Google v2 issue with providing parameters

[Dmitrii Fedorov]

Hello,

can we get some assistance on how to setup google plugin with v2 api version in order to make it work without recognizers.


Currently API v2 requires to set us "recognizer-id" but as per google docs we can set it to "_" which will generate


projects/apifonica/locations/global/recognizers/_
recognizer path and allow to work with any language without preset.
Below the grammar file generated that returns and error

<?xml version="1.0" encoding="UTF-8"?>
mrcp              | <grammar xmlns="http://www.w3.org/2001/06/grammar" xml:lang="de-DE" version="1.0" mode="voice" root="input">
mrcp               |   <meta name="model" content="default"/>
mrcp               |   <meta name="single-utterance" content="false"/>
mrcp               |   <meta name="recognizer-id" content="_"/>
mrcp               |   <meta name="api" content="v2"/>
mrcp               |   <meta name="speech-complete" content="false"/>
mrcp               |   <rule id="input">
mrcp               |     <one-of>
mrcp               |       <item>sure</item>
mrcp               |       <item/>
mrcp               |     </one-of>
mrcp               |   </rule>
mrcp               | </grammar>

Error log and config



mrcp               | 2024-03-18 15:06:14:041338 [INFO]   Speech Detector State Transition NO-INPUT -> IN-PROGRESS [2680 ms] <b56c54a8adb84941>
mrcp               | 2024-03-18 15:06:14:041347 [INFO]   Start Input Timer [20000 ms] <b56c54a8adb84941>
mrcp               | 2024-03-18 15:06:14:041549 [INFO]   Send Config <b56c54a8adb84941@gsr>
mrcp               | {"recognizer":"projects/apifonica/locations/eu/recognizers/_","streamingConfig":{"config":{"features":{"maxAlternatives":1},"explicitDecodingConfig":{"encoding":"LINEAR16","sampleRateHertz":8000,"audioChannelCount":1}},"streamingFeatures":{"interimResults":true}}}
mrcp               | 2024-03-18 15:06:14:093661 [INFO]   Received Response: status [0] type [SPEECH_EVENT_TYPE_UNSPECIFIED] result-count [0] <b56c54a8adb84941@gsr>
mrcp               | {}
mrcp               | 2024-03-18 15:06:14:101400 [INFO]   Input Complete [stopped] size=6400 bytes, dur=490 ms <b56c54a8adb84941@gsr>
mrcp               | 2024-03-18 15:06:14:101708 [NOTICE] RDRM total file entries [15] timers [15]
mrcp               | 2024-03-18 15:06:14:101719 [INFO]   Process RECOGNITION-COMPLETE Event <b56c54a8adb84941@speechrecog> [2]
mrcp               | 2024-03-18 15:06:14:101721 [INFO]   State Transition RECOGNIZING -> RECOGNIZED <b56c54a8adb84941@speechrecog>
mrcp               | 2024-03-18 15:06:14:101751 [INFO]   Send MRCPv2 Data 10.168.8.111:1545 <-> 10.168.8.201:47878 [138 bytes]
mrcp               | MRCP/2.0 138 RECOGNITION-COMPLETE 2 COMPLETE
mrcp               | Channel-Identifier: b56c54a8adb84941@speechrecog
mrcp               | Completion-Cause: 006 recognizer-error
mrcp               |
mrcp               |
mrcp               | 2024-03-18 15:06:14:101822 [WARN]   gRPC Status: Invalid `config.language_codes`: field must be non-empty. <b56c54a8adb84941@gsr>

 

 

The question is how can we properly set config.language_codes in order to use recognizer without creating recognizer for every language or use of other workarounds?

Apifonica belongs to SmartTel Plus OÜ. This is a confidential e-mail intended solely for the use of the addressee(s). 
Unauthorized publication, use, dissemination or disclosure of this message, either in whole or in part is strictly prohibited. If you have received this message in error please send it back to the sender and delete it.