I have been trying to warn the user that their session is about to end using dialogflow CX. But I can't get it to respond to me, without the user making a request first.
Is there any way to do this without the user making a request first?
For your use case, if you’ve integrated your agent to your custom application, you can do your own implementation to create a timer that tracks the session time of the user and display a warning on your custom application that the user’s session is ending soon. This approach must be done on your custom application’s side and does not require the use of Dialogflow.
Alternatively, you can use the timer that tracks the session time of the user on your custom application to send a detectIntent request containing the current session ID of the user to your agent to trigger an event. This lets your agent send a response to the user within the same session without having the user to first send a request.
Inside your flow, select the page you want to add a custom event to. Then, click the “Add route type” button to add the Event handlers if not yet added.
Click on the + sign beside “event handlers” field and select any event.
Tick the check box beside “Use custom event”.
Add the name of the custom event you want to use.
Here’s a sample detectIntent request that triggers a custom event using REST API:
Sample URL for detect intent API:
POST
https://dialogflow.googleapis.com/v3beta1/projects/project-id/locations/us/agents/agent-id/sessions/session-id:detectIntent
Make the following replacements for the URL:
project-id: your GCP project ID
agent-id: your agent ID
session-id: your session ID
Sample JSON Request Body should look like this:
{
"queryInput": {
"event": {
"event": "custom-event-name" // custom event to be triggered
},
"languageCode": "en"
},
"queryParams": {
"timeZone": "America/Los_Angeles"
}
}
You can refer below for more information on:
Sessions
Custom events
For more samples on detect intent calls, please refer to this link.
To integrate your agent with your own application, you can use Dialogflow CX’s Client Libraries, Rest API, or RPC API.
Related
I have created a bot with Amazon lex and it's validation & fulfillment with Python and MongoDb.
Bot is working as expected.
Now I am working to integrate my Bot with an ipad application.
Currently my bot asks user about his account id and then bot validate that id in DB and according responses.
Now after integration instead of asking the account id from user, that id should be passed from ipad application to the bot and then bot should responds according.
My question is about this. How can we pass account id from ipad app to bot and then how can my bot or lambda function can get that?
Please suggest if anyone has done similar functionality.
You will want to use requestAttributes or sessionAttributes to pass information like an account ID to your bot with the initial input.
Your bot can then retrieve these from event.requestAttributes or event.sessionAttributes
References: Lex-Lambda Input Event and Response Format
sessionAttributes – Application-specific session attributes that the client sends in the request. If you want Amazon Lex to include them in the response to the client, your Lambda function should send these back to Amazon Lex in the response. For more information, see Setting Session Attributes
requestAttributes – Request-specific attributes that the client sends in the request. Use request attributes to pass information that doesn't need to persist for the entire session. If there are no request attributes, the value will be null. For more information, see Setting Request Attributes
Additional Info
You will want to handle the passing of userInput to your Lex bot yourself in order to include requestAttributes data. To do this, you will need to use PostContent (text or audio input) or PostText (text input only) to send data to your Lex bot.
Your Lex bot will interpret the input and pass along the requestAttributes to your Lambda function, where you can handle the logic based on the Account ID.
Sending user input data as JSON object via PostText:
POST /bot/botName/alias/botAlias/user/userId/text HTTP/1.1
Content-type: application/json
{
"inputText": "Hello Bot",
"requestAttributes": {
"accountID" : "xxxxxxxx"
},
"sessionAttributes": {
"name" : "John Smith"
}
}
To see what Lex will pass to your Lambda Function and how to retrieve the requestAttributes there, see this question where I've answered that in more depth:
AWS Lex Python Codehook references
Following the steps for creating incoming webhooks via add to slack button. The final response will look like this:
{
"access_token": "xoxp-XXXXXXXX-XXXXXXXX-XXXXX",
"scope": "incoming-webhook,commands,bot",
"team_name": "Team Installing Your Hook",
"team_id": "XXXXXXXXXX",
"incoming_webhook": {
"url": "https://hooks.slack.com/TXXXXX/BXXXXX/XXXXXXXXXX",
"channel": "#channel-it-will-post-to",
"configuration_url": "https://teamname.slack.com/services/BXXXXX"
},
....
}
This provides access to team_name, channel the webhook will post to, and a url for configuration.
While this is working great and after initially storing the 'configuration' to my own webapp it displays correctly. But after a someone uses the 'configuration_url' link from the response these previous values like the channel could become obsolete. For example the webhook could be deleted from this link.
My Question is:
Is there a way to reflect this changes in my external app? As far as I can see there is no way to query information for 'incoming webhooks' via an api call from an external web application?
I also could not find any api calls that could change webhook configuration directly without using the provided link.
Its not possible to change the webhook configuration after it is created. The only possible change is that the workspace that installed your Slack app (which contains the webhook) is uninstalled.
If that happens the webhook will no longer work and you will get an error in response.
You can also get notified when you app is uninstalled by listening to the app_uninstalled event.
Or you can also test if your app and webhook is still valid by checking the token you received during the Add to Slack process. Just call the API method auth.test with that token. If it returns the following error than the webhook is no longer valid:
{
"ok": false,
"error": "invalid_auth"
}
What I am doing:
I am integrating Google Picker on my page. This will allow users to select files from their Google Drive to be used in the web app. In the app, people in a group share a common google drive (i.e. they all can select files from account example#email.com) which was created by group admin by his email address. When the admin signs-up for the account we do OAuth and get access_token with refresh_token against our app on google (with offline access enabled). I plan to use the access_token and refresh-token of the admin, on other group user's account when they try to use picker to select files.
What I have done:
I have integrated the Google Picker successfully in my app using the basic code provided in docs. Then to achieve what I wanted, I removed following code from the example code:
gapi.load('auth', {'callback': onAuthApiLoad});
and
function onAuthApiLoad() {
window.gapi.auth.authorize(
{
'client_id': clientId,
'scope': scope,
'immediate': false
},
handleAuthResult);
}
and
function handleAuthResult(authResult) {
if (authResult && !authResult.error) {
oauthToken = authResult.access_token;
createPicker();
}
}
and instead of .setOAuthToken(oauthToken) I pass refreshed access_token directly as string (I get that from my server with an ajax call).
.setOAuthToken("<access_token>")
But every time I call picker.setVisible(true); I get a screen in an iframe saying In order to select an item from your online storage, please sign in.
Problem:
Try to add sign in listener. Listeners provide a way to automatically respond to changes in the current user's Sign-In session. For example, after your startup method initializes the Google Sign-In auth2 object, you can set up listeners to respond to events like auth2.isSignedIn state changes, or changes in auth2.currentUser.
Validating the token might be a possibility before using the token each time but that might add a lot of extra overhead for a rare use-case each time we load the picker and when calling the API endpoints with a token after the re-authentication issue, there was no key about the token being invalid. You can validate a token by making a web service request to an endpoint on the Google Authorization Server and performing a string match on the results of that web service request.
I was wondering : is it possible to intercept the event that checks authorization when attempting to edit a document (eg that returns error Authorization error: Unauthorized access to URL: )
and add my own checks (eg based on specific document ids, external user groupings etc)?
In ContentService.Save() there's an option to include a user ID. I assume the saving is then done in the context of that user and that users permissions.
Also, you can hook into ContentService.Saving event and do whatever checks you want (and cancel the save if your checks "fail").
https://our.umbraco.org/documentation/reference/events/contentservice-events
Our app is importing the next 1000 events from a user's Google calendar API. We ran into the problem where nginx would timeout. To get around this I'm putting the pagination data into a session variable and making separate HTTP requests to the API. This works except for one problem: every time we make a new HTTP request the API asks the user to choose which account they want to use (one user with multiple gmail accounts). I would have thought that the pagination data would include account selection but this is apparently not the case. How can I programmatically select the email account within the HTTP request?
You can store it once
public static void setmCredential(GoogleAccountCredential _mCredential) {
mCredential = _mCredential;
mService = new com.google.api.services.calendar.Calendar.Builder(
transport, jsonFactory, mCredential)
.setApplicationName("YourApplicationName")
.build();
}
And then when caliing pass it like this
new MakeRequestTask(AccountCredential.mService).execute();