/// <summary>
/// Speech-to-text conversion using HTTP mode.
/// </summary>
/// <param name="options">Recognition settings.</param>
/// <param name="mediaStream">Audio stream used for recognition.</param>
/// <param name="cancellationToken">The cancellation token to cancel operation.</param>
/// <exception cref="ObjectDisposedException"></exception>
/// <exception cref="ArgumentNullException"></exception>
/// <exception cref="ArgumentOutOfRangeException"></exception>
/// <exception cref="OperationCanceledException"></exception>
public async Task <SpeechToTextResult> SpeechToTextAsync(SpeechRecognitionOptions options, Stream mediaStream, CancellationToken cancellationToken = default(CancellationToken))
{
if (options == null)
{
throw new ArgumentNullException(nameof(options));
}
if (mediaStream == null)
{
throw new ArgumentNullException(nameof(mediaStream));
}
cancellationToken.ThrowIfCancellationRequested();
ThrowIfDisposed();
var queryParams = new Dictionary <string, string>
{
["uuid"] = _userId.ToUuid(),
["key"] = _apiKey,
["topic"] = options.SpeechModel.GetEnumString(),
["lang"] = options.Language.GetEnumString()
};
var queryString = string.Join("&", queryParams.Select(pair => $"{pair.Key}={pair.Value}"));
var uri = new Uri($"https://{Configuration.RecognitionEndpointAddress}/asr_xml?{queryString}");
var message = new HttpRequestMessage(HttpMethod.Post, uri)
{
Content = new StreamContent(mediaStream),
};
if (message.Content.Headers.ContentLength > 1024 * 1024)
{
throw new ArgumentOutOfRangeException(nameof(mediaStream), message.Content.Headers.ContentLength, "Content length must be less then 1 MB.");
}
message.Content.Headers.ContentType = MediaTypeHeaderValue.Parse(options.AudioFormat.GetEnumString());
message.Headers.TransferEncodingChunked = true;
try
{
var response = await _httpClient.SendAsync(message, cancellationToken).ConfigureAwait(false);
if (!response.IsSuccessStatusCode)
{
return new SpeechToTextResult
{
TransportStatus = TransportStatus.Ok,
StatusCode = response.StatusCode
}
}
;
using (var contentStream = await response.Content.ReadAsStreamAsync())
{
return(new SpeechToTextResult
{
TransportStatus = TransportStatus.Ok,
StatusCode = response.StatusCode,
Result = new UtteranceRecognitionResult(XmlMessageSerializer.DeserializeResponse(contentStream))
});
}
}
catch (HttpRequestException)
{
return(new SpeechToTextResult {
TransportStatus = TransportStatus.SocketError
});
}
}
