-
Notifications
You must be signed in to change notification settings - Fork 28
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
RequestIntent interrupts conversation #14
Comments
Hi Martin,
If you use the dashboard to edit the “TextClassifier”, you will find the table that converts intents from conversation into the actual “IIntent” based class. Check the last entry, it should translate any intent “*” into a QuestionIntent, if that’s a RequestIntent then that may be your problem and you can just change it to a QuestionIntent class instead.
…-Richard
From: MartinConlon [mailto:notifications@github.com]
Sent: Sunday, July 16, 2017 11:47 AM
To: watson-intu/self <self@noreply.github.com>
Cc: Subscribed <subscribed@noreply.github.com>
Subject: [watson-intu/self] RequestIntent interrupts conversation (#14)
Hi,
Using the latest develop branch the "Conversation" functionality is different than RC4 and I need some guidance.
Self will dialog with me until I give it something that is recognized as a #request intent, so, for example, using the Intu-Dialog-Starter Kit conversation, I can do the usual "hello", "how are you", "what's your name" but when I ask it "can you stand" it replies with the boilerplate from etc/bootstrap.json, which is ultimately stored in the graph in knowledge.db:
{
"Type_" : "RequestAgent",
"m_EntityFilter" : [ "verb", "noun", "directionPerpendicular", "directionParallel", "adjective" ],
"m_RequestFailedText" : [ "I do not know how to %s." ],
"m_RequestFiles" : [ "shared/self_requests.json" ]
}
I see in the Dashboard (by clicking the "Advanced" link on the top right of its reply) that this is recognized as a "RequestIntent" rather than (for example) "QuestionIntent". The sentence is parsed properly and I see entities tagged in the JSON:
"entities": [
{
"entity": "Possesive",
"value": "you"
},
{
"entity": "Possessive",
"value": "you"
},
{
"entity": "verb",
"value": "stand"
}
]
This is how the Watson Conversation "Try it out" panel parses the text as well, however, if I add a Dialog node for the "#request" intent to respond with (say) "I see you've made a request" Watson's Conversation test panel uses that response, but Self does not.
In other words, I expect Self to use the same Dialog nodes as the Watson Conversation "Try it out" panel, but the RequestAgent seems to be eating the request (and also doesn't know how to act on it).
Is this expected behaviour?
Thank you,
Martin
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub <#14> , or mute the thread <https://github.com/notifications/unsubscribe-auth/APfat5gTeHpPp_4xEPqAVnnQxH8lKeVOks5sOj6BgaJpZM4OZWSX> . <https://github.com/notifications/beacon/APfat1-VJkBiZnHoPc-Oc6_ua6-QmYBVks5sOj6BgaJpZM4OZWSX.gif>
!DSPAM:596b988a10651375212999!
|
Hi Richard, Good call -- it was RequestIntent. Is that the new default? I should have a vanilla install of the develop branch... Thank you, Martin |
Comparing body.json with bootstrap.json, I see that the "catchall" intent changed between the two (RC4 to RC5) from "QuestionIntent" to "RequestIntent." Martin |
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi,
Using the latest develop branch the "Conversation" functionality is different than RC4 and I need some guidance.
Self will dialog with me until I give it something that is recognized as a #request intent, so, for example, using the Intu-Dialog-Starter Kit conversation, I can do the usual "hello", "how are you", "what's your name" but when I ask it "can you stand" it replies with the boilerplate from etc/bootstrap.json, which is ultimately stored in the graph in knowledge.db:
I see in the Dashboard (by clicking the "Advanced" link on the top right of its reply) that this is recognized as a "RequestIntent" rather than (for example) "QuestionIntent". The sentence is parsed properly and I see entities tagged in the JSON:
This is how the Watson Conversation "Try it out" panel parses the text as well, however, if I add a Dialog node for the "#request" intent to respond with (say) "I see you've made a request" Watson's Conversation test panel uses that response, but Self does not.
In other words, I expect Self to use the same Dialog nodes as the Watson Conversation "Try it out" panel, but the RequestAgent seems to be eating the request (and also doesn't know how to act on it).
Is this expected behaviour?
Thank you,
Martin
The text was updated successfully, but these errors were encountered: