper_print_timings(context)
+                }
+            }
+        }
+    }
+    
+    func getTranscription() -> String {
+        var transcription = ""
+        for i in 0..<whisper_full_n_segments(context) {
+            transcription += String.init(cString: whisper_full_get_segment_text(context, i))
+        }
+        return transcription
+    }
+    
+    static func createContext(path: String) throws -> WhisperContext {
+        let context = whisper_init(path)
+        if let context {
+            return WhisperContext(context: context)
+        } else {
+            print("Couldn't load model at \(path)")
+            throw WhisperError.couldNotInitializeContext
+        }
+    }
+}
+
+fileprivate func cpuCount() -> Int {
+    ProcessInfo.processInfo.processorCount
+}

examples/whisper.swiftui/whisper.cpp.swift/WhisperCppDemo-Bridging-Header.h

@@ -0,0 +1,4 @@
+//
+//  Use this file to import your target's public headers that you would like to expose to Swift.
+//
+#import "whisper.h"

examples/whisper.swiftui/whisper.swiftui.demo/Models/WhisperState.swift

@@ -0,0 +1,162 @@
+import Foundation
+import SwiftUI
+import AVFoundation
+
+@MainActor
+class WhisperState: NSObject, ObservableObject, AVAudioRecorderDelegate {
+    @Published var isModelLoaded = false
+    @Published var messageLog = ""
+    @Published var canTranscribe = false
+    @Published var isRecording = false
+    
+    private var whisperContext: WhisperContext?
+    private let recorder = Recorder()
+    private var recordedFile: URL? = nil
+    private var audioPlayer: AVAudioPlayer?
+    
+    private var modelUrl: URL? {
+        Bundle.main.url(forResource: "ggml-tiny.en", withExtension: "bin", subdirectory: "models")
+    }
+    
+    private var sampleUrl: URL? {
+        Bundle.main.url(forResource: "jfk", withExtension: "wav", subdirectory: "samples")
+    }
+    
+    private enum LoadError: Error {
+        case couldNotLocateModel
+    }
+    
+    override init() {
+        super.init()
+        do {
+            try loadModel()
+            canTranscribe = true
+        } catch {
+            print(error.localizedDescription)
+            messageLog += "\(error.localizedDescription)\n"
+        }
+    }
+    
+    private func loadModel() throws {
+        messageLog += "Loading model...\n"
+        if let modelUrl {
+            whisperContext = try WhisperContext.createContext(path: modelUrl.path())
+            messageLog += "Loaded model \(modelUrl.lastPathComponent)\n"
+        } else {
+            messageLog += "Could not locate model\n"
+        }
+    }
+    
+    func transcribeSample() async {
+        if let sampleUrl {
+            await transcribeAudio(sampleUrl)
+        } else {
+            messageLog += "Could not locate sample\n"
+        }
+    }
+    
+    private func transcribeAudio(_ url: URL) async {
+        if (!canTranscribe) {
+            return
+        }
+        guard let whisperContext else {
+            return
+        }
+        
+        do {
+            canTranscribe = false
+            messageLog += "Reading wave samples...\n"
+            let data = try readAudioSamples(url)
+            messageLog += "Transcribing data...\n"
+            await whisperContext.fullTranscribe(samples: data)
+            let text = await whisperContext.getTranscription()
+            messageLog += "Done: \(text)\n"
+        } catch {
+            print(error.localizedDescription)
+            messageLog += "\(error.localizedDescription)\n"
+        }
+        
+        canTranscribe = true
+    }
+    
+    private func readAudioSamples(_ url: URL) throws -> [Float] {
+        stopPlayback()
+        try startPlayback(url)
+        return try decodeWaveFile(url)
+    }
+    
+    func toggleRecord() async {
+        if isRecording {
+            await recorder.stopRecording()
+            isRecording = false
+            if let recordedFile {
+                await transcribeAudio(recordedFile)
+            }
+        } else {
+            requestRecordPermission { granted in
+                if granted {
+                    Task {
+                        do {
+                            self.stopPlayback()
+                            let file = try FileManager.default.url(for: .documentDirectory, in: .userDomainMask, appropriateFor: nil, create: true)
+                                .appending(path: "output.wav")
+                            try await self.recorder.startRecording(toOutputFile: file, delegate: self)
+                            self.isRecording = true
+                            self.recordedFile = file
+                        } catch {
+                            print(error.localizedDescription)
+                            self.messageLog += "\(error.localizedDescription)\n"
+                            self.isRecording = false
+                        }
+                    }
+                }
+            }
+        }
+    }
+    
+    private func requestRecordPermission(response: @escaping (Bool) -> Void) {
+#if os(macOS)
+        response(true)
+#else
+        AVAudioSession.sharedInstance().requestRecordPermission { granted in
+            response(granted)
+        }
+#endif
+    }
+    
+    private func startPlayback(_ url: URL) throws {
+        audioPlayer = try AVAudioPlayer(contentsOf: url)
+        audioPlayer?.play()
+    }
+    
+    private func stopPlayback() {
+        audioPlayer?.stop()
+        audioPlayer = nil
+    }
+    
+    // MARK: AVAudioRecorderDelegate
+    
+    nonisolated func audioRecorderEncodeErrorDidOccur(_ recorder: AVAudioRecorder, error: Error?) {
+        if let error {
+            Task {
+                await handleRecError(error)
+            }
+        }
+    }
+    
+    private func handleRecError(_ error: Error) {
+        print(error.localizedDescription)
+        messageLog += "\(error.localizedDescription)\n"
+        isRecording = false
+    }
+    
+    nonisolated func audioRecorderDidFinishRecording(_ recorder: AVAudioRecorder, successfully flag: Bool) {
+        Task {
+            await onDidFinishRecording()
+        }
+    }
+    
+    private func onDidFinishRecording() {
+        isRecording = false
+    }
+}

examples/whisper.swiftui/whisper.swiftui.demo/Supporting files/Assets.xcassets/AccentColor.colorset/Contents.json

@@ -0,0 +1,11 @@
+{
+  "colors" : [
+    {
+      "idiom" : "universal"
+    }
+  ],
+  "info" : {
+    "author" : "xcode",
+    "version" : 1
+  }
+}

examples/whisper.swiftui/whisper.swiftui.demo/Supporting files/Assets.xcassets/AppIcon.appiconset/Contents.json

@@ -0,0 +1,63 @@
+{
+  "images" : [
+    {
+      "idiom" : "universal",
+      "platform" : "ios",
+      "size" : "1024x1024"
+    },
+    {
+      "idiom" : "mac",
+      "scale" : "1x",
+      "size" : "16x16"
+    },
+    {
+      "idiom" : "mac",
+      "scale" : "2x",
+      "size" : "16x16"
+    },
+    {
+      "idiom" : "mac",
+      "scale" : "1x",
+      "size" : "32x32"
+    },
+    {
+      "idiom" : "mac",
+      "scale" : "2x",
+      "size" : "32x32"
+    },
+    {
+      "idiom" : "mac",
+      "scale" : "1x",
+      "size" : "128x128"
+    },
+    {
+      "idiom" : "mac",
+      "scale" : "2x",
+      "size" : "128x128"
+    },
+    {
+      "idiom" : "mac",
+      "scale" : "1x",
+      "size" : "256x256"
+    },
+    {
+      "idiom" : "mac",
+      "scale" : "2x",
+      "size" : "256x256"
+    },
+    {
+      "idiom" : "mac",
+      "scale" : "1x",
+      "size" : "512x512"
+    },
+    {
+      "idiom" : "mac",
+      "scale" : "2x",
+      "size" : "512x512"
+    }
+  ],
+  "info" : {
+    "author" : "xcode",
+    "version" : 1
+  }
+}

examples/whisper.swiftui/whisper.swiftui.demo/Supporting files/Assets.xcassets/Contents.json

@@ -0,0 +1,6 @@
+{
+  "info" : {
+    "author" : "xcode",
+    "version" : 1
+  }
+}

examples/whisper.swiftui/whisper.swiftui.demo/Supporting files/Preview Content/Preview Assets.xcassets/Contents.json

@@ -0,0 +1,6 @@
+{
+  "info" : {
+    "author" : "xcode",
+    "version" : 1
+  }
+}

examples/whisper.swiftui/whisper.swiftui.demo/Supporting files/WhisperCppDemo.entitlements

@@ -0,0 +1,12 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+	<key>com.apple.security.app-sandbox</key>
+	<true/>
+	<key>com.apple.security.device.audio-input</key>
+	<true/>
+	<key>com.apple.security.files.user-selected.read-only</key>
+	<true/>
+</dict>
+</plist>

examples/whisper.swiftui/whisper.swiftui.demo/UI/ContentView.swift

@@ -0,0 +1,43 @@
+import SwiftUI
+import AVFoundation
+
+struct ContentView: View {
+    @StateObject var whisperState = WhisperState()
+    
+    var body: some View {
+        NavigationStack {
+            VStack {
+                HStack {
+                    Button("Transcribe", action: {
+                        Task {
+                            await whisperState.transcribeSample()
+                        }
+                    })
+                    .buttonStyle(.bordered)
+                    .disabled(!whisperState.canTranscribe)
+                    
+                    Button(whisperState.isRecording ? "Stop recording" : "Start recording", action: {
+                        Task {
+                            await whisperState.toggleRecord()
+                        }
+                    })
+                    .buttonStyle(.bordered)
+                    .disabled(!whisperState.canTranscribe)
+                }
+                
+                ScrollView {
+                    Text(verbatim: whisperState.messageLog)
+                        .frame(maxWidth: .infinity, alignment: .leading)
+                }
+            }
+            .navigationTitle("Whisper SwiftUI Demo")
+            .padding()
+        }
+    }
+}
+
+struct ContentView_Previews: PreviewProvider {
+    static var previews: some View {
+        ContentView()
+    }
+}