This section explains how to verify the correct operation of the Real-time Virtual Appliance using the Websockets Speech API.
The first time that the Real-time Virtual Appliance is started up there are no persistent workers configured. This means that workers will be spun up dynamically to the limit of the maxworkers limit. If you want to pre-allocate workers so that they are ready for incoming streams, you can select a language to configure as a persistent worker. You can find out the list of available languages on your appliance using the REST API:
This will return as JSON output the list of persistent workers (by language code) and the number of instances; initially they will all be zero. The list of supported languages are available on the Speechmatics website https://www.speechmatics.com/language-support/, or you can consult the release notes.
Use the Management API to set persistent_workers with a count of at least 1 and the language code to use. For example, to set French as a persistent worker, use the following method:
{ "persistent_workers": [ { "count": "1", "id": "fr" } ] }This will return an updated list of persistent workers with entry for French (fr) updated to 1.
curl -s -L -X POST "http://${APPLIANCE_HOST}:8080/v1/management/persistentworkers" \
-H 'Accept: application/json' \
-d '{ "persistent_workers":
[ { "count": "1", "id": "fr" } ]
}'
Check that all the Speechmatics services within the appliance are up and running before trying to open a WebSockets connection. The Management API can be used for this.
curl -s -L -X GET "http://${APPLIANCE_HOST}:8080/v1/management/services" \
-H 'Accept: application/json'
The Real-time Virtual Appliance is now ready to use.
The Speech API Guide provides details of how to use WebSockets to stream audio to the Speechmatics engine in real time and obtain transcripts. For all WebSocket communication, ensure that the IP address identified in the steps above is used.