Первый релиз WhisperKey

2026-07-29 14:46:12 +06:00
commit 5751743277
19 changed files with 1436 additions and 0 deletions
@@ -0,0 +1,183 @@
+import Foundation
+
+enum LocalWhisper {
+    private static let server = LocalWhisperServer()
+
+    static func validate(settings: AppSettings) throws {
+        guard FileManager.default.isExecutableFile(atPath: settings.localBinaryPath) else {
+            throw WhisperKeyError.missingDependency(settings.localBinaryPath)
+        }
+        guard FileManager.default.fileExists(atPath: settings.localModelPath) else {
+            throw WhisperKeyError.missingDependency(settings.localModelPath)
+        }
+        guard FileManager.default.fileExists(atPath: vadModelPath) else {
+            throw WhisperKeyError.missingDependency(vadModelPath)
+        }
+        guard FileManager.default.isExecutableFile(atPath: serverBinaryPath) else {
+            throw WhisperKeyError.missingDependency(serverBinaryPath)
+        }
+    }
+
+    private static let vadModelPath =
+        "\(NSHomeDirectory())/Library/Application Support/WhisperKey/ggml-silero-v6.2.0.bin"
+    private static let serverBinaryPath =
+        "\(NSHomeDirectory())/Library/Application Support/WhisperKey/whisper-server"
+
+    static func preload(settings: AppSettings) async throws {
+        try validate(settings: settings)
+        try await server.ensureRunning(
+            binaryPath: serverBinaryPath,
+            modelPath: settings.localModelPath,
+            vadModelPath: vadModelPath
+        )
+    }
+
+    static func transcribe(wav: URL, settings: AppSettings) async throws -> String {
+        try validate(settings: settings)
+        return try await server.transcribe(
+            wav: wav,
+            binaryPath: serverBinaryPath,
+            modelPath: settings.localModelPath,
+            vadModelPath: vadModelPath
+        )
+    }
+}
+
+private actor LocalWhisperServer {
+    private let baseURL = URL(string: "http://127.0.0.1:8178")!
+    private var process: Process?
+
+    func ensureRunning(binaryPath: String, modelPath: String, vadModelPath: String) async throws {
+        if await isHealthy() { return }
+        if process?.isRunning != true {
+            let process = Process()
+            process.executableURL = URL(fileURLWithPath: binaryPath)
+            process.arguments = [
+                "-m", modelPath,
+                "--host", "127.0.0.1",
+                "--port", "8178",
+                "-l", "ru",
+                "-nt",
+                "-nf",
+                "-sns",
+                "--no-speech-thold", "0.35",
+                "--vad",
+                "--vad-model", vadModelPath,
+                "--vad-threshold", "0.55",
+                "--vad-min-speech-duration-ms", "250",
+                "--vad-min-silence-duration-ms", "300",
+                "--vad-speech-pad-ms", "100",
+                "-t", "8",
+            ]
+            process.standardOutput = FileHandle.nullDevice
+            process.standardError = FileHandle.nullDevice
+            try process.run()
+            self.process = process
+        }
+
+        for _ in 0..<300 {
+            if await isHealthy() { return }
+            try await Task.sleep(for: .milliseconds(100))
+        }
+        throw WhisperKeyError.processFailed("Локальный Whisper не успел загрузить модель.")
+    }
+
+    func transcribe(
+        wav: URL,
+        binaryPath: String,
+        modelPath: String,
+        vadModelPath: String
+    ) async throws -> String {
+        try await ensureRunning(
+            binaryPath: binaryPath,
+            modelPath: modelPath,
+            vadModelPath: vadModelPath
+        )
+
+        let boundary = "WhisperKey-\(UUID().uuidString)"
+        var request = URLRequest(url: baseURL.appendingPathComponent("inference"))
+        request.httpMethod = "POST"
+        request.timeoutInterval = 60
+        request.setValue("multipart/form-data; boundary=\(boundary)", forHTTPHeaderField: "Content-Type")
+
+        var body = Data()
+        body.appendField("language", value: "ru", boundary: boundary)
+        body.appendField("temperature", value: "0", boundary: boundary)
+        body.appendField("no_speech_thold", value: "0.35", boundary: boundary)
+        body.appendField("response_format", value: "json", boundary: boundary)
+        body.appendFile(
+            "file",
+            filename: "dictation.wav",
+            mime: "audio/wav",
+            data: try Data(contentsOf: wav),
+            boundary: boundary
+        )
+        body.append("--\(boundary)--\r\n".data(using: .utf8)!)
+        request.httpBody = body
+
+        let (data, response) = try await URLSession.shared.data(for: request)
+        guard let http = response as? HTTPURLResponse, (200..<300).contains(http.statusCode) else {
+            throw WhisperKeyError.badResponse(String(data: data, encoding: .utf8) ?? "Ошибка локального Whisper.")
+        }
+        return try JSONDecoder().decode(TranscriptionResponse.self, from: data).text
+    }
+
+    private func isHealthy() async -> Bool {
+        var request = URLRequest(url: baseURL.appendingPathComponent("health"))
+        request.timeoutInterval = 0.5
+        guard let (_, response) = try? await URLSession.shared.data(for: request),
+              let http = response as? HTTPURLResponse else { return false }
+        return (200..<300).contains(http.statusCode)
+    }
+}
+
+enum OpenAITranscriber {
+    static func transcribe(wav: URL, settings: AppSettings) async throws -> String {
+        guard let key = settings.apiKey(), !key.isEmpty else { throw WhisperKeyError.missingAPIKey }
+        let boundary = "WhisperKey-\(UUID().uuidString)"
+        var request = URLRequest(url: URL(string: "https://api.openai.com/v1/audio/transcriptions")!)
+        request.httpMethod = "POST"
+        request.setValue("Bearer \(key)", forHTTPHeaderField: "Authorization")
+        request.setValue("multipart/form-data; boundary=\(boundary)", forHTTPHeaderField: "Content-Type")
+
+        var body = Data()
+        body.appendField("model", value: settings.cloudModel, boundary: boundary)
+        body.appendField("language", value: "ru", boundary: boundary)
+        body.appendField("temperature", value: "0", boundary: boundary)
+        body.appendField(
+            "prompt",
+            value: "Дословная русская транскрипция. Не добавляй слова, которых нет в аудио. Термины: \(settings.glossary)",
+            boundary: boundary
+        )
+        body.appendFile("file", filename: "dictation.wav", mime: "audio/wav", data: try Data(contentsOf: wav), boundary: boundary)
+        body.append("--\(boundary)--\r\n".data(using: .utf8)!)
+        request.httpBody = body
+
+        let (data, response) = try await URLSession.shared.data(for: request)
+        guard let http = response as? HTTPURLResponse, (200..<300).contains(http.statusCode) else {
+            let message = (try? JSONSerialization.jsonObject(with: data) as? [String: Any])
+                .flatMap { $0["error"] as? [String: Any] }?["message"] as? String
+            throw WhisperKeyError.badResponse(message ?? "OpenAI API вернул ошибку.")
+        }
+        let result = try JSONDecoder().decode(TranscriptionResponse.self, from: data)
+        return result.text
+    }
+}
+
+private struct TranscriptionResponse: Decodable { let text: String }
+
+private extension Data {
+    mutating func appendField(_ name: String, value: String, boundary: String) {
+        append("--\(boundary)\r\n".data(using: .utf8)!)
+        append("Content-Disposition: form-data; name=\"\(name)\"\r\n\r\n".data(using: .utf8)!)
+        append("\(value)\r\n".data(using: .utf8)!)
+    }
+
+    mutating func appendFile(_ name: String, filename: String, mime: String, data: Data, boundary: String) {
+        append("--\(boundary)\r\n".data(using: .utf8)!)
+        append("Content-Disposition: form-data; name=\"\(name)\"; filename=\"\(filename)\"\r\n".data(using: .utf8)!)
+        append("Content-Type: \(mime)\r\n\r\n".data(using: .utf8)!)
+        append(data)
+        append("\r\n".data(using: .utf8)!)
+    }
+}