mikeincb avatar image
mikeincb asked

slot values returned not consistent in spelling

Not sure if this is a bug, or just the way alexa works. If I have a slot of LITERAL type called contactName, and then I include a bunch of sample utterances that include all the names I want to recognize, the recognizer works fine and picks up the name, but the spelling, the actual text value, returned for these names in the slot varies from test to test. That causes problems for my code trying to match on the name returned. For example, is I have something like this in my sample utterances: {trever|contactName}, alexa will return a value of 'trever' sometimes, and sometime 'trevor'. It does this with multiple names that I have tried, and the problem is i don't know all the variations that alexa will return for each name, so I can only test and test and see what it does. This is no tenable for a large set of name. In traditioinal speech grammars, you specify the value to return for each 'slot' so you control it and then your code can always work. With alexa, I would have thought it would have returned the spelling I use (consistently) in my utterance file, but it doesn't seem to. Is this how the slots are supposed to work? I can see how this would be a problem for many applications using different sets of values in their slots. Even the color sample. Say I wanted to do something special if the user spoke 'purple' and so my code would be looking for that, and then alexa returns 'perple' (i know, not likely, but you get my point) . If this is how it is supposed to work, is there some way to 'work around it'? Thanks! Mike
alexa skills kitvoice-user interface
10 |5000

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

0 Answers