OpenAI-DotNet 5.1.0 (RageAgainstThePixel#37)

- Closes RageAgainstThePixel#34 adds Audio Endpoint to support 
  - audio speech to text
  - audio translation
- updated unit tests

OpenAI-DotNet-Tests/TestFixture_03_Edits.cs → OpenAI-DotNet-Tests/TestFixture_04_Edits.cs

@@ -5,7 +5,7 @@
 
 namespace OpenAI.Tests
 {
- internal class TestFixture_03_Edits
+ internal class TestFixture_04_Edits
  {
  [Test]
  public async Task Test_1_GetBasicEdit()

OpenAI-DotNet-Tests/TestFixture_04_Images.cs → OpenAI-DotNet-Tests/TestFixture_05_Images.cs

@@ -6,7 +6,7 @@
 
 namespace OpenAI.Tests
 {
- internal class TestFixture_04_Images
+ internal class TestFixture_05_Images
  {
  [Test]
  public async Task Test_1_GenerateImages()

OpenAI-DotNet-Tests/TestFixture_05_Embeddings.cs → OpenAI-DotNet-Tests/TestFixture_06_Embeddings.cs

@@ -3,7 +3,7 @@
 
 namespace OpenAI.Tests
 {
- internal class TestFixture_05_Embeddings
+ internal class TestFixture_06_Embeddings
  {
  [Test]
  public async Task Test_1_CreateEmbedding()

OpenAI-DotNet-Tests/TestFixture_07_Audio.cs

@@ -0,0 +1,35 @@
+using NUnit.Framework;
+using OpenAI.Audio;
+using System;
+using System.IO;
+using System.Threading.Tasks;
+
+namespace OpenAI.Tests
+{
+ internal class TestFixture_07_Audio
+ {
+ [Test]
+ public async Task Test_1_Transcription()
+ {
+ var api = new OpenAIClient(OpenAIAuthentication.LoadFromEnv());
+ Assert.IsNotNull(api.AudioEndpoint);
+ var transcriptionAudio = Path.GetFullPath("..\\..\\..\\Assets\\T3mt39YrlyLoq8laHSdf.mp3");
+ var request = new AudioTranscriptionRequest(transcriptionAudio, language: "en");
+ var result = await api.AudioEndpoint.CreateTranscriptionAsync(request);
+ Assert.IsNotNull(result);
+ Console.WriteLine(result);
+ }
+
+ [Test]
+ public async Task Test_2_Translation()
+ {
+ var api = new OpenAIClient(OpenAIAuthentication.LoadFromEnv());
+ Assert.IsNotNull(api.AudioEndpoint);
+ var translationAudio = Path.GetFullPath("..\\..\\..\\Assets\\Ja-botchan_1-1_1-2.mp3");
+ var request = new AudioTranslationRequest(Path.GetFullPath(translationAudio));
+ var result = await api.AudioEndpoint.CreateTranslationAsync(request);
+ Assert.IsNotNull(result);
+ Console.WriteLine(result);
+ }
+ }
+}

OpenAI-DotNet-Tests/TestFixture_06_Files.cs → OpenAI-DotNet-Tests/TestFixture_08_Files.cs

@@ -6,7 +6,7 @@
 
 namespace OpenAI.Tests
 {
- internal class TestFixture_06_Files
+ internal class TestFixture_08_Files
  {
  [Test]
  public async Task Test_01_UploadFile()

OpenAI-DotNet-Tests/TestFixture_07_FineTunes.cs → OpenAI-DotNet-Tests/TestFixture_09_FineTuning.cs

@@ -9,7 +9,7 @@
 
 namespace OpenAI.Tests
 {
- internal class TestFixture_07_FineTuning
+ internal class TestFixture_09_FineTuning
  {
  private async Task<FileData> CreateTestTrainingDataAsync(OpenAIClient api)
  {

OpenAI-DotNet-Tests/TestFixture_08_Moderations.cs → OpenAI-DotNet-Tests/TestFixture_10_Moderations.cs

@@ -3,7 +3,7 @@
 
 namespace OpenAI.Tests
 {
- internal class TestFixture_08_Moderations
+ internal class TestFixture_10_Moderations
  {
  [Test]
  public async Task Test_1_Moderate()

OpenAI-DotNet/Audio/AudioEndpoint.cs

@@ -0,0 +1,112 @@
+using System.IO;
+using System.Net.Http;
+using System.Text.Json;
+using System.Text.Json.Serialization;
+using System.Threading;
+using System.Threading.Tasks;
+
+namespace OpenAI.Audio
+{
+ /// <summary>
+ /// Speech to text.
+ /// </summary>
+ public sealed class AudioEndpoint : BaseEndPoint
+ {
+ private class AudioResponse
+ {
+ public AudioResponse(string text)
+ {
+ Text = text;
+ }
+
+ [JsonPropertyName("text")]
+ public string Text { get; }
+ }
+
+ /// <inheritdoc />
+ public AudioEndpoint(OpenAIClient api) : base(api) { }
+
+ /// <inheritdoc />
+ protected override string GetEndpoint()
+ => $"{Api.BaseUrl}audio";
+
+ /// <summary>
+ /// Transcribes audio into the input language.
+ /// </summary>
+ /// <param name="request"><see cref="AudioTranscriptionRequest"/>.</param>
+ /// <param name="cancellationToken">Optional, <see cref="CancellationToken"/>.</param>
+ /// <returns>The transcribed text.</returns>
+ public async Task<string> CreateTranscriptionAsync(AudioTranscriptionRequest request, CancellationToken cancellationToken = default)
+ {
+ using var content = new MultipartFormDataContent();
+ using var audioData = new MemoryStream();
+ await request.Audio.CopyToAsync(audioData, cancellationToken);
+ content.Add(new ByteArrayContent(audioData.ToArray()), "file", request.AudioName);
+ content.Add(new StringContent(request.Model), "model");
+
+ if (!string.IsNullOrWhiteSpace(request.Prompt))
+ {
+ content.Add(new StringContent(request.Prompt), "prompt");
+ }
+
+ var responseFormat = request.ResponseFormat;
+ content.Add(new StringContent(responseFormat.ToString().ToLower()), "response_format");
+
+ if (request.Temperature.HasValue)
+ {
+ content.Add(new StringContent(request.Temperature.ToString()), "temperature");
+ }
+
+ if (!string.IsNullOrWhiteSpace(request.Language))
+ {
+ content.Add(new StringContent(request.Language), "language");
+ }
+
+ request.Dispose();
+
+ var response = await Api.Client.PostAsync($"{GetEndpoint()}/transcriptions", content, cancellationToken);
+ var responseAsString = await response.ReadAsStringAsync(cancellationToken);
+
+ return responseFormat == AudioResponseFormat.Json
+ ? JsonSerializer.Deserialize<AudioResponse>(responseAsString)?.Text
+ : responseAsString;
+ }
+
+ /// <summary>
+ /// Translates audio into into English.
+ /// </summary>
+ /// <param name="request"></param>
+ /// <param name="cancellationToken"></param>
+ /// <returns>The translated text.</returns>
+ public async Task<string> CreateTranslationAsync(AudioTranslationRequest request, CancellationToken cancellationToken = default)
+ {
+ using var content = new MultipartFormDataContent();
+ using var audioData = new MemoryStream();
+ await request.Audio.CopyToAsync(audioData, cancellationToken);
+ content.Add(new ByteArrayContent(audioData.ToArray()), "file", request.AudioName);
+ content.Add(new StringContent(request.Model), "model");
+
+ if (!string.IsNullOrWhiteSpace(request.Prompt))
+ {
+ content.Add(new StringContent(request.Prompt), "prompt");
+ }
+
+ var responseFormat = request.ResponseFormat;
+ content.Add(new StringContent(responseFormat.ToString().ToLower()), "response_format");
+
+ if (request.Temperature.HasValue)
+ {
+ content.Add(new StringContent(request.Temperature.ToString()), "temperature");
+ }
+
+ request.Dispose();
+
+ var response = await Api.Client.PostAsync($"{GetEndpoint()}/translations", content, cancellationToken);
+ var responseAsString = await response.ReadAsStringAsync(cancellationToken);
+
+ return responseFormat == AudioResponseFormat.Json
+ ? JsonSerializer.Deserialize<AudioResponse>(responseAsString)?.Text
+ : responseAsString;
+ }
+ }
+}

OpenAI-DotNet/Audio/AudioResponseFormat.cs

@@ -0,0 +1,11 @@
+namespace OpenAI.Audio
+{
+ public enum AudioResponseFormat
+ {
+ Json = 0,
+ Verbose_Json,
+ Text,
+ Srt,
+ Vtt
+ }
+}