Found a weird bug, and I don’t think it is Jovo. I think the Alexa NLU is doing it. Normally, the slots are all stored as lowercase. Here’s an example from my logs:
"stainType": {
"name": "stainType",
"value": "red wine",
"resolutions": {
"resolutionsPerAuthority": [
{
"authority": "amzn1.er-authority.echo-sdk.amzn1.ask.skill.21c97416-c086-498c-a0a5-ea3bf4918ca5.xxxxxxxxx",
"status": {
"code": "ER_SUCCESS_MATCH"
},
"values": [
{
"value": {
"name": "red wine",
"id": "07d4b2aa6a900c554874fa6d09b2243a"
}
}
]
}
]
},
"confirmationStatus": "NONE",
"source": "USER"
}
The ‘red wine’ slot works fine. But today I noticed my app wasn’t responding. I have another slot ‘mud’ that seems be capitalized. Also, it interpreted it as ‘Mudd’ - with the double-d on the end. I’ve added this weird spelling to the inputs just in case:
"stainType": {
"name": "stainType",
"alexaSkill": {
"name": "stainType",
"value": "Mudd",
"resolutions": {
"resolutionsPerAuthority": [
{
"authority": "amzn1.er-authority.echo-sdk.amzn1.ask.skill.21c97416-c086-498c-a0a5-ea3bf4918ca5.xxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
"status": {
"code": "ER_SUCCESS_MATCH"
},
"values": [
{
"value": {
"name": "mud",
"id": "c95e3dd9ec9230450a558cb17be2106d"
}
}
]
}
]
},
"confirmationStatus": "NONE",
"source": "USER"
},
"value": "Mudd",
"key": "mud",
"id": "c95e3dd9ec9230450a558cb17be2106d"
}
Strangely, from this it’s showing “ER_SUCCESS_MATCH” so it sees this as ‘mud’ anyway. But it messes up the app because I need that ‘mud’ value to process other things. All my inputs are lowercase in the en-GB.json
model file:
{
"value": "red wine",
},
Also here:
{
"value": "mud",
"synonyms": [
"mudd",
"dirt",
"soil"
]
},
Any ideas why this might be happening?
Many thanks,
Simon