search

snarfed / bridgy

📣 Connects your web site to social media. Likes, retweets, mentions, cross-posting, and more...
https://brid.gy
Creative Commons Zero v1.0 Universal
703 stars 52 forks source link

FAQ entry for >10 Instagram likes #722

Closed chrisjonesio closed 7 years ago

chrisjonesio commented 7 years ago

Just a request for an update to the FAQ. Instagram only returns usernames for the first 10 likes of a post. Beyond that it's just a generic number. So if people are wondering why they're only getting 10 likes backfed to their site, I think that's why.

snarfed commented 7 years ago

thanks for the nudge!

out of curiosity, do you know for sure that it's the first 10? or could it be the latest 10, or some other group? e.g. twitter only shows 25 at once, but it's the latest 25, so we still get them all as long as there aren't more than 25 new favorites between posts: https://github.com/snarfed/bridgy/issues/73#issuecomment-168412227

snarfed commented 7 years ago

based on spot checking a few different posts, it looks like it's always the first 10, or at least the 10 never change.

chrisjonesio commented 7 years ago

I'm inclined to agree with you. From my testing and poking around it seems like the first ten likes come through the API as separate users, thus getting pulled by bridgy, then once it rolls over to 11, no more of that. I'm not sure what the API response looks like but I'd guess it turns from an array of like objects each with user info into a single object with total likes and no user info.

There are some articles here and there discussing how IG only shows a number of likes, never names, even under 10. That doesn't appear to be the case now, or for a while now, as all my <10 liked photos (I am SO popular) have usernames listed, not a total number of likes.

I can see why so many devs flipped their desks and walked away from working with IG.

snarfed commented 7 years ago

most devs walked away from IG because they dramatically closed their API to most use cases, including bridgy. so we're not using the API at all now, we're scraping. :/ see #603.

regardless, thanks for reporting!