how to get variable from extensions.conf

[Natphitchayuk]

in extensions.conf same = n,Set(SPEECH_ENGINE(VOSK_CONFIG)={"config" : {""}}

how to get VOSK_CONFIG in res_speech_vosk.c

I have tried to use ast_variable_retrieve but it has not worked for me.

[Natphitchayuk]

#define EXTENSIONS_CONFIG "/etc/asterisk/extensions.conf"
struct ast_flags config_flags = { 0 };
struct ast_config *cfg = ast_config_load(EXTENSIONS_CONFIG, config_flags);
const char *value = NULL;
if((value = ast_variable_retrieve(cfg, "from-internal", "VOSK_CONFIG")) != NULL) {
	ast_log(LOG_DEBUG, "from-internal.config=%s\n", value);
	vosk_engine.config = ast_strdup(value);
}

[nshmyrev]

Something like this, yes. You'd better pass arguments as parameters to the engine, not through variable. You can substitute variable in parameters in speech command if you need them.

[Natphitchayuk]

Can you please give me SOME examples?

[nshmyrev]

You'd better describe what you want to implement overall, why do you need to change config.

[Natphitchayuk]

I added max_speech_timeout & complete_timeout to asr_server.py. (vosk-server)

        message = await websocket.recv()
        if isinstance(message, str) and 'config' in message:
            jobj = json.loads(message)['config']
            logging.info("Config %s", jobj)
            if 'phrase_list' in jobj:
                phrase_list = jobj['phrase_list']
            if 'sample_rate' in jobj:
                sample_rate = float(jobj['sample_rate'])
            if 'words' in jobj:
                show_words = bool(jobj['words'])
            if 'max_alternatives' in jobj:
                max_alternatives = int(jobj['max_alternatives'])
            if 'max_speech_timeout' in jobj:
                max_speech_timeout = now + datetime.timedelta(seconds=float(jobj['max_speech_timeout']))
            if 'complete_timeout' in jobj:
                complete_timeout = datetime.timedelta(seconds=float(jobj['complete_timeout']))
            continue

and send config with ast_websocket_write_string (res-speech-vosk.c)

vosk_engine.config = ast_strdup("{\"config\" : { \"max_speech_timeout\" : 10, \"complete_timeout\" : 2.5 } }");
ast_websocket_write_string(vosk_speech->ws,vosk_engine.config);

how to obtain VOSK_CONFIG (extensions.conf) on res_speech_vosk.c

same = n,Set(SPEECH_ENGINE(VOSK_CONFIG)={"config" : { "max_speech_timeout" : 20, "complete_timeout" : 5 } })

[Natphitchayuk]

I want to set a different timeout for each input. For example,

1. When I ask for address the timeout should be 10 seconds
2. When I ask for phone number the time should only be 5 seconds.

exten   = 5,1,Answer
same    = n,Wait(1)
same 	= n,Set(SPEECH_ENGINE(VOSK_CONFIG)={"config" : { "max_speech_timeout" : 20, "complete_timeout" : 20 } })
same    = n,SpeechCreate
same    = n,SpeechBackground(hello, 30)
same    = n,Verbose(0,Result      : ${SPEECH(results})
same    = n,Verbose(0,Status      : ${SPEECH(status)})
same    = n,Verbose(0,Spoke       : ${SPEECH(spoke)})
same    = n,Verbose(0,Transcribe  : ${SPEECH_TEXT(0)})
same    = n,SpeechDestroy
same 	= n,Set(SPEECH_ENGINE(VOSK_CONFIG)={"config" : { "max_speech_timeout" : 5, "complete_timeout" : 5 } })
same    = n,SpeechCreate
same    = n,SpeechBackground(hello, 30)
same    = n,Verbose(0,Result      : ${SPEECH(results})
same    = n,Verbose(0,Status      : ${SPEECH(status)})
same    = n,Verbose(0,Spoke       : ${SPEECH(spoke)})
same    = n,Verbose(0,Transcribe  : ${SPEECH_TEXT(0)})
same    = n,SpeechDestroy

[helpinghandindia1]

Hello @Natphitchayuk Have you successfully implemented vosk server with asterisk. we facing some issue with integration. pls share steps.