Add TextStream to Go SDK to improve responsiveness

[wanghengwen]

Currently, in the Cognitive Services Speech SDK for Go, there is no equivalent to the TextStream feature available in other languages like Python or C#. This feature is critical for streaming text inputs in smaller chunks during TTS operations, which significantly improves responsiveness for scenarios involving large or dynamically generated content.

It would be great to have support for TextStream or a similar mechanism in the Go SDK, enabling incremental text input for real-time or large-scale text-to-speech use cases. This enhancement would align the Go SDK's capabilities with those of other languages and improve overall performance and usability.

[wanghengwen]

I have already implemented this feature myself. Please merge the following patch.

001.patch
002.patch

[pankopon]

@yulin-li Please review.

[wanghengwen]

I think there's a bug related to SpeechSDK or the server. my sdk version is : 1.42.0，I tested it with the C++ version and was able to reproduce the same issue. Here's the test code:
request, err := speech.NewTextStreamingRequest() if err != nil { log.Printf("NewTextStreamingRequest error: %v", err) return } defer request.Close() request.InputStream.Write("Hello, How are you?") task := speechSynthesizer.StartSpeakingAsync(request)
If I change the input data for the Request to the following:
request.InputStream.Write("Welcome to the world of AI, I am Ava, your personal assistant. How can I help you today?")
it woks fine

[Zeouterlimits]

Hi, is there any update on this feature getting added to the golang sdk please?

[pwli0755]

any updates? @yulin-li @pankopon