Prior to today, I used the following URL to get the Facebook share/like/comment count of a URL:
https://api.facebook.com/method/links.getStats?format=json&urls=http://stackoverflow.com/
Today, Facebook removed this feature. So now I'm using this:
https://graph.facebook.com/v2.7?emc=rss&fields=og_object{engagement},share&access_token=<ACCESS_TOKEN_GOES_HERE>&id=http://stackoverflow.com/
Which outputs:
{
"og_object": {
"engagement": {
"count": 45267,
"social_sentence": "45K people like this."
},
"id": "10150180465825637"
},
"share": {
"comment_count": 12,
"share_count": 45267
},
"id": "http://stackoverflow.com/"
}
The problem is, share_count is the total of likes + comments + shares (as far as I know).
Is there a way to get the number of likes, comments, and shares separately?
Looks like a bug in fb-api.
My solution:
receive the number of shares/comments as you describe in your question with
graph.facebook.com/?fields=og_object{id},share&id=https://stackoverflow.com/
save shares/comments count
save fb object-id of url
og_object->id
get likes count with (max limit is 1000, then you can use paging):
graph.facebook.com/OBJECT_ID/likes?access_token=ACCESS_TOKEN&pretty=1&limit=1000
UPD 2016-08-22
I found a solution that allows you to get the number of likes/reposts/comments in one step:
https://graph.facebook.com/?fields=og_object{likes.limit(0).summary(true)},share&ids=http://google.com,http://twitter.com
It is not currently possible.
Facebook is adding likes/shares/comments for URLs and returning them as shared_count.
Not possible to return correct number of individual likes/shares/comments, and that is not clear in the Graph API documentation. Not clear if it is a bug or a documentation error.
The old Rest API that provided this data was turned off on the 18th August.
There is a comment from a Facebook engineer explaining this in reply to a bug report, in the answer to this Stack Overflow question:
Getting Facebook Share, Like and Comment Counts for a Given URL with API Graph v2.6
You can also subscribe to this bug report at Facebook, but is a bit old with still no solution:
https://developers.facebook.com/bugs/748651458568287/
I am rather new to Facebook Graph API and I'm trying to find out if a user has liked a specific post.
Right now, in order to achieve this, I get all the likes for that post and then I iterate through them to find a match. But this procedure requires a lot of requests when the number of likes is too high, as the maximum number of likes per request appears to be 100.
Is there any better way to check if the user has liked the post?
Request the summary of the likes (post_id?fields=likes.limit(0).summary(1)), then you will get a data structure that looks like this:
{
"likes": {
"data": [
],
"summary": {
"total_count": 1,
"can_like": true,
"has_liked": false
}
},
"id": "…"
}
The properties can_like and has_liked reflect the values for the user who’s access token you used to make the request.
UPDATE
The latest version of Graph API today is v2.8. In this version you can get the reactions also. It is better to use reactions than likes as reactions also includes the Love, WOW and HAHA etc reactions which likes does not.
You can add the field for reactions's summary instead of likes's summary as below
fields=reactions.limit(0).summary(true)
Using this you will get the reactions summary as below
"reactions": {
"data": [
],
"summary": {
"total_count": 6,
"viewer_reaction": "LOVE"
}
}, ......
Notice the "viewer_reaction": "LOVE" which means you have given a reaction to the post which is LOVE. It's value could be any one of LIKE, LOVE, HAHA, WOW, SAD, ANGRY, NONE. The NONE means you have not liked the post.
Introduction
/me/books.reads returns books[1].
It includes an array of books and the following fields for each book:
title
type
id
url
Problem
I'd like to get the author name(s) at least. I know that written_by is an existing field for books.
I'd like to get ISBN, if possible.
Current situation
I tried this:
/me/books.reads?fields=data.fields(author)
or
/me/books.reads?fields=data.fields(book.fields(author))
But the error response is:
"Subfields are not supported by data"
The books.reads response looks like this (just one book included):
{
"data": [
{
"id": "00000",
"from": {
"name": "User name",
"id": "11111"
},
"start_time": "2013-07-18T23:50:37+0000",
"publish_time": "2013-07-18T23:50:37+0000",
"application": {
"name": "Books",
"id": "174275722710475"
},
"data": {
"book": {
"id": "192511337557794",
"url": "https://www.facebook.com/pages/A-Semantic-Web-Primer/192511337557794",
"type": "books.book",
"title": "A Semantic Web Primer"
}
},
"type": "books.reads",
"no_feed_story": false,
"likes": {
"count": 0,
"can_like": true,
"user_likes": false
},
"comments": {
"count": 0,
"can_comment": true,
"comment_order": "chronological"
}
}
}
If I take the id of a book, I can get its metadata from the open graph, for example http://graph.facebook.com/192511337557794 returns something like this:
{
"category": "Book",
"description": "\u003CP>The development of the Semantic Web...",
"genre": "Computers",
"is_community_page": true,
"is_published": true,
"talking_about_count": 0,
"were_here_count": 0,
"written_by": "Grigoris Antoniou, Paul Groth, Frank Van Harmelen",
"id": "192511337557794",
"name": "A Semantic Web Primer",
"link": "http://www.facebook.com/pages/A-Semantic-Web-Primer/192511337557794",
"likes": 1
}
The response includes ~10 fields, including written_by which has the authors of the book.
Curiously, link field seems to map to url of the books.reads response. However, the field names are different, so I'm starting to loose hope that I would be able to ask for written_by in books.reads request..
The only reference that I've found about /me/books is https://developers.facebook.com/docs/reference/opengraph/object-type/books.book/
This is essentially about user sharing that he/she has read a book, not the details of the book itself.
The data structure is focused on the occasion of reading a book: when reading was started, when this story was published, etc.
[1] I know this thanks to How to get "read books"
FQl does not looks very promising – although you can request books from the user table, it seems to deliver just a string value with only the book titles comma-separated.
You can search page table by name – but I doubt it will work with name in (subquery) when what that subquery delivers is just one string of the format 'title 1,title 2,…'.
Can’t really test this right now, because I have read only one book so far (ahm, one that I have set as “books I read” on FB, not in general …) – but using that to search the page table by name already delivers a multitude of pages, and even if I narrow that selection down by AND is_community_page=1, I still get several, so no real way of telling which would be the right one, I guess.
So, using the Graph API and a batch request seems to be more promising.
Similar to an FQL multi-query, batch requests also allow you to refer data from the previous “operation” in a batch, by giving operations a “name”, and then referring to data from the first operation by using JSONPath expression format (see Specifying dependencies between operations in the request for details).
So a batch query for this could look like this,
[
{"method":"GET","name":"get-books","relative_url":"me\/books?fields=id"},
{"method":"GET","relative_url":"?ids={result=get-books:$.data.*.id}
&fields=description,name,written_by"}
]
Here all in one line, for easier copy&paste, so that line breaks don’t cause syntax errors:
[{"method":"GET","name":"get-books","relative_url":"me\/books?fields=id"},{"method":"GET","relative_url":"?ids={result=get-books:$.data.*.id}&fields=description,name,written_by"}]
So, to test this:
Go to Graph API Explorer.
Change method to POST via the dropdown, and clear whatever is in the field right next to it.
Click “Add a field”, and input name batch, and as value insert the line copy&pasted from above.
Since that will also get you a lot of “headers” you might not be interested in, you can add one more field, name include_headers and value false to get rid of those.
In the result, you will get a field named body, that contains the JSON-encoded data for the second query. If you want more fields, add them to the fields parameter of the second query, or leave that parameter out completely if you want all of them.
OK, after some trial-and-error I managed to create a direct link to Graph API Explorer to test this – the right amount of URL-encoding to use is a little fiddly to figure out :-)
(I left out the fields parameter for the second operation here, so this will give you all the info for the book that there is.)
As I said, I only got one book on FB, but this should work for a user with multiple books the same way (since the second operation just takes however many IDs it is given from the first one).
But I can’t tell you off the top of my head how this will work for a lot of books – how slow the second operation might get with that, when you set a high limit for the first one. And I also don’t know how this will behave in regard to pagination, which you might run into when me/books delivers a lot of books for a user.
But I think this should be a good enough starting point for you to figure the rest out by trying it on users with more data. HTH.
Edit: ISBN does not seem to be part of the info for a book’s community page, at least not for the ones I checked. And also written_by is optional – my book doesn’t have it. So you’ll only get that info if it is actually provided.
It's easy to get people who have liked post: https://graph.facebook.com/19292868552_10150189643478553/likes
How can I get people who have shared this post ?
Using the Graph Explorer with a post that has shares:
https://graph.facebook.com/20531316728_10151733211311729?fields=shares
The output is:
{
"shares": {
"count": 1806
},
"id": "20531316728_10151733211311729",
"created_time": "2013-02-10T12:59:25+0000"
}
Of course if you remove the fields parameter you see all the fields associated with that object.
EDIT: The post on Facebook - https://www.facebook.com/photo.php?fbid=10151733211231729&set=a.376995711728.190761.20531316728&type=1
Consider the following piece of dummy data returned from the Graph API for Facebook's Commenting Social Plugin:
{
"id": "387539086359_14842802",
"from": {
"name": "Manja Elis",
"id": "1168990251"
},
"message": "http://www.facebook.com/pages/Manja-Elis/151029671610040",
"created_time": "2010-11-24T07:06:06+0000",
"comments": {
"data": [
{
"id": "494442921359",
"from": {
"name": "J Reigy T II",
"id": "100001234540335"
},
"message": "~ www.jareta.name ~",
"created_time": "2011-01-30T00:04:30+0000"
}
],
"count": 1
}
}
How can I generate the permalink to this specific comment so that it can be bumped to the top of the discussion?
I know that the first id# "387539086359_14842802" is the ID of the comment thread as a whole, but now I need the specific ID of the actual first comment so that I can properly link to it. The problem is, that 3rd piece of ID information is nowhere to be found.
What I find funny is that the reply comment with id# "494442921359" is the correct ID for that reply, but there is no equivalent offered for the original post - why??? Worse yet, Facebook ignores links to any comments that aren't the first in a new sub-thread so linking to the reply instead is useless to me.
So to clarify:
Comment id of first post in the thread is : 387539086359_14842802_????
Comment id of the reply to this post would be: 387539086359_14842802_494442921359 but unfortunately FB ignores reply IDs so it is useless to me.
Can anyone tell me how I programatically get this ???? ID for permalink generation purposes?