While building an app using actions-on-google, i ask user for device location.
I need google assistant to remember if user grants this permission to avoid asking for the same permission again when user uses the app later. How do I do this?
Thanks.
You can't save the permission, but you can save the location as Yuksel has suggested.
This works well if you're expecting the location from a relatively fixed device such as a Google Home, but not as well if you're getting the location from a mobile device.
If you are planning to get the location from a mobile device, you need to request it each time.
Once the user grants you access to the information you need, you can store that information in UserStorage. That way you can check if a user has information stored in their UserStorage, before asking for a permission.
Please not that if you're storing personal data about the user, you may need to ask for their permission,not only to query the data but also to store it as well. As stated in the docs:
Some countries have regulations that require developers to obtain
consent from the user before they can access, or save certain
information (e.g. personal information) in the userStorage. If you
operate in one of these countries and you want to access, or save such
information in userStorage, you must use the Confirmation helper to
ask consent to the user and obtain the consent before you can start
storing such information in userStorage.
You also need to state that you're saving data in your privacy policy:
If you store the user's information, remember to disclose it in your
privacy policy when submitting your project for review.
As discussed in the comments. You CAN'T get users permission once and query their location multiple times. You need a new permission each time you query the data. If you suspect that users location might change every time they use your action (i.e. Actions that mainly target smart phones), you can handle it as a dialog turn. Here's an example:
System: Would you like to know about the weather in $userSavedLocation?
If user says yes, no need for permission. If they say no, then ask for permission again.
Facebook provides an OAuth authentication system for third-party apps but they appear to not provide any fine-grained user management capabilities.
This seems like a security problem. I need to be able to:
Disable delete unused accounts
Disable or delete abusive users
White list users for early access to a new app.
But AFAICT FB provides no tool to do this.
I can probably hook of of the OAuth ID and put an entry in my database for the first two requirements, but I can't figure out how to white list Facebook users.
Google is unhelpful because all the results are about let users manage applications their user account is attached to, and nothing about allowing apps to manage what user accounts they are attached to...
Usually is a bad idea to delete or disable unused accounts, because maybe user was inactive for a long time, this doesn't mean the user deleted or disabled his account.
On the other hand, you can't avoid malicious users join your app via facebook authentication. What you can do is to detect suspicious activities with your own implementation (after Facebook accept them to join your app), and suspend, block or disable those accounts on the server (blacklisting them), after the user has joined client-side and sent the required access-token (optionally request the ID) trying to be accepted by the server.
Whitelisting users may depend on the requirements of your application, you can check profiles data and reject those that doesn't provide the requirements that your app needs (remember to ask the user for the fields that your app requires, like public profile, pictures, etc; these must be accepted by the user before signing in).
Here's how you white list users for an app on Facebook:
Keep the app in Development mode
Add users from your friend network in the Testers role
When you are comfortable, open your application to the world.
AFAICT you cannot white list when your app is published, but that's the point of publishing. Also the whitelist is not a request-grant method, it's invite only.
The other requirement to delete or disable users you must handle yourself in your database. You would tie the user's status to the FB's unique ID.
I can't really figure out what exactly the facebook API and its services are capable of in terms of user management. Given that I would completely rely on facebook for registration and login, does facebook only provide the authentication and registration process and return me data to store in my own database or does it also itself store a list of already registered user accounts which I can query later on so that I could completely outsource user management from my servers?
If you only need to authorize a User, you don´t need to store anything. You only need to store them in a database if you need to identify returning Users, or if you need to use his data while he is not online.
There is no user management though, you can´t just get a list of all users who authorized your App from Facebook. That´s what you have to store on your own. Make sure to implement the possibility to remove a User from your database, Facebook offers a deauthorization-callback in the App settings for that.
I'm aware that there are many questions about Facebook access-tokens and the grief they cause, but despite much experimentation and reading many frustratingly vague blog articles (FB and otherwise), I'm still struggling to get a clear answer to my needs. Let me succinctly break down my process so far:
I am creating a site that, server-side, needs to pull the posts/statuses from a single Facebook Page
I am an admin of that Facebook Page
I have created a Facebook App
Using the Facebook Graph API Explorer, I've generated a short-lived key, connected to my app and my account, that grants permission to my account to view the access-tokens for my pages
I've converted my short-lived key to a long-lived key (60 days) ala scenario 4 from this
And here's where I am stuck. My 60 day key works fine for my server to pull the info needed from the page, but as far I can tell, there's no way to programmatically extend that 60 day key. I also do not know of a way to generate a new short-lived key without manually going to the Facebook Graph API Explorer and creating one.
Since it is my server making the requests to the Facebook API and not a user-based system (where I could easily request that a user authorize the Facebook app again), this creates a very clunky system. Since Facebook deprecated offline_access, is there really no permanent way to have my server pull info from my own page? Will I really have to create a new key by hand and manually update my server with it every 60 days?
Or is there something I'm missing?
Update:
The step-by-step guide that was previously found here has been migrated down into its own answer.
These are the steps that were previously in the question - they have been migrated to this answer.
Having found that it is possible to generate a Facebook Page Access Token that does not expire (with help from #Igy), here is a clear, step-by-step quide for all those looking to the same:
Make sure you are the admin of the FB page you wish to pull info from
Create a FB App (should be with the same user account that is the page admin)
Head over to the Facebook Graph API Explorer
On the top right, select the FB App you created from the "Application" drop down list
Click "Get Access Token"
Make sure you add the manage_pages permission
Convert this short-lived access token into a long-lived one by making this Graph API call:
https://graph.facebook.com/oauth/access_token?client_id=<your FB App ID >&client_secret=<your FB App secret>&grant_type=fb_exchange_token&fb_exchange_token=<your short-lived access token>
Grab the new long-lived access token returned back
Make a Graph API call to see your accounts using the new long-lived access token: https://graph.facebook.com/me/accounts?access_token=<your long-lived access token>
Grab the access_token for the page you'll be pulling info from
Lint the token to see that it is set to Expires: Never!
That should do it. You should now have a Facebook Page Access Token that doesn't expire, unless:
You change your Facebook account password
You lose admin access for the target page
You delete or de-authorize your Facebook App
Any of these will cause the access token to become invalid.
If you are getting (#100) Tried accessing nonexisting field (accounts) on node type (Page), go to the Access Token Debugger, copy the value of User ID, and use it to replace the "me" part of the URL in step 9.
This is covered in the Offline Access deprecation document
Use the 60-day token for the page admin to retrieve a Page Access Token (via /PAGE_ID?fields=access_token or /me/accounts) - the Page access token will not have an expiry time
An approach that works in 2019
I was recently trying to achieve something similar (to the use case described in this thread), but I wanted to make sure to respect Facebook's current policies, so I did a little research and here I'm sharing what I found.
My use case
So, as I said already, my use case is very similar to the one described here; that is:
I'm doing some work for a school district.
They are using a software tool to manage pretty much everything that relates to school transportation.
That tool allows them to send email notifications (to subscribers) when they publish bus delay alerts and school closure alerts.
A lot of people in the community follow the organization on their Facebook page, and that's the only place they look for those alerts.
So an employee of the organization has to manually publish each notification on the Facebook page (in addition to creating it in the transportation software). Moreover, those notifications eventually expire (or are simply deleted before they expire), so the employee has to go back later on to delete them manually as well.
It's a waist of time, so what we are trying to do here is to develop as simple system that periodically polls the software tool's database for new (and expired) notifications and update them (i.e. add and remove) on the Facebook page.
This is, in my view, a legitimate use case, but I wasn't sure how to implement it in a way that's in line with Facebook's policies.
The accepted answer
I followed the steps of the accepted answer and it worked, except that things appear to have changed: now, even though the generated page token does not expire, access to data does expire after around 60 days. You will see that as well if you follow the procedure and inspect the page token in the FB Token Debugger Tool.
Besides, the fact that the generated page tokens are tied to the user account is also unfortunate, because if the user updates his/her password, then the page token also gets invalidated.
How to do it in 2019
After several hours of research, I stumbled upon the following Facebook documentation article: Business Login for Direct Businesses.
It turns out that it is now possible, following the steps described in the above article, to generate a page token that is not associated to any particular Facebook user account and which will not expire (unless the FB App gets deleted or the underlying application token gets deleted, you know...)
So here are the steps and the most important parts:
You need a Business Manager account.
Verification will be required and a digital contract will have to be signed.
You need to add the target Facebook page to that account.
You need to create a Facebook App, and transfer that app to the same Business Manager account as well.
The app will have to go through Facebook's review process, because the following permissions will be needed: manage_pages and publish_pages.
Important note For the posts made using the generate page token to be visible to users other than the application administrators, that app will need to have been published and approved.
You may still experiment with the concept without submitting for review, but the posts won't be publicly visible.
In the Business Manager account (only after your app and page have been added to the account), you need to create what's called a System User, and give that user admin role (or permissions) to the target Facebook page.
A system user is owned by the Business Manager account, and isn't tied to a specific user. My current understanding is that one major use case for a system user is programmatic access to Facebook's Graph API (just what we need).
Then, for that system user, you need to generate a access token (which will be never-expiring). You will be prompted to select for which app. You will then select your target app.
You will then need to use the generated app token to generate a page token, which will also be never-expiring. The procedure is described in this article as:
GET /<PAGE_ID>?fields=access_token&access_token=<SYSTEM_USER_ACCESS_TOKEN>
That's it.
That token will never expire, and it won't be tied to a particular Facebook user, so it's exactly what we need!
The last part is to make sure that your Facebook app gets approved by Facebook. It's in fact the most important part, because the whole procedure is worthless if people don't see our posts.
I wanted to know for sure that I could rely on the above procedure to build something for my client without Facebook rejecting it in the end, so, beforehand (i.e. before starting to work on my client's project), I went through the whole process of creating a page, an app, a Business Manager account, etc. I verified my business. I submitted my app for review. In my request, I was very specific about my use case and emphasized that the app was for "self-use" (i.e. that the organization is developing an app for itself, not for other Facebook users). I got approved without less than 24 hours.
A few other notes about the app review process:
I had to select a platform for the app, so I selected website.
I had to indicate why the app needed the two permissions and how it was going to use them.
I had to indicate why the reviewer would not be able to sign into my app and try it (i.e. because the app will be used by a worker process).
For the mandatory screencasts, I simply presented manual operations in the terminal using the curl utility (to generate the page token and make posts to the Facebook page). I also showed how I was using Business Manager to link the system user to the page and generate a token, and so on.
Again, I was very specific about my use case, and I think that that helped.
I hope this information will be useful to people with similar use cases.
Many thanks to #redhotvengeance for step-by-step guide.
After some time, now there is clearly described in Facebook documentation:
https://developers.facebook.com/docs/facebook-login/access-tokens/expiration-and-extension
Extending Page Access Tokens
Apps can retrieve a page access token from Page admin users when they
authenticate with the manage_pages permission. If the user access
token used to retrieve this page access token is short-lived, the page
access token will also be short-lived.
To get a longer-lived page access token, exchange the User access
token for a long-lived one, as above, and then request the Page access
token. The resulting page access token will not have any expiry time.
You can also copy and past from the app dashboard on facebook.
The steps:
Go to https://developers.facebook.com
Select your app in the top right corner of the page
(pic of what it looks like)
Click on Messenger from the options on the left (it will go to setting automatically) (pic of what it looks like)
Go to the "Token Generation" section in the page. Select what page you want to generate the token for. (pic of what that section looks like)
The copy and past your page token where ever you need it.
Keep in mind that while in theory your token won't expire, that it is directly tied to what ever facebook account your logged into. So say you change your password or you remove the permissions from between your account and your app then your token won't be valid any more.
My product users are using the Facebbok apps for authorisation and generating Keys and tokens.
When I remove the app, my users are still able to connect and fetch the posts from facebook. How is that..? The old tokens should nto work as I assumed , but they are still working.. Does it takes time to reflect.?
Please help me..:)
If you have made a post to facebook through one of your apps, you can also delete that post from the same app. You need to get the post ID (it will be returned when you sucessfully do a post in the first place). Then issue a "DELETE" request. http://developers.facebook.com/docs/reference/api/post/
If you have deleted the app, or the user posted the items themselves (through a share dialog), or you can't find the Post ID, or the user has revoked the "publish_stream" permission, then the post is there to stay.
When one of your users removes an app from the privacy settings interface, the app's access tokens are invalidated. This should be immediate.
However, nothing stops you from going back into the app and authorising it again, and content produced by that app will still be visible on your Timeline. The other answer here shows how you can delete feed stories made by the app, if you want to do that for some reason.
Also, check your users aren't either removing stories by the app from their feed, or hiding the app from Timeline, neither of which removes their ability to use the app or would invalidate the tokens.
You can also programmatically remove the app from a user's account with a HTTP DELETE request to /USER_ID/permissions using the user's access token or an app access token