I am trying to add Additional invocation phrases on the Directory Information tab for a Smart Action on the Actions Console. After filling phrases in and pressing the Save button I receive a notification "Assistant app information saved successfully". But filled in phrases do not appear in the list under Additional invocation phrases. Also, I cannot see filled in phrases in the "Explore all actions" -> "What it can do" tab of Google Assistant for the Smart Action. I can find there only auto-generated phrases.
I've tried to fill in phrases with and without Action name or Device name, but nothing worked.
How can I add Additional invocation phrases to Smart Action? Auto-generated phrases partially don't match the functionality of my Action.
I cannot see filled in phrases in the "Explore all actions" -> "What it can do" tab of Google Assistant for the Smart Action. I can find there only auto-generated phrases.
This functionality is not currently supported for smart home Actions. Example invocations on the "Explore" pages are auto-generated from the device type/trait information provided by your fulfillment. Feel free to log this as a feature request in the public issue tracker.
Auto-generated phrases partially don't match the functionality of my Action.
If there is a concern here, you can file an issue on the public issue tracker with your agent (project) id and we can forward it to the certification team for review.
Related
I have two queries related to Alexa Skill -
I am using the Account Linking feature, But when I logged in with one account I am unable to login with another account, any idea? for more information on this please refer here
https://forums.developer.amazon.com/questions/233563/how-to-delinking-an-linked-account-from-my-alexa-s.html
For User Full Name I have created a custom Slot, but unable to fetch Exact Full Name spoken to Echo show device by the user, for example, Use speaks Danny Stone but Alexa device capture it as "Danny", Dannys tone", "Danni" etc.
Any help would be appreciated
For the name, do a two or three slot response pattern on that intent. First name can use Amazon's built in first name slot type. For the other(s), you'll need custom slots with lots of examples.
For unlinking an account, check out this option: https://developer.amazon.com/en-US/docs/alexa/account-linking/skill-activation-api.html#disable-and-unlink
I read your query on the other forum. I would suggest you clear the cache of your Alexa app if you are logging on to your phone. If it does not help then open an incognito browser and disable enable the skill and link the desired account.
Use custom AMAZON.FirstName intent or AMAZON.Person intent for the names. I am not sure that it will be able to recognize some tricky names.
how to add middleware/intercepter for every request in on dialogflow webhook server while using Action-on-google library and working with dialogflow application,
the purpose is i want to authenticate on every request that key is still valid or not, and also i want to check if that user is already managing a group then get all members of group and put in /userEntity,
now i'm doing this in wellcome intent, so when user say talk to xyz app in wellcome intent i check in database if user is managing a group then get all members of that specific group and put in user entity,
but this logic become trash when user directly say a command such as if user don't say talk to my xyz app and instead he says ask my xyz app john wink is present or not then app is unable to recognize this name, note that i cannot user system name entity because in my case these are not english names
for now i have restricted direct commands with context combination but it is not good e.g: user cannot say direct command unless WELCOME_DONE context which is context out of wellcome intent
The "direct command" you mention (ask my xyz app john wink is present or not) is what we call an "action invocation phrase". Since you're depending on a user entity in order to extract a name from the action phrase, you'll only be able to do this successfully if another intent has recently been matched, since the user entities you add are only available for 30 minutes.
One way around this might be the following. We are going to do something clever to run the user query through Dialogflow a second time, after adding the user entity.
In the intent that handles your "action invocation", use #sys.any (which is essentially a wildcard entity) to capture the name. Enable the webhook for this intent.
Create another intent with User says examples that match the expected "action phrase", similar to the first intent, but don't add it to your "Additional triggering intents". Enable the webhook for this intent.
Create a function in your webhook that handles the intent from Step 1.
In this function, look up the user and their group and add the members via /userEntity.
Now that the user entity has been created, call the Dialogflow /query API with whatever the user said.
Since the user entity has now been created, the intent created in Step 2 will be matched, so the /query API response will contain the name from the entity you created in Step 4.
Use the name to generate whatever response you wanted to generate.
Back in the day Facebook used to let you define URLs that were something like http://facebook.com/send_a_message/user/55/pre_fill/Hi there. When you clicked that link Facebook would open up its Messenger with a "Hi there" message ready to be sent to user 55.
But then Facebook decided that "pre-filling" was bad, and outlawed it across all their APIs. Now there are a bunch of different ways to initiate a Messenger chat session, but they all amount to "start a chat with user 55", not "start a chat with user 55 using some information".
The thing is, I'd like to somehow pass information along with the link. For instance, I might want to have three different links on three different pages, and I want my chatbot to be able to respond differently based on which page the user came from. Since I can't "pre-fill" a message (eg. "Hello, I just came from page A"), I'm wondering if there is any other way to pass auxiliary information.
TLDR
I'm wondering if there's any way I can possibly provide a user with a link that, when they click it, starts up a chat session with my chatbot user, and somehow passes information to it. Using an intermediate proxy (ie. having the link go to my server, and then have my server start the chat somehow) would work, but only if I can associate the information with the user (ie. it doesn't help to know "page A" if I can't associate that with the user once they start chatting).
After researching further it would appear that the answer to my question is no: there is currently no way to pass-along supplemental information.
But here's the good news: when I filed a bug about this (which it turns out I should have filed as a feature request, sorry Facebook) I got a response saying something to the effect of "we couldn't tell you that we are currently developing this, even if we were", which I'm going to optimistically choose to interpret as a sign hope means that this functionality may be coming in the future.
I need to log some actions in Drupal. i.e: published content type of type X, deleted comment y, assigned role W to user Z etc.
It could be very easy to do this using watchdog, but i don't want to do it because i want to keep this logs without deleting rows and because i don't want to mix business actions with error logs.
I think there must be a module that does this, maybe also only an API module.
Ideally it should use a new entity to store logs and it should be integrated vith the views module.
Thank you in advance
I've been looking for something similar. But I just don't see anything with that type of subscription. I did find this: http://drupal.org/project/logging_alerts . It adds e-mail logging and allows rules to add watchdog statements.
With the idea of rules a heartbeat monitor may actually do the job as well. http://drupal.org/project/heartbeat It is supposed to give you Facebook style home feeds but you can choose what is logged to it with rules so it will only log node creation info etc.
Just some thoughts that I was going through when I stumbled upon this question in google.
Scenario:
I get a mail from Foo Bar about a problem. I add the issue to Bugzilla and would like Foo Bar to be the reporter of the issue so it will be possible to do searches on issues reported by this individual etc.
I understand that I most likley will have to create dummy users for every reporting person since Bugzilla only works with the users in it's system. But if I could live with that, would it be possible to accomplish this?
The only way to do it would be to actually log in as the dummy user when reporting the bug. (Or if you have permission, use the impersonation feature.)
To make the reporter field editable on the bug submission form, you'd have to edit the source code.
The impersonation feature is documented here. Quoting briefly:
You may start a session by going to
the Edit Users page, Searching for a
user and clicking on their login. You
should see a link below their login
name titled "Impersonate this user".
Click on the link. This will take you
to a page where you will see a
description of the feature and
instructions for using it. After
reading the text, simply enter the
login of the user you would like to
impersonate, provide a short message
explaining why you are doing this, and
press the button.
There is no way in the web interface to do this.
It can be achieved via the XMLRPC mehtod only, but this is not suitable for most use cases:
bugzilla.changeAssignment($bugid, { reporter => new#reporter.com },
$username, $password);
Source: https://bugzilla.redhat.com/show_bug.cgi?id=432238