[question] export all likes on a profile

[nikartz]

Is there a way to export all profiles, that liked pictures on a given profile? I'm thinking of a way similar to this approach for followers. I've played around with get_likes from structures.py, but I've had no luck. Maybe someone could help!

If anyone is interested: I am trying to get a list of people, who follow someone (done, thanks to get_followers or get_followees) and compare that list to all profiles, that liked a picture (or maybe a picture in the last 6 months or so). This way I want to filter all followers, that haven't liked anything (=ghost followers) and manually block them (I don't want to block someone automated, so that I can really select on which profiles to block).

[nikartz]

So I now found a working solution and want to share it with anyone interested:

I now use InstaPy to export alle the likes on my profile. If set the quickstart.py up with:

try:
    session.login()
    session.set_dont_unfollow_active_users(enabled=True, posts=1000, boundary=50000)
    session.unfollow_users(amount=0, onlyInstapyFollowed = True, onlyInstapyMethod = 'FIFO', sleep_delay=600)

which works for me to output a list of profiles that liked my images. Of course the output needs to be logged somewhere and in the instapy.py you need to add print(active_users) somewhere around line 1977 (at least in my case) in def set_dont_unfollow_active_users so that it looks something like this:

def set_dont_unfollow_active_users(self, enabled=False, posts=4, boundary=500):
        """Prevents unfollow followers who have liked one of
        your latest X posts"""

        # do nothing
        if not enabled:
            return

        # list of users who liked our media
        active_users = get_active_users(self.browser,
                                        self.username,
                                        posts,
                                        boundary,
                                        self.logger)

        print(active_users)

        for user in active_users:
            # include active user to not unfollow list
            self.dont_include.append(user)

After that the output file should contain a list of all profiles that liked your posts.

No head on to outputting all followers. Again the output has to be logged somewhere, but my instaloader (I made a file called export_followers and threw it in the instaloader folder) looks like this:

import instaloader
import time

L = instaloader.Instaloader()

USER = 'your_account'
PASSWORD = 'your_password'
PROFILE = USER

L.login(USER, PASSWORD)

profile = instaloader.Profile.from_username(L.context, PROFILE)

for follower in profile.get_followers():
    print(follower.username)

That should output a file containing all your followers each in a new line.

That file I take to word and automatically replace every newline character with ', ' which means I search for ^p and replace it with ', '. After that there has to be done some cleaning, like adding [' at the beginning and taking a look at the beginning and the end if everything looks like a proper python list.

Those two lists I take into another simple python-script, which compares two lists and outputs everything that doesn't match up into a file. The script looks like this:

#list of followers:
follower = []

#list of likes:
liker = []

#compare
follower.sort()
liker.sort()

matches_literal = [set(follower) & set(liker)]

matches = str(matches_literal).replace('{','').replace('}','')

print('All matches:')
print(matches)
print()
print()
print('Inactive followers:')

ban = [n for n in follower if n not in liker]

print(ban)

print(ban, file=open("/YOUR PATH/inactive-users.txt", "w"))

print()
print()
print('Done')

Where ist says follower = [] and liker = [] you of course need to add your own lists, that have been outputted.

This way there after all of that there is a file containing all inactive users, that follow you. Now I go ahead and manually decide, if I want to ban them in order to get rid of ghost followers.

Maybe this approach is helpful to someone. I know, that it is a bit of work required, but I a beginner at python and couldn't automate the process more.

[Thammus]

Hello nikartz, your goal can easily be achieved using Instaloader only. No need to use other python modules or text editing software. To store inactive followers into a file you can use following approach:

import instaloader

L = instaloader.Instaloader()

USER = 'your_account'
PROFILE = USER

# Your preferred way of logging in:
L.load_session_from_file(USER)

profile = instaloader.Profile.from_username(L.context, PROFILE)

likes = set()
print('Fetching likes of all posts of profile {}.'.format(profile.username))
for post in profile.get_posts():
    print(post)
    likes = likes | set(post.get_likes())

print('Fetching followers of profile {}.'.format(profile.username))
followers = set(profile.get_followers())

ghosts = followers - likes

print('Storing ghosts into file.')
with open('/YOUR PATH/inactive-users.txt', 'w') as f:
    for ghost in ghosts:
        print(ghost.username, file=f)

[kenstowe]

@Thammus is there a way to have instaloader only pull likes from the latest five posts? Like @nikartz is utilizing instapy to do?

I'm trying to compare likes to ghosts on an account with 500 posts. It obviously takes a very long time to scrape and likes that are older than a couple weeks are stale and don't really prove currently active users.

[aandergr]

@Thammus is there a way to have instaloader only pull likes from the latest five posts? Like @nikartz is utilizing instapy to do?

Sure. profile.get_posts() returns an iterator, which can be sliced with islice() from itertools. So instead of

for post in profile.get_posts():
    ...

you can use

from itertools import islice
for post in islice(profile.get_posts(), 5):
    ...

You can also use the post's age as stop condition, where takewhile() comes handy. For example,

from datetime import datetime, timedelta
from itertools import takewhile
NOW = datetime.now()
for post in takewhile(lambda p: NOW - p.date < timedelta(days=7), profile.get_posts()):
    ...

[kenstowe]

Thanks @aandergr for so many options. I'll try them out. :)

[kenstowe]

The islice() worked perfectly for finding recent activity.

I would like to have a list of all likes also as in Thammus' original example. But my profile has 500 posts and when I run .get_posts() it always errors out with the 429 too many requests halfway-ish through. How do I insert longer wait times between requests? Or is there a better way to prevent the 429 errors? I have no other instances of instaloader or anything related to instagram running on this machine or even from the same IP.

[aandergr]

A general note about the notorious 429 - Too Many Requests: Instaloader has a logic to keep track of its requests to Instagram and to obey their rate limits. Since they are nowhere documented, we try them out experimentally. We have a daily cron job running to confirm that Instaloader still stays within the rate limits. Nevertheless, the rate control logic assumes that

• at one time, Instaloader is the only application that consumes requests. I.e. neither the Instagram browser interface, nor a mobile app, nor another Instaloader instance is running in parallel,
• no requests had been consumed when Instaloader starts.

The latter one implies that restarting or reinstantiating Instaloader often within short time is prone to cause a 429. When a request is denied with a 429, Instaloader retries the request as soon as the temporary ban is assumed to be expired.

(copy of my recent comment in https://github.com/instaloader/instaloader/issues/128#issuecomment-396171282)