.NET’te AI Desteğiyle Text-to-Speech (TTS) ve Speech-to-Text (STT) Operasyonları

using AI_TextToSpeech_And_SpeechToText_Example;
using Microsoft.Extensions.FileProviders;
using Microsoft.Net.Http.Headers;
using System.Net.Mime;
using System.Text;
using System.Text.Json;

var builder = WebApplication.CreateBuilder(args);

builder.Services.AddHttpClient("HuggingFaceAPI", httpClient =>
{
    httpClient.BaseAddress = new Uri("https://api-inference.huggingface.co");
    httpClient.DefaultRequestHeaders.Add(HeaderNames.Authorization, $"Bearer {builder.Configuration["HuggingFace:APIKey"]}");
});

builder.Services.AddCors(options =>
    options.AddDefaultPolicy(policy =>
    {
        policy.AllowAnyOrigin().AllowAnyHeader().AllowAnyMethod();
    }));

builder.Services.AddDirectoryBrowser();

var app = builder.Build();
app.UseCors();

#region index.html - style.css
app.UseDefaultFiles();
app.UseStaticFiles();
#endregion

#region wwwroot/audio_outputs
string webRootPath = Path.Combine(Directory.GetCurrentDirectory(), "wwwroot", "audio_outputs");
if (!Directory.Exists(webRootPath))
    Directory.CreateDirectory(webRootPath);

IFileProvider fileProvider = new PhysicalFileProvider(Path.Combine(builder.Environment.WebRootPath, "audio_outputs"));

app.UseStaticFiles(new StaticFileOptions
{
    FileProvider = fileProvider,
    RequestPath = "/audio_outputs"
});

app.UseDirectoryBrowser(new DirectoryBrowserOptions
{
    FileProvider = fileProvider,
    RequestPath = "/audio_outputs"
});
#endregion

//Endpoints...

app.Run();

record TextToSpeechRequest(string Text);

app.MapPost("/text-to-speech", async (IHttpClientFactory httpClientFactory, HttpContext httpContext) =>
{
    try
    {
        HttpClient httpClient = httpClientFactory.CreateClient("HuggingFaceAPI");

        using StreamReader streamReader = new StreamReader(httpContext.Request.Body);
        string requestBody = await streamReader.ReadToEndAsync();
        TextToSpeechRequest request = JsonSerializer.Deserialize<TextToSpeechRequest>(requestBody)!;

        var requestData = new
        {
            inputs = request.Text
        };

        StringContent stringContent = new(
            JsonSerializer.Serialize(requestData),
            Encoding.UTF8,
            MediaTypeNames.Application.Json
            );

        Console.WriteLine("Ses oluşturuluyor...");

        HttpResponseMessage response = await httpClient.PostAsync("models/facebook/mms-tts-tur", stringContent);
        if (response.IsSuccessStatusCode)
        {
            byte[] audioBytes = await response.Content.ReadAsByteArrayAsync();
            string outputPath = Path.Combine(Directory.GetCurrentDirectory(), "wwwroot", "audio_outputs", $"output_{DateTime.Now:yyyyMMddHHmmss}.wav");

            //Ses dosyası kaydediliyor
            File.WriteAllBytes(outputPath, audioBytes);
            httpContext.Response.ContentType = "audio/wav";
            await httpContext.Response.Body.WriteAsync(audioBytes);
        }
        else
        {
            string error = await response.Content.ReadAsStringAsync();
            Console.WriteLine($"API Hatası\t: {response.StatusCode}");
            Console.WriteLine($"Hata Detayı\t: {error}");
        }
    }
    catch (Exception ex)
    {
        Console.WriteLine($"Hata oluştu\t: {ex.Message}");
    }
});

app.MapPost("/speech-to-text", async (IHttpClientFactory httpClientFactory, HttpContext httpContext) =>
{
    try
    {
        IFormCollection form = await httpContext.Request.ReadFormAsync();
        IFormFile? file = form.Files.GetFile("audio");

        if (file is { Length: 0 } or null)
        {
            httpContext.Response.StatusCode = 400;
            await httpContext.Response.WriteAsJsonAsync(new { Error = "Ses dosyası bulunamadı" });
            return;
        }

        using MemoryStream memoryStream = new();
        await file.CopyToAsync(memoryStream);
        byte[] audioBytes = memoryStream.ToArray();

        HttpClient httpClient = httpClientFactory.CreateClient("HuggingFaceAPI");

        ByteArrayContent byteArrayContent = new(audioBytes);
        byteArrayContent.Headers.Add(HeaderNames.ContentType, "audio/wav");

        //Whisper small modeli kullanılıyor | Hızlı ve güvenilir!
        HttpResponseMessage response = await httpClient.PostAsync("models/openai/whisper-large-v3-turbo", byteArrayContent);
        string responseContent = await response.Content.ReadAsStringAsync();
        Console.WriteLine($"API Yanıtı\t: {responseContent}");
        if (!response.IsSuccessStatusCode)
        {
            httpContext.Response.StatusCode = (int)response.StatusCode;
            await httpContext.Response.WriteAsJsonAsync(new { Error = $"API Hatası\t:{responseContent}" });
            return;
        }

        //Whisper modeli için özel yanıt formatı ayarlanıyor
        Dictionary<string, string>? result = JsonSerializer.Deserialize<Dictionary<string, string>>(responseContent);
        string? text = result?.GetValueOrDefault("text", "");

        if (string.IsNullOrEmpty(text))
        {
            httpContext.Response.StatusCode = 500;
            await httpContext.Response.WriteAsJsonAsync(new { Error = "Metin çevrilemedi!" });
            return;
        }

        await httpContext.Response.WriteAsJsonAsync(new { Text = text });
    }
    catch (Exception ex)
    {
        Console.WriteLine($"Hata\t: {ex.Message}");
        httpContext.Response.StatusCode = 500;
        await httpContext.Response.WriteAsJsonAsync(new { Error = $"Sunucu hatası\t: {ex.Message}" });
    }
})
    .DisableAntiforgery();

<!DOCTYPE html>
<html lang="tr">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Ses-Metin Dönüştürücü</title>
    <link href="style.css" rel="stylesheet" />
</head>
<body>
    <div class="tab-container">
        <h1 style="text-align: center; margin-bottom: 20px;">Ses Dönüştürücü</h1>
        <div class="tab-buttons">
            <button onclick="showTab('tts')" class="tab-button active">Metin-Ses</button>
            <button onclick="showTab('stt')" class="tab-button">Ses-Metin</button>
        </div>
    </div>

    <div id="tts" class="tab-content container active">
        <h2>Metin-Ses Dönüştürücü</h2>
        <textarea id="inputText" placeholder="Türkçe metin girin..."></textarea>
        <button onclick="convertToSpeech()">Sese Dönüştür</button>
        <div id="ttsStatus" class="status"></div>
        <audio id="audioPlayer" controls style="display: none; margin-top: 20px; width: 100%;"></audio>
    </div>

    <div id="stt" class="tab-content container">
        <h2>Ses-Metin Dönüştürücü</h2>
        <div id="audioControls">
            <button id="recordButton" onclick="toggleRecording()">Kayıt Başlat</button>
            <input type="file" id="audioFile" accept="audio/*" style="display: none;">
            <button onclick="document.getElementById('audioFile').click()">Ses Dosyası Seç</button>
        </div>
        <div id="sttStatus" class="status"></div>
        <textarea id="outputText" placeholder="Dönüştürülen metin burada görünecek..." readonly></textarea>
    </div>
    <script src="script.js"></script>
</body>
</html> 

let mediaRecorder;
let audioChunks = [];
let isRecording = false;

function showTab(tabId) {
    document.querySelectorAll('.tab-content').forEach(tab => tab.classList.remove('active'));
    document.querySelectorAll('.tab-button').forEach(btn => btn.classList.remove('active'));
    document.getElementById(tabId).classList.add('active');
    document.querySelector(`button[onclick="showTab('${tabId}')"]`).classList.add('active');
}

async function convertToSpeech() {
    const inputText = document.getElementById('inputText').value;
    const statusDiv = document.getElementById('ttsStatus');
    const audioPlayer = document.getElementById('audioPlayer');
    const button = document.querySelector('#tts button');

    if (!inputText) {
        statusDiv.className = 'status error';
        statusDiv.textContent = 'Lütfen bir metin girin.';
        return;
    }

    try {
        button.disabled = true;
        statusDiv.className = 'status';
        statusDiv.textContent = 'Ses oluşturuluyor...';

        const response = await fetch('/text-to-speech', {
            method: 'POST',
            headers: {
                'Content-Type': 'application/json',
            },
            body: JSON.stringify({ Text: inputText })
        });

        if (!response.ok) {
            throw new Error(`HTTP error! status: ${response.status}`);
        }

        const audioBlob = await response.blob();
        const audioUrl = URL.createObjectURL(audioBlob);

        audioPlayer.src = audioUrl;
        audioPlayer.style.display = 'block';
        await audioPlayer.play();

        statusDiv.className = 'status success';
        statusDiv.textContent = 'Ses başarıyla oluşturuldu!';
    } catch (error) {
        statusDiv.className = 'status error';
        statusDiv.textContent = `Hata oluştu: ${error.message}`;
        console.error('Error:', error);
    } finally {
        button.disabled = false;
    }
}

async function toggleRecording() {
    const recordButton = document.getElementById('recordButton');
    const statusDiv = document.getElementById('sttStatus');

    if (!isRecording) {
        try {
            const stream = await navigator.mediaDevices.getUserMedia({ audio: true });
            mediaRecorder = new MediaRecorder(stream);
            audioChunks = [];

            mediaRecorder.ondataavailable = (event) => {
                audioChunks.push(event.data);
            };

            mediaRecorder.onstop = async () => {
                const audioBlob = new Blob(audioChunks, { type: 'audio/wav' });
                await convertSpeechToText(audioBlob);
            };

            mediaRecorder.start();
            isRecording = true;
            recordButton.textContent = 'Kayıt Durdur';
            recordButton.classList.add('recording');
            statusDiv.className = 'status success';
            statusDiv.textContent = 'Kayıt yapılıyor...';
        } catch (error) {
            statusDiv.className = 'status error';
            statusDiv.textContent = `Mikrofon erişim hatası: ${error.message}`;
        }
    } else {
        mediaRecorder.stop();
        isRecording = false;
        recordButton.textContent = 'Kayıt Başlat';
        recordButton.classList.remove('recording');
    }
}

document.getElementById('audioFile').addEventListener('change', async (event) => {
    const file = event.target.files[0];
    if (file) {
        await convertSpeechToText(file);
    }
});

async function convertSpeechToText(audioData) {
    const statusDiv = document.getElementById('sttStatus');
    const outputText = document.getElementById('outputText');

    try {
        statusDiv.className = 'status';
        statusDiv.textContent = 'Ses metne dönüştürülüyor...';

        const formData = new FormData();
        formData.append('audio', audioData);

        const response = await fetch('/speech-to-text', {
            method: 'POST',
            body: formData
        });

        if (!response.ok) {
            throw new Error(`HTTP error! status: ${response.status}`);
        }

        const result = await response.json();
        outputText.value = result.text;
        statusDiv.className = 'status success';
        statusDiv.textContent = 'Ses başarıyla metne dönüştürüldü!';
    } catch (error) {
        statusDiv.className = 'status error';
        statusDiv.textContent = `Hata oluştu: ${error.message}`;
        console.error('Error:', error);
    }
}