LinTo-Skills-Core
Here are the recommended LinTo skill, without them a process can't be guaranteed to work the intended way
Skills
This entry provides information about these skills in the context of the project LinTO
Category - STT
Regroup the STT (Speech To Text) nodes that will allow to transcript an audio file or buffer to text. The following describes the possible inputs and outputs these skill
Input:
{
audio : Buffer Base64,
conversationData : { } //require but can be empty, used if conversation skill is trigger
}
Output :
{
transcript : "text transcript",
confidence : "Confidence score of the transcription",
conversationData : { } //will store the input conversationData
}
LinStt
The LinStt node allow to request to Linagora Speech To Text API The require node configuration are :
- URL : The URL of the LinSTT API
-
Option : Setup the request for a
stream
orfile
, at this moment onlyfile
is supported by LinSTT
Bing
Bing node allow to request to Bing Speech To Text API The require node configuration are :
- Key : Subscription key of Bing
-
Language : Language that you speak to LinTO (recommended based on what LinTO speak
en-US
orfr-FR
- Mode : Recognition mode for the transcription
Category - NLU
Regroup all NLU (Natural Language Understanding) node that will allow to detect which skill to trigger The following describes the possible inputs and outputs these skill
Input :
{
transcript : "text transcript",
confidence : "Confidence score of the transcription",
conversationData : { } //require but can be empty, used if conversation skill is trigger
}
Output
{
transcript : 'text transcript',
nlu : {
intent : 'intentDetected',
entitiesNumber : 1, //number of entities
entities : [{
entity: 'entity type',
value: 'entity name'
}]
},
conversationData : { } //optional json from the previous intention if a conversation is required
}
Tock
The node Tock (The Open Conversation Kit) will retrieve the data from an NLU model has an intent detection The require node configuration are :
-
URL : URL of the API location followed by
/rest/nlp/parse
- Application Name : Application name to use from TOCK
Category - Settings
These category will help to use LinTO. They will mainly use to setup variable or format the data for other node
Prepare
These skill format the MQTT topic to be used has output:
input : <clienId>/fromlinto/<seriaNumber>/nlp/file/<lintoId>
output : <clienId>/tolinto/<seriaNumber>/nlp/file/<lintoId>