Speech to text not working with websockets

[pandrews11]

Describe the bug
Speech to text is not working with websockets. Works fine with synchronous recognize.

To Reproduce
Here is the code:

#!/usr/bin/env ruby

require "ibm_watson/speech_to_text_v1"
require "ibm_watson/websocket/recognize_callback"
require "ibm_watson/authenticators"
require "json"

authenticator = IBMWatson::Authenticators::IamAuthenticator.new(
  apikey: ENV['API_KEY']
)

speech_to_text = IBMWatson::SpeechToTextV1.new(
  authenticator: authenticator
)

speech_to_text.service_url = ENV['API_URL']

File.open(Dir.getwd + "/resources/speech.wav") do |audio_file|
  recognition = speech_to_text.recognize(
    audio: audio_file,
    content_type: "audio/wav",
    timestamps: true,
    word_confidence: true
  ).result
  puts JSON.pretty_generate(recognition)
end

# Example using websockets
class MyRecognizeCallback < IBMWatson::RecognizeCallback
  def initialize
    super
  end

  def on_transcription(transcript:)
    puts JSON.pretty_generate(transcript)
  end

  def on_connected
    puts "Connection was successful"
  end

  def on_error(error:)
    puts "Error received: #{error}"
    puts "Error message: #{error.message}"
  end

  def on_inactivity_timeout(error:)
    puts "Inactivity timeout: #{error}"
  end

  def on_listening
    puts "Service is listening"
  end

  def on_transcription_complete
    puts "Transcription completed"
  end

  def on_hypothesis(hypothesis:)
    puts hypothesis.to_s
  end

  def on_data(data:)
    puts data.to_s
  end
end

mycallback = MyRecognizeCallback.new
File.open(Dir.getwd + "/resources/speech.wav") do |audio_file|
  speech_to_text.recognize_using_websocket(
    audio: audio_file,
    recognize_callback: mycallback,
    content_type: "audio/wav"
  ).start
end

Expected behavior
I am expecting to see a successful connected message from the recognize callback.

Actual behavior

Error received: #<Faye::WebSocket::API::ErrorEvent:0x00000000029eaa28>
Error message: Error during WebSocket handshake: Unexpected response code: 401

Desktop (please complete the following information):

• Docker
• ruby-2.4.3
• ibm_cloud_sdk_core-1.1.1
• ibm_watson-1.2.0

Additional context
By adding the authentication step before sending websocket request I was able to make it work:

diff --git a/lib/ibm_watson/speech_to_text_v1.rb b/lib/ibm_watson/speech_to_text_v1.rb
index 7451924..ba3c20a 100644
--- a/lib/ibm_watson/speech_to_text_v1.rb
+++ b/lib/ibm_watson/speech_to_text_v1.rb
@@ -567,6 +567,7 @@ module IBMWatson

       require_relative("./websocket/speech_to_text_websocket_listener.rb")
       headers = {}
+      @authenticator.authenticate(headers)
       headers = conn.default_options.headers.to_hash unless conn.default_options.headers.to_hash.empty?
       service_url = @service_url.gsub("https:", "wss:")
       params = {

Thanks!

[mamoonraja]

Hi @pandrews11, thanks a lot for opening the issue, we will have the fix out soon.

[mamoonraja]

Fixed in #94