Showing 6 results.
Items per Page 50
of 1
First Previous Next Last

CVP Forum

« Back to General Discussion

accessing return values from external grammars

Combination View Flat View Tree View
Threads [ Previous | Next ]
accessing return values from external grammars
Answer
4/2/03 5:54 AM
I'm using Nuance VWS, and found (pg 46 of the user guide) that for external grammars the slot must be called foundation_fld. I did that and got the slot value returned in the [b:d78eedb62c]value [/b:d78eedb62c]element data.
So far so good.
I also would like to get the actual words spoken, in addition to this [b:d78eedb62c]interpretation [/b:d78eedb62c]or slot value.
I note in the audium log that the [b:d78eedb62c]uttereance [/b:d78eedb62c]is logged (the transcription of the actual words spoken). The user element specs (forms element) say [quote:d78eedb62c]The matched utterance and interpretation are saved for the duration of the session. [/quote:d78eedb62c] but there is no documentation of what element or session data contains it.

Also, does audium support multi-slot external grammars?
If so how? (I don't know how to specify a slot name, since in the case above it was implicitly foundation_fld)
Flag

RE: accessing return values from external grammars
Answer
4/5/03 6:00 AM as a reply to Anonymous.
Hello,

Currently, the Audium Form (with Confirm) elements do not store the actual utterance as element data but it is logged in the application log files. Future releases will add functionality and support for more shadow variables for the Form element. However, it is possible to build your own custom element with the current release that will provide your with the functionality you are looking for, or you may also use a vxml insert element.

If you are using VWS 2.0 or NVP with the Form element, you can use a grammar that fills multiple slots. However, in this case, you should NOT have any slot that is named "foundation_fld". The value that is stored in "value" will be a string that contains all the slot name/value pairs. For example, if the grammar has as slot names "lastname", "name" & "firstname", you would get the following result:

voice input - "mary lee"
stored in "value" - {lastname="lee" name="mary lee" firstname="mary"}
utterance logged - mary lee

Most likely, you will want to process the element data at some point after the element has collected the input.

Regards,

Joe
Flag