openai/openai-dotnet
Publicmirrored from https://github.com/openai/openai-dotnetAvailable
src/Custom/Audio/AudioClient.cs
469lines · modecode
| 1 | using Microsoft.TypeSpec.Generator.Customizations; |
| 2 | using System; |
| 3 | using System.ClientModel; |
| 4 | using System.ClientModel.Primitives; |
| 5 | using System.Diagnostics.CodeAnalysis; |
| 6 | using System.IO; |
| 7 | using System.Threading; |
| 8 | using System.Threading.Tasks; |
| 9 | |
| 10 | namespace OpenAI.Audio; |
| 11 | |
| 12 | // CUSTOM: |
| 13 | // - Renamed. |
| 14 | // - Suppressed constructor that takes endpoint parameter; endpoint is now a property in the options class. |
| 15 | // - Suppressed methods that only take the options parameter. |
| 16 | /// <summary> The service client for OpenAI audio operations. </summary> |
| 17 | [CodeGenType("Audio")] |
| 18 | [CodeGenSuppress("AudioClient", typeof(ClientPipeline), typeof(Uri))] |
| 19 | [CodeGenSuppress("GenerateSpeechAsync", typeof(SpeechGenerationOptions), typeof(CancellationToken))] |
| 20 | [CodeGenSuppress("GenerateSpeech", typeof(SpeechGenerationOptions), typeof(CancellationToken))] |
| 21 | public partial class AudioClient |
| 22 | { |
| 23 | private readonly string _model; |
| 24 | |
| 25 | // CUSTOM: Added as a convenience. |
| 26 | /// <summary> Initializes a new instance of <see cref="AudioClient"/>. </summary> |
| 27 | /// <param name="model"> The name of the model to use in requests sent to the service. To learn more about the available models, see <see href="https://platform.openai.com/docs/models"/>. </param> |
| 28 | /// <param name="apiKey"> The API key to authenticate with the service. </param> |
| 29 | /// <exception cref="ArgumentNullException"> <paramref name="model"/> or <paramref name="apiKey"/> is null. </exception> |
| 30 | /// <exception cref="ArgumentException"> <paramref name="model"/> is an empty string, and was expected to be non-empty. </exception> |
| 31 | public AudioClient(string model, string apiKey) : this(model, new ApiKeyCredential(apiKey), new OpenAIClientOptions()) |
| 32 | { |
| 33 | } |
| 34 | |
| 35 | // CUSTOM: |
| 36 | // - Added `model` parameter. |
| 37 | // - Used a custom pipeline. |
| 38 | // - Demoted the endpoint parameter to be a property in the options class. |
| 39 | /// <summary> Initializes a new instance of <see cref="AudioClient"/>. </summary> |
| 40 | /// <param name="model"> The name of the model to use in requests sent to the service. To learn more about the available models, see <see href="https://platform.openai.com/docs/models"/>. </param> |
| 41 | /// <param name="credential"> The <see cref="ApiKeyCredential"/> to authenticate with the service. </param> |
| 42 | /// <exception cref="ArgumentNullException"> <paramref name="model"/> or <paramref name="credential"/> is null. </exception> |
| 43 | /// <exception cref="ArgumentException"> <paramref name="model"/> is an empty string, and was expected to be non-empty. </exception> |
| 44 | public AudioClient(string model, ApiKeyCredential credential) : this(model, credential, new OpenAIClientOptions()) |
| 45 | { |
| 46 | } |
| 47 | |
| 48 | // CUSTOM: |
| 49 | // - Added `model` parameter. |
| 50 | // - Used a custom pipeline. |
| 51 | // - Demoted the endpoint parameter to be a property in the options class. |
| 52 | /// <summary> Initializes a new instance of <see cref="AudioClient"/>. </summary> |
| 53 | /// <param name="model"> The name of the model to use in requests sent to the service. To learn more about the available models, see <see href="https://platform.openai.com/docs/models"/>. </param> |
| 54 | /// <param name="credential"> The <see cref="ApiKeyCredential"/> to authenticate with the service. </param> |
| 55 | /// <param name="options"> The options to configure the client. </param> |
| 56 | /// <exception cref="ArgumentNullException"> <paramref name="model"/> or <paramref name="credential"/> is null. </exception> |
| 57 | /// <exception cref="ArgumentException"> <paramref name="model"/> is an empty string, and was expected to be non-empty. </exception> |
| 58 | public AudioClient(string model, ApiKeyCredential credential, OpenAIClientOptions options) : this(model, OpenAIClient.CreateApiKeyAuthenticationPolicy(credential), options) |
| 59 | { |
| 60 | } |
| 61 | |
| 62 | // CUSTOM: Added as a convenience. |
| 63 | /// <summary> Initializes a new instance of <see cref="AudioClient"/>. </summary> |
| 64 | /// <param name="model"> The name of the model to use in requests sent to the service. To learn more about the available models, see <see href="https://platform.openai.com/docs/models"/>. </param> |
| 65 | /// <param name="authenticationPolicy"> The authentication policy used to authenticate with the service. </param> |
| 66 | /// <exception cref="ArgumentNullException"> <paramref name="model"/> or <paramref name="authenticationPolicy"/> is null. </exception> |
| 67 | /// <exception cref="ArgumentException"> <paramref name="model"/> is an empty string, and was expected to be non-empty. </exception> |
| 68 | [Experimental("OPENAI001")] |
| 69 | public AudioClient(string model, AuthenticationPolicy authenticationPolicy) : this(model, authenticationPolicy, new OpenAIClientOptions()) |
| 70 | { |
| 71 | } |
| 72 | |
| 73 | // CUSTOM: Added as a convenience. |
| 74 | /// <summary> Initializes a new instance of <see cref="AudioClient"/>. </summary> |
| 75 | /// <param name="model"> The name of the model to use in requests sent to the service. To learn more about the available models, see <see href="https://platform.openai.com/docs/models"/>. </param> |
| 76 | /// <param name="authenticationPolicy"> The authentication policy used to authenticate with the service. </param> |
| 77 | /// <param name="options"> The options to configure the client. </param> |
| 78 | /// <exception cref="ArgumentNullException"> <paramref name="model"/> or <paramref name="authenticationPolicy"/> is null. </exception> |
| 79 | /// <exception cref="ArgumentException"> <paramref name="model"/> is an empty string, and was expected to be non-empty. </exception> |
| 80 | [Experimental("OPENAI001")] |
| 81 | public AudioClient(string model, AuthenticationPolicy authenticationPolicy, OpenAIClientOptions options) |
| 82 | { |
| 83 | Argument.AssertNotNullOrEmpty(model, nameof(model)); |
| 84 | Argument.AssertNotNull(authenticationPolicy, nameof(authenticationPolicy)); |
| 85 | options ??= new OpenAIClientOptions(); |
| 86 | |
| 87 | _model = model; |
| 88 | Pipeline = OpenAIClient.CreatePipeline(authenticationPolicy, options); |
| 89 | _endpoint = OpenAIClient.GetEndpoint(options); |
| 90 | } |
| 91 | |
| 92 | // CUSTOM: |
| 93 | // - Added `model` parameter. |
| 94 | // - Used a custom pipeline. |
| 95 | // - Demoted the endpoint parameter to be a property in the options class. |
| 96 | // - Made protected. |
| 97 | /// <summary> Initializes a new instance of <see cref="AudioClient"/>. </summary> |
| 98 | /// <param name="pipeline"> The HTTP pipeline to send and receive REST requests and responses. </param> |
| 99 | /// <param name="model"> The name of the model to use in requests sent to the service. To learn more about the available models, see <see href="https://platform.openai.com/docs/models"/>. </param> |
| 100 | /// <param name="options"> The options to configure the client. </param> |
| 101 | /// <exception cref="ArgumentNullException"> <paramref name="pipeline"/> or <paramref name="model"/> is null. </exception> |
| 102 | /// <exception cref="ArgumentException"> <paramref name="model"/> is an empty string, and was expected to be non-empty. </exception> |
| 103 | protected internal AudioClient(ClientPipeline pipeline, string model, OpenAIClientOptions options) |
| 104 | { |
| 105 | Argument.AssertNotNull(pipeline, nameof(pipeline)); |
| 106 | Argument.AssertNotNullOrEmpty(model, nameof(model)); |
| 107 | options ??= new OpenAIClientOptions(); |
| 108 | |
| 109 | _model = model; |
| 110 | Pipeline = pipeline; |
| 111 | _endpoint = OpenAIClient.GetEndpoint(options); |
| 112 | } |
| 113 | |
| 114 | /// <summary> |
| 115 | /// Gets the name of the model used in requests sent to the service. |
| 116 | /// </summary> |
| 117 | [Experimental("OPENAI001")] |
| 118 | public string Model => _model; |
| 119 | |
| 120 | /// <summary> |
| 121 | /// Gets the endpoint URI for the service. |
| 122 | /// </summary> |
| 123 | [Experimental("OPENAI001")] |
| 124 | public Uri Endpoint => _endpoint; |
| 125 | |
| 126 | #region GenerateSpeech |
| 127 | |
| 128 | /// <summary> Generates a life-like, spoken audio recording of the input text. </summary> |
| 129 | /// <remarks> |
| 130 | /// The default format of the generated audio is <see cref="GeneratedSpeechFormat.Mp3"/> unless otherwise specified |
| 131 | /// via <see cref="SpeechGenerationOptions.ResponseFormat"/>. |
| 132 | /// </remarks> |
| 133 | /// <param name="text"> The text to generate audio for. </param> |
| 134 | /// <param name="voice"> The voice to use in the generated audio. </param> |
| 135 | /// <param name="options"> The options to configure the audio generation. </param> |
| 136 | /// <param name="cancellationToken"> A token that can be used to cancel this method call. </param> |
| 137 | /// <exception cref="ArgumentNullException"> <paramref name="text"/> is null. </exception> |
| 138 | /// <returns> The generated audio in the specified output format. </returns> |
| 139 | public virtual async Task<ClientResult<BinaryData>> GenerateSpeechAsync(string text, GeneratedSpeechVoice voice, SpeechGenerationOptions options = null, CancellationToken cancellationToken = default) |
| 140 | { |
| 141 | Argument.AssertNotNull(text, nameof(text)); |
| 142 | |
| 143 | options ??= new(); |
| 144 | CreateSpeechGenerationOptions(text, voice, ref options); |
| 145 | |
| 146 | using BinaryContent content = options.ToBinaryContent(); |
| 147 | ClientResult result = await GenerateSpeechAsync(content, cancellationToken.ToRequestOptions()).ConfigureAwait(false); |
| 148 | return ClientResult.FromValue(result.GetRawResponse().Content, result.GetRawResponse()); |
| 149 | } |
| 150 | |
| 151 | /// <summary> Generates a life-like, spoken audio recording of the input text. </summary> |
| 152 | /// <remarks> |
| 153 | /// The default format of the generated audio is <see cref="GeneratedSpeechFormat.Mp3"/> unless otherwise specified |
| 154 | /// via <see cref="SpeechGenerationOptions.ResponseFormat"/>. |
| 155 | /// </remarks> |
| 156 | /// <param name="text"> The text to generate audio for. </param> |
| 157 | /// <param name="voice"> The voice to use in the generated audio. </param> |
| 158 | /// <param name="options"> The options to configure the audio generation. </param> |
| 159 | /// <param name="cancellationToken"> A token that can be used to cancel this method call. </param> |
| 160 | /// <exception cref="ArgumentNullException"> <paramref name="text"/> is null. </exception> |
| 161 | /// <returns> The generated audio in the specified output format. </returns> |
| 162 | public virtual ClientResult<BinaryData> GenerateSpeech(string text, GeneratedSpeechVoice voice, SpeechGenerationOptions options = null, CancellationToken cancellationToken = default) |
| 163 | { |
| 164 | Argument.AssertNotNull(text, nameof(text)); |
| 165 | |
| 166 | options ??= new(); |
| 167 | CreateSpeechGenerationOptions(text, voice, ref options); |
| 168 | |
| 169 | using BinaryContent content = options.ToBinaryContent(); |
| 170 | ClientResult result = GenerateSpeech(content, cancellationToken.ToRequestOptions()); ; |
| 171 | return ClientResult.FromValue(result.GetRawResponse().Content, result.GetRawResponse()); |
| 172 | } |
| 173 | |
| 174 | #endregion |
| 175 | |
| 176 | #region TranscribeAudio |
| 177 | |
| 178 | /// <summary> Transcribes the input audio. </summary> |
| 179 | /// <param name="audio"> The audio stream to transcribe. </param> |
| 180 | /// <param name="audioFilename"> |
| 181 | /// The filename associated with the audio stream. The filename's extension (for example: .mp3) will be used to |
| 182 | /// validate the format of the input audio. The request may fail if the filename's extension and the actual |
| 183 | /// format of the input audio do not match. |
| 184 | /// </param> |
| 185 | /// <param name="options"> The options to configure the audio transcription. </param> |
| 186 | /// <param name="cancellationToken"> A token that can be used to cancel this method call. </param> |
| 187 | /// <exception cref="ArgumentNullException"> <paramref name="audio"/> or <paramref name="audioFilename"/> is null. </exception> |
| 188 | /// <exception cref="ArgumentException"> <paramref name="audioFilename"/> is an empty string, and was expected to be non-empty. </exception> |
| 189 | public virtual async Task<ClientResult<AudioTranscription>> TranscribeAudioAsync(Stream audio, string audioFilename, AudioTranscriptionOptions options = null, CancellationToken cancellationToken = default) |
| 190 | { |
| 191 | Argument.AssertNotNull(audio, nameof(audio)); |
| 192 | Argument.AssertNotNullOrEmpty(audioFilename, nameof(audioFilename)); |
| 193 | |
| 194 | using MultiPartFormDataBinaryContent content |
| 195 | = CreatePerCallTranscriptionOptions(options) |
| 196 | .ToMultipartContent(audio, audioFilename); |
| 197 | |
| 198 | ClientResult result = await TranscribeAudioAsync(content, content.ContentType, cancellationToken.ToRequestOptions()).ConfigureAwait(false); |
| 199 | return ClientResult.FromValue(AudioTranscription.FromResponse(result.GetRawResponse()), result.GetRawResponse()); |
| 200 | } |
| 201 | |
| 202 | /// <summary> Transcribes the input audio. </summary> |
| 203 | /// <param name="audio"> The audio stream to transcribe. </param> |
| 204 | /// <param name="audioFilename"> |
| 205 | /// The filename associated with the audio stream. The filename's extension (for example: .mp3) will be used to |
| 206 | /// validate the format of the input audio. The request may fail if the filename's extension and the actual |
| 207 | /// format of the input audio do not match. |
| 208 | /// </param> |
| 209 | /// <param name="options"> The options to configure the audio transcription. </param> |
| 210 | /// <param name="cancellationToken"> A token that can be used to cancel this method call. </param> |
| 211 | /// <exception cref="ArgumentNullException"> <paramref name="audio"/> or <paramref name="audioFilename"/> is null. </exception> |
| 212 | /// <exception cref="ArgumentException"> <paramref name="audioFilename"/> is an empty string, and was expected to be non-empty. </exception> |
| 213 | public virtual ClientResult<AudioTranscription> TranscribeAudio(Stream audio, string audioFilename, AudioTranscriptionOptions options = null, CancellationToken cancellationToken = default) |
| 214 | { |
| 215 | Argument.AssertNotNull(audio, nameof(audio)); |
| 216 | Argument.AssertNotNullOrEmpty(audioFilename, nameof(audioFilename)); |
| 217 | |
| 218 | using MultiPartFormDataBinaryContent content |
| 219 | = CreatePerCallTranscriptionOptions(options) |
| 220 | .ToMultipartContent(audio, audioFilename); |
| 221 | |
| 222 | ClientResult result = TranscribeAudio(content, content.ContentType, cancellationToken.ToRequestOptions()); |
| 223 | return ClientResult.FromValue(AudioTranscription.FromResponse(result.GetRawResponse()), result.GetRawResponse()); |
| 224 | } |
| 225 | |
| 226 | /// <summary> Transcribes the input audio. </summary> |
| 227 | /// <param name="audioFilePath"> |
| 228 | /// The path of the audio file to transcribe. The provided file path's extension (for example: .mp3) will be |
| 229 | /// used to validate the format of the input audio. The request may fail if the file path's extension and the |
| 230 | /// actual format of the input audio do not match. |
| 231 | /// </param> |
| 232 | /// <param name="options"> The options to configure the audio transcription. </param> |
| 233 | /// <exception cref="ArgumentNullException"> <paramref name="audioFilePath"/> is null. </exception> |
| 234 | /// <exception cref="ArgumentException"> <paramref name="audioFilePath"/> is an empty string, and was expected to be non-empty. </exception> |
| 235 | public virtual async Task<ClientResult<AudioTranscription>> TranscribeAudioAsync(string audioFilePath, AudioTranscriptionOptions options = null) |
| 236 | { |
| 237 | Argument.AssertNotNullOrEmpty(audioFilePath, nameof(audioFilePath)); |
| 238 | |
| 239 | using FileStream audioStream = File.OpenRead(audioFilePath); |
| 240 | return await TranscribeAudioAsync(audioStream, audioFilePath, options).ConfigureAwait(false); |
| 241 | } |
| 242 | |
| 243 | /// <summary> Transcribes the input audio. </summary> |
| 244 | /// <param name="audioFilePath"> |
| 245 | /// The path of the audio file to transcribe. The provided file path's extension (for example: .mp3) will be |
| 246 | /// used to validate the format of the input audio. The request may fail if the file path's extension and the |
| 247 | /// actual format of the input audio do not match. |
| 248 | /// </param> |
| 249 | /// <param name="options"> The options to configure the audio transcription. </param> |
| 250 | /// <exception cref="ArgumentNullException"> <paramref name="audioFilePath"/> is null. </exception> |
| 251 | /// <exception cref="ArgumentException"> <paramref name="audioFilePath"/> is an empty string, and was expected to be non-empty. </exception> |
| 252 | public virtual ClientResult<AudioTranscription> TranscribeAudio(string audioFilePath, AudioTranscriptionOptions options = null) |
| 253 | { |
| 254 | Argument.AssertNotNullOrEmpty(audioFilePath, nameof(audioFilePath)); |
| 255 | |
| 256 | using FileStream audioStream = File.OpenRead(audioFilePath); |
| 257 | return TranscribeAudio(audioStream, audioFilePath, options); |
| 258 | } |
| 259 | |
| 260 | // CUSTOM: Added Experimental attribute. |
| 261 | [Experimental("OPENAI001")] |
| 262 | public virtual AsyncCollectionResult<StreamingAudioTranscriptionUpdate> TranscribeAudioStreamingAsync(Stream audio, string audioFilename, AudioTranscriptionOptions options = null, CancellationToken cancellationToken = default) |
| 263 | { |
| 264 | Argument.AssertNotNull(audio, nameof(audio)); |
| 265 | Argument.AssertNotNullOrEmpty(audioFilename, nameof(audioFilename)); |
| 266 | |
| 267 | EnsureModelSupportsStreaming(); |
| 268 | |
| 269 | MultiPartFormDataBinaryContent content |
| 270 | = CreatePerCallTranscriptionOptions(options, stream: true) |
| 271 | .ToMultipartContent(audio, audioFilename); |
| 272 | |
| 273 | return new AsyncSseUpdateCollection<StreamingAudioTranscriptionUpdate>( |
| 274 | async () => await TranscribeAudioAsync(content, content.ContentType, cancellationToken.ToRequestOptions(streaming: true)).ConfigureAwait(false), |
| 275 | StreamingAudioTranscriptionUpdate.DeserializeStreamingAudioTranscriptionUpdate, |
| 276 | cancellationToken); |
| 277 | } |
| 278 | |
| 279 | // CUSTOM: Added Experimental attribute. |
| 280 | [Experimental("OPENAI001")] |
| 281 | public virtual AsyncCollectionResult<StreamingAudioTranscriptionUpdate> TranscribeAudioStreamingAsync(string audioFilePath, AudioTranscriptionOptions options = null, CancellationToken cancellationToken = default) |
| 282 | { |
| 283 | Argument.AssertNotNullOrEmpty(audioFilePath, nameof(audioFilePath)); |
| 284 | |
| 285 | EnsureModelSupportsStreaming(); |
| 286 | |
| 287 | FileStream inputStream = File.OpenRead(audioFilePath); |
| 288 | |
| 289 | MultiPartFormDataBinaryContent content |
| 290 | = CreatePerCallTranscriptionOptions(options, stream: true) |
| 291 | .ToMultipartContent(inputStream, audioFilePath); |
| 292 | |
| 293 | AsyncSseUpdateCollection<StreamingAudioTranscriptionUpdate> result = new( |
| 294 | async () => await TranscribeAudioAsync(content, content.ContentType, cancellationToken.ToRequestOptions(streaming: true)).ConfigureAwait(false), |
| 295 | StreamingAudioTranscriptionUpdate.DeserializeStreamingAudioTranscriptionUpdate, |
| 296 | cancellationToken); |
| 297 | result.AdditionalDisposalActions.Add(() => inputStream?.Dispose()); |
| 298 | return result; |
| 299 | } |
| 300 | |
| 301 | // CUSTOM: Added Experimental attribute. |
| 302 | [Experimental("OPENAI001")] |
| 303 | public virtual CollectionResult<StreamingAudioTranscriptionUpdate> TranscribeAudioStreaming(Stream audio, string audioFilename, AudioTranscriptionOptions options = null, CancellationToken cancellationToken = default) |
| 304 | { |
| 305 | Argument.AssertNotNull(audio, nameof(audio)); |
| 306 | Argument.AssertNotNullOrEmpty(audioFilename, nameof(audioFilename)); |
| 307 | |
| 308 | EnsureModelSupportsStreaming(); |
| 309 | |
| 310 | MultiPartFormDataBinaryContent content |
| 311 | = CreatePerCallTranscriptionOptions(options, stream: true) |
| 312 | .ToMultipartContent(audio, audioFilename); |
| 313 | |
| 314 | return new SseUpdateCollection<StreamingAudioTranscriptionUpdate>( |
| 315 | () => TranscribeAudio(content, content.ContentType, cancellationToken.ToRequestOptions(streaming: true)), |
| 316 | StreamingAudioTranscriptionUpdate.DeserializeStreamingAudioTranscriptionUpdate, |
| 317 | cancellationToken); |
| 318 | } |
| 319 | |
| 320 | // CUSTOM: Added Experimental attribute. |
| 321 | [Experimental("OPENAI001")] |
| 322 | public virtual CollectionResult<StreamingAudioTranscriptionUpdate> TranscribeAudioStreaming(string audioFilePath, AudioTranscriptionOptions options = null, CancellationToken cancellationToken = default) |
| 323 | { |
| 324 | Argument.AssertNotNullOrEmpty(audioFilePath, nameof(audioFilePath)); |
| 325 | |
| 326 | EnsureModelSupportsStreaming(); |
| 327 | |
| 328 | FileStream inputStream = File.OpenRead(audioFilePath); |
| 329 | |
| 330 | MultiPartFormDataBinaryContent content |
| 331 | = CreatePerCallTranscriptionOptions(options, stream: true) |
| 332 | .ToMultipartContent(inputStream, audioFilePath); |
| 333 | |
| 334 | SseUpdateCollection<StreamingAudioTranscriptionUpdate> result = new( |
| 335 | () => TranscribeAudio(content, content.ContentType, cancellationToken.ToRequestOptions(streaming: true)), |
| 336 | StreamingAudioTranscriptionUpdate.DeserializeStreamingAudioTranscriptionUpdate, |
| 337 | cancellationToken); |
| 338 | result.AdditionalDisposalActions.Add(() => inputStream?.Dispose()); |
| 339 | return result; |
| 340 | } |
| 341 | |
| 342 | private void EnsureModelSupportsStreaming() |
| 343 | { |
| 344 | if (string.Equals(_model, "whisper-1", StringComparison.OrdinalIgnoreCase)) |
| 345 | { |
| 346 | string isEnabled = Environment.GetEnvironmentVariable("OPENAI_ENABLE_WHISPER_1_STREAMING"); |
| 347 | if (!string.Equals(isEnabled, "true", StringComparison.OrdinalIgnoreCase)) |
| 348 | { |
| 349 | throw new NotSupportedException( |
| 350 | "The selected model 'whisper-1' does not support streaming transcription. " + |
| 351 | "Please use a compatible model or set the environment variable 'OPENAI_ENABLE_WHISPER_1_STREAMING=true' to bypass this check."); |
| 352 | } |
| 353 | } |
| 354 | } |
| 355 | |
| 356 | #endregion |
| 357 | |
| 358 | #region TranslateAudio |
| 359 | |
| 360 | /// <summary> Translates the input audio into English. </summary> |
| 361 | /// <param name="audio"> The audio stream to translate. </param> |
| 362 | /// <param name="audioFilename"> |
| 363 | /// The filename associated with the audio stream. The filename's extension (for example: .mp3) will be used to |
| 364 | /// validate the format of the input audio. The request may fail if the filename's extension and the actual |
| 365 | /// format of the input audio do not match. |
| 366 | /// </param> |
| 367 | /// <param name="options"> The options to configure the audio translation. </param> |
| 368 | /// <param name="cancellationToken"> A token that can be used to cancel this method call. </param> |
| 369 | /// <exception cref="ArgumentNullException"> <paramref name="audio"/> or <paramref name="audioFilename"/> is null. </exception> |
| 370 | /// <exception cref="ArgumentException"> <paramref name="audioFilename"/> is an empty string, and was expected to be non-empty. </exception> |
| 371 | public virtual async Task<ClientResult<AudioTranslation>> TranslateAudioAsync(Stream audio, string audioFilename, AudioTranslationOptions options = null, CancellationToken cancellationToken = default) |
| 372 | { |
| 373 | Argument.AssertNotNull(audio, nameof(audio)); |
| 374 | Argument.AssertNotNullOrEmpty(audioFilename, nameof(audioFilename)); |
| 375 | |
| 376 | options ??= new(); |
| 377 | CreateAudioTranslationOptions(audio, audioFilename, ref options); |
| 378 | |
| 379 | using MultiPartFormDataBinaryContent content = options.ToMultipartContent(audio, audioFilename); |
| 380 | ClientResult result = await TranslateAudioAsync(content, content.ContentType, cancellationToken.ToRequestOptions()).ConfigureAwait(false); |
| 381 | return ClientResult.FromValue(AudioTranslation.FromResponse(result.GetRawResponse()), result.GetRawResponse()); |
| 382 | } |
| 383 | |
| 384 | /// <summary> Translates the input audio into English. </summary> |
| 385 | /// <param name="audio"> The audio stream to translate. </param> |
| 386 | /// <param name="audioFilename"> |
| 387 | /// The filename associated with the audio stream. The filename's extension (for example: .mp3) will be used to |
| 388 | /// validate the format of the input audio. The request may fail if the filename's extension and the actual |
| 389 | /// format of the input audio do not match. |
| 390 | /// </param> |
| 391 | /// <param name="options"> The options to configure the audio translation. </param> |
| 392 | /// <param name="cancellationToken"> A token that can be used to cancel this method call. </param> |
| 393 | /// <exception cref="ArgumentNullException"> <paramref name="audio"/> or <paramref name="audioFilename"/> is null. </exception> |
| 394 | /// <exception cref="ArgumentException"> <paramref name="audioFilename"/> is an empty string, and was expected to be non-empty. </exception> |
| 395 | public virtual ClientResult<AudioTranslation> TranslateAudio(Stream audio, string audioFilename, AudioTranslationOptions options = null, CancellationToken cancellationToken = default) |
| 396 | { |
| 397 | Argument.AssertNotNull(audio, nameof(audio)); |
| 398 | Argument.AssertNotNullOrEmpty(audioFilename, nameof(audioFilename)); |
| 399 | |
| 400 | options ??= new(); |
| 401 | CreateAudioTranslationOptions(audio, audioFilename, ref options); |
| 402 | |
| 403 | using MultiPartFormDataBinaryContent content = options.ToMultipartContent(audio, audioFilename); |
| 404 | ClientResult result = TranslateAudio(content, content.ContentType, cancellationToken.ToRequestOptions()); |
| 405 | return ClientResult.FromValue(AudioTranslation.FromResponse(result.GetRawResponse()), result.GetRawResponse()); |
| 406 | } |
| 407 | |
| 408 | /// <summary> Translates the input audio into English. </summary> |
| 409 | /// <param name="audioFilePath"> |
| 410 | /// The path of the audio file to translate. The provided file path's extension (for example: .mp3) will be |
| 411 | /// used to validate the format of the input audio. The request may fail if the file path's extension and the |
| 412 | /// actual format of the input audio do not match. |
| 413 | /// </param> |
| 414 | /// <param name="options"> The options to configure the audio translation. </param> |
| 415 | /// <exception cref="ArgumentNullException"> <paramref name="audioFilePath"/> was null. </exception> |
| 416 | /// <exception cref="ArgumentException"> <paramref name="audioFilePath"/> is an empty string, and was expected to be non-empty. </exception> |
| 417 | public virtual ClientResult<AudioTranslation> TranslateAudio(string audioFilePath, AudioTranslationOptions options = null) |
| 418 | { |
| 419 | Argument.AssertNotNullOrEmpty(audioFilePath, nameof(audioFilePath)); |
| 420 | |
| 421 | using FileStream audioStream = File.OpenRead(audioFilePath); |
| 422 | return TranslateAudio(audioStream, audioFilePath, options); |
| 423 | } |
| 424 | |
| 425 | /// <summary> Translates the input audio into English. </summary> |
| 426 | /// <param name="audioFilePath"> |
| 427 | /// The path of the audio file to translate. The provided file path's extension (for example: .mp3) will be |
| 428 | /// used to validate the format of the input audio. The request may fail if the file path's extension and the |
| 429 | /// actual format of the input audio do not match. |
| 430 | /// </param> |
| 431 | /// <param name="options"> The options to configure the audio translation. </param> |
| 432 | /// <exception cref="ArgumentNullException"> <paramref name="audioFilePath"/> was null. </exception> |
| 433 | /// <exception cref="ArgumentException"> <paramref name="audioFilePath"/> is an empty string, and was expected to be non-empty. </exception> |
| 434 | public virtual async Task<ClientResult<AudioTranslation>> TranslateAudioAsync(string audioFilePath, AudioTranslationOptions options = null) |
| 435 | { |
| 436 | Argument.AssertNotNull(audioFilePath, nameof(audioFilePath)); |
| 437 | |
| 438 | using FileStream audioStream = File.OpenRead(audioFilePath); |
| 439 | return await TranslateAudioAsync(audioStream, audioFilePath, options); |
| 440 | } |
| 441 | |
| 442 | #endregion |
| 443 | |
| 444 | private void CreateSpeechGenerationOptions(string text, GeneratedSpeechVoice voice, ref SpeechGenerationOptions options) |
| 445 | { |
| 446 | options.Input = text; |
| 447 | options.Voice = voice; |
| 448 | options.Model = _model; |
| 449 | } |
| 450 | |
| 451 | internal virtual AudioTranscriptionOptions CreatePerCallTranscriptionOptions(AudioTranscriptionOptions userOptions, bool stream = false) |
| 452 | { |
| 453 | AudioTranscriptionOptions copiedOptions = userOptions is null ? new() : userOptions.GetClone(); |
| 454 | |
| 455 | copiedOptions.Model = _model; |
| 456 | |
| 457 | if (stream) |
| 458 | { |
| 459 | copiedOptions.Stream = true; |
| 460 | } |
| 461 | |
| 462 | return copiedOptions; |
| 463 | } |
| 464 | |
| 465 | private void CreateAudioTranslationOptions(Stream audio, string audioFilename, ref AudioTranslationOptions options) |
| 466 | { |
| 467 | options.Model = _model; |
| 468 | } |
| 469 | } |
| 470 | |