Toak/Core/DaemonService.cs

using System.Net.Sockets;
using Toak.Configuration;
using Toak.Api;
using Toak.Core.Interfaces;
using Toak.Audio;
using Toak.IO;

namespace Toak.Core;

public static class DaemonService
{
    public static string GetSocketPath()
    {
        return Constants.Paths.GetSocketPath();
    }

    private static FileStream? _lockFile;

    public static async Task StartAsync(bool verbose)
    {
        var lockPath = Constants.Paths.DaemonLockFile;
        try
        {
            Directory.CreateDirectory(Path.GetDirectoryName(lockPath)!);
            _lockFile = new FileStream(lockPath, FileMode.OpenOrCreate, FileAccess.ReadWrite, FileShare.None);
        }
        catch (IOException)
        {
            Console.WriteLine("Toak daemon is already running.");
            return;
        }

        Logger.Verbose = verbose;
        var socketPath = GetSocketPath();

        if (File.Exists(socketPath))
        {
            try { File.Delete(socketPath); } catch { }
        }

        var configManager = new ConfigManager();
        var config = configManager.LoadConfig();
        if (config.WhisperProvider == "groq" && string.IsNullOrWhiteSpace(config.GroqApiKey))
        {
            Console.WriteLine("Groq API Key is not configured for Whisper. Run 'toak onboard'.");
            return;
        }
        if (config.WhisperProvider == "fireworks" && string.IsNullOrWhiteSpace(config.FireworksApiKey))
        {
            Console.WriteLine("Fireworks API Key is not configured for Whisper. Run 'toak onboard'.");
            return;
        }

        var stateTracker = new StateTracker();
        var notifications = new Notifications();

        ISpeechClient speechClient = config.WhisperProvider == "fireworks"
            ? new OpenAiCompatibleClient(config.FireworksApiKey, "https://api.fireworks.ai/inference/v1/")
            : new OpenAiCompatibleClient(config.GroqApiKey);

        ILlmClient llmClient = config.LlmProvider switch
        {
            "together" => new OpenAiCompatibleClient(config.TogetherApiKey, "https://api.together.xyz/v1/", config.ReasoningEffort),
            "cerebras" => new OpenAiCompatibleClient(config.CerebrasApiKey, "https://api.cerebras.ai/v1/", config.ReasoningEffort),
            "fireworks" => new OpenAiCompatibleClient(config.FireworksApiKey, "https://api.fireworks.ai/inference/v1/", config.ReasoningEffort),
            _ => new OpenAiCompatibleClient(config.GroqApiKey, "https://api.groq.com/openai/v1/", config.ReasoningEffort)
        };

        IAudioRecorder recorder = config.AudioBackend == "ffmpeg"
            ? new FfmpegAudioRecorder(stateTracker, notifications)
            : new PipewireAudioRecorder(stateTracker, notifications);

        var orchestrator = new TranscriptionOrchestrator(
            speechClient,
            llmClient,
            configManager,
            recorder,
            notifications,
            new TextInjector(notifications),
            new HistoryManager(),
            new ClipboardManager(notifications),
            stateTracker
        );

        using var socket = new Socket(AddressFamily.Unix, SocketType.Stream, ProtocolType.Unspecified);
        var endPoint = new UnixDomainSocketEndPoint(socketPath);

        try
        {
            socket.Bind(endPoint);
            socket.Listen(10);
            Logger.LogDebug($"Daemon listening on {socketPath}");
            Console.WriteLine($"Toak daemon started, listening on {socketPath}");

            while (true)
            {
                var client = await socket.AcceptAsync();
                _ = Task.Run(() => HandleClientAsync(client, orchestrator, stateTracker));
            }
        }
        catch (Exception ex)
        {
            Logger.LogDebug($"Daemon error: {ex.Message}");
        }
        finally
        {
            if (File.Exists(socketPath))
            {
                File.Delete(socketPath);
            }
        }
    }

    private static async Task HandleClientAsync(Socket client, ITranscriptionOrchestrator orchestrator, IRecordingStateTracker stateTracker)
    {
        try
        {
            var buffer = new byte[3];
            var bytesRead = await client.ReceiveAsync(buffer, SocketFlags.None);
            if (bytesRead > 0)
            {
                var cmd = buffer[0];
                var pipeToStdout = bytesRead > 1 && buffer[1] == 1;
                var copyToClipboard = bytesRead > 2 && buffer[2] == 1;

                if (cmd == 1) // START
                {
                    await orchestrator.ProcessStartRecordingAsync();
                }
                else if (cmd == 2) // STOP
                {
                    await orchestrator.ProcessStopRecordingAsync(client, pipeToStdout, copyToClipboard);
                }
                else if (cmd == 3) // ABORT
                {
                    orchestrator.ProcessAbortAsync();
                }
                else if (cmd == 4) // TOGGLE
                {
                    if (stateTracker.IsRecording())
                        await orchestrator.ProcessStopRecordingAsync(client, pipeToStdout, copyToClipboard);
                    else
                        await orchestrator.ProcessStartRecordingAsync();
                }
                else if (cmd == 5) // STATUS
                {
                    var json = pipeToStdout; // buffer[1] == 1 is json
                    var isRecording = stateTracker.IsRecording();
                    var stateStr = isRecording ? "Recording" : "Idle";

                    if (json)
                    {
                        var start = stateTracker.GetRecordingStartTime();
                        double durationMs = 0;
                        if (isRecording && start.HasValue)
                        {
                            durationMs = (DateTime.UtcNow - start.Value).TotalMilliseconds;
                        }
                        var jsonStr = $"{{\"state\": \"{stateStr}\", \"duration\": {Math.Round(durationMs)}}}";
                        await client.SendAsync(System.Text.Encoding.UTF8.GetBytes(jsonStr), SocketFlags.None);
                    }
                    else
                    {
                        await client.SendAsync(System.Text.Encoding.UTF8.GetBytes(stateStr), SocketFlags.None);
                    }
                }
            }
        }
        catch (Exception ex)
        {
            Logger.LogDebug($"HandleClient error: {ex.Message}");
        }
        finally
        {
            client.Close();
        }
    }
}