my englich is verry vairy teribel
@romromdc Syn Bot framework follows a community and enterprise edition models. Please visit the developer network to learn more.
@d0vich Could you help us know what sort of assistance you need in setting up Syn Speech
Hi, I followed the instructions to using recognition without grammar, and it works well reading from files, but when I tried to switch to a live speech recognizer, my result hypothesis was consistently an empty string. I'm on a mac and my microphone seems to be working fine. Any insight into what might be happening? Thanks
@mdalzin As your query involves reproduction of test case. We highly recommend that you make use of our Community Forum to seek expert advice.
Or better question is: How do you achieve multiple patterns match from one input with a comma?
@juraj.piar_gitlab If you're working with Patterns in SIML Classic you may specify Comma as a child <Item> element of Splitter tag
ok….. so... the way I’ve figured it out is to map the json from GetState to c# objects , change the variable that needs changing, then serialise it back to json and pass that to SetState. It seems to work but also seems kinda expensive. Is this the intended way of setting a variable in the user state? The GetState’s output, for example, has some unparsable keys (in "$type"), so I’m not sure if my approach does not brake anything in the bot when passing the re-serealised state. thanks
@juraj.piar_gitlab Please note that from version to version we do keep optimizing the serialization process to yield faster results. So there is a possibility that if you modify the JSON value externally then it may not work in a newer version of the framework.
e.g. “‘variables’: {
‘$values’: [
{
’name’: ’Name’,
‘values’: {
‘$values’: [ ‘John’ ]
}
]
}"
‘$values’: [
{
’name’: ’Name’,
‘values’: {
‘$values’: [ ‘John’ ]
}
]
}"
@juraj.piar_gitlab The framework is built around the design concept that a single user message holds a single intent. This is one of the reasons for the SuggestedIntent property on EvaluationResult class. However, the execution of intents is left solely to the developer. If the developer wishes to execute more than one intent, they can check the Score and call the Invoke() method.
@senthilnathan0203 Could you please elaborate your query. What do you mean by "go to" previous context? You could also pose the same question in our Developer Forum to receive answers from experts.
@senthilnathan0203 Coincidentally if you are referring to Contextual Reversal, the feature is new to 3.4.0 and is currently available under the release candidate version.
@SynOfficial I'll describe you clearly. When the user types "I want to know my name?"(Entity: want,my,name) In this scenario the bot will reply the user name. But when the user types, "I don't want my name".. Eventhough the bot is replying the same here. Is there any way to handle this easily.?
@senthilnathan0203 You can either use
OscovaBot.Configuration to set global minimum intent scores. For specifying isolated values you can refer the tutorial here
Any suggestion using oscova on the same
