unispeech / unimrcp

Open source cross-platform implementation of MRCP protocol
http://www.unimrcp.org
Apache License 2.0
380 stars 167 forks source link

GDF RECOGNITION-COMPLETE Input-Mode #274

Closed nacartwright closed 4 years ago

nacartwright commented 4 years ago

Results presented back to the MRCP client on the RECOGNITION-COMPLETE should match mode set in voice grammars for compatibility. Cisco (CCX/CCE) voice browsers only accept either voice or dtmf as input modes. This either should be the default or give the option to override for those platforms.

361881: Jul 07 18:01:04.502 EDT %MIVR-SS_MRCP_ASR-7-UNK: [MIVR_ENG_TASKS-31-12-TASK:0x9502f9026_vxml_sample.aef] com.cisco.wf.mrcp.asr.channel.impl.MRCPSpeechChannelImpl MRCPSpeechChannel(Channel id:2, Contact id:38, Task id:40000000038): processRecResults(): NLSLM Result is: <?xml version="1.0"?>
<result>
          <interpretation grammar="session:field4@field.grammar" confidence="100">
                    <instance>
                              <query_text>startDialog</query_text>
                              <action>input.welcome</action>
                              <parameters>
                              </parameters>
                              <all_required_params_present>true</all_required_params_present>
                              <fulfillment_text>Good day! What can I do for you today?</fulfillment_text>
                              <fulfillment_messages>
                                        <text>
                                                  <text>Good day! What can I do for you today?</text>
                                        </text>
                              </fulfillment_messages>
                              <intent>
                                        <name>projects/badgerlinkassistant-a918f/agent/intents/d3c1994b-ff27-468c-a291-497893f569f9</name>
                                        <display_name>Default Welcome Intent</display_name>
                              </intent>
                              <intent_detection_confidence>1</intent_detection_confidence>
                              <language_code>en</language_code>
                                        <output_audio_uri>http://localhost/utterances/umsgdf-a6cf5cdd3de14c79-1108.wav                                        </output_audio_uri>
                    </instance>
                    **<input mode="event">startDialog</input>**
          </interpretation>
</result>

361904: Jul 07 18:01:04.520 EDT %MIVR-SS_VB-7-UNK: [MIVR_ENG_TASKS-31-12-TASK:0x9502f9026_vxml_sample.aef] com.cisco.wfapi.impl.WFTracingLogImpl Task:40000000038 playAndRecognize done.
361905: Jul 07 18:01:04.520 EDT %MIVR-SS_VB-7-UNK: [MIVR_ENG_TASKS-31-12-TASK:0x9502f9026_vxml_sample.aef] com.cisco.wfapi.impl.WFTracingLogImpl Task:40000000038 VBEventHandler::handleEvent event =: Exception=com.cisco.voicebrowser.VBEvent: error.unsupported.format**: Invalid input mode; nested exception is: 
    org.xml.sax.SAXParseException; lineNumber: 24; columnNumber: 41; Unrecognized mode attribute: event**
361906: Jul 07 18:01:04.520 EDT %MIVR-SS_VB-7-EXCEPTION:com.cisco.voicebrowser.VBEvent: error.unsupported.format: Invalid input mode; nested exception is: 
361907: Jul 07 18:01:04.520 EDT %MIVR-SS_VB-7-EXCEPTION:    org.xml.sax.SAXParseException; lineNumber: 24; columnNumber: 41; Unrecognized mode attribute: event
achaloyan commented 4 years ago

The resolution is provided in this post.