ig-pixelfed-mirror/igmirror.py

#!/usr/bin/python3
import threading
import requests
import hashlib
import string
import random
import time
import json
import os
import re

CONFIG = {}

def add_igaccount(acc_id):
    accfile = './db/accounts/{}'.format(acc_id)

    # user_create script must exist before running the API server
    if not os.path.exists('./scripts/user_create'):
        print('E| You may need to initialize the server environment first')
        return 1

    if not os.path.exists(accfile):
        # get all profile data from instagram acc
        data = getig_user_data(acc_id)

        # this tells us the user probably don't exist (or also might be a network error?)
        if len(data.keys()) == 0:
            print('E| User "{}" does not exist on Instagram'.format(acc_id))
            return 2

        # we don't and can't mirror private profiles
        #  (well, we can if we follow them and follow back, but we just don't need that)
        if data['graphql']['user']['is_private']:
            print('E| User "{}" is a private account. We just won\'t mirror that!'.format(acc_id))
            return 3

        # get account display name to create it
        name = data['graphql']['user']['full_name']
        name = re.sub(r'[^a-zA-Z0-9_\s]', '', name)
        account = {
            'name': name,
            'username': acc_id,
            'password': random_string()
        }
        # we are completely sure the parameters can't contain unwanted characters
        #  a shell exploit is not possible here :)
        os.system('./scripts/user_create \'{}\' \'{}\' \'{}\''.format(\
            account['name'], account['username'], account['password']))

        # save the account login information for updates and mirroring
        db_set('accounts', acc_id, account)
        db_set('posts', acc_id, [])

        # set Pixelfed account data for the username
        pixelfed_setpic(acc_id, data['graphql']['user']['profile_pic_url'])
        pixelfed_setinfo(acc_id, data['graphql']['user']['biography'],\
            data['graphql']['user']['external_url'])
        update_igaccount_async(acc_id)
    else:
        print('W| User "{}" already exists in local database'.format(acc_id))
    return 0

def update_igaccount_async(acc_id):
    threading.Thread(target=update_igaccount, args=(acc_id,)).start()

def update_igaccount(acc_id):
    # if account does not exist, we stop the mirroring process
    accfile = './db/accounts/{}'.format(acc_id)
    if not os.path.exists(accfile):
        print('E| User "'+acc_id+'" has not been created yet, maybe you wanted to call /<username>/add ?')
        return 1

    data = getig_user_data(acc_id)
    pixelfed_setpic(acc_id, data['graphql']['user']['profile_pic_url'])
    pixelfed_setinfo(acc_id, data['graphql']['user']['biography'],\
            data['graphql']['user']['external_url'])
    pixelfed_dlposts(acc_id, data['graphql']['user'])

def update_allaccounts_async():
    threading.Thread(target=update_allaccounts).start()

def update_allaccounts():
    # update all accounts with a timeout of 20 seconds
    for acc_id in os.listdir('./db/accounts'):
        print('I| mirroring account "{}"...'.format(acc_id))
        update_igaccount(acc_id)
        print('I| timeout 20 seconds')
        time.sleep(20)
        print()
    print('I| done updating all accounts')


def pixelfed_islogged(acc_id, accdata=None):
    if accdata is None:
        accdata = db_get('accounts', acc_id)
    return 'cookie' in accdata

def pixelfed_login(acc_id, force=False):
    # check account is already logged in if not "force"
    accdata = db_get('accounts', acc_id)
    if not force and pixelfed_islogged(acc_id, accdata):
        return

    # obtain one time tokens for the pixelfed instance
    _cookies, _token = pixelfed_token_url()

    # do the login post and retrieve the raw cookies, the rest of API calls will have this cookies
    r = requests.post( 'https://'+config()['instance']+'/login' ,\
        data={
            '_token': _token, 'email': 'pixelfed.'+acc_id+'@localhost',
            'password': accdata['password'], 'remember': 'on'
        },
        cookies=_cookies
    )

    # add the raw cookies to the account data for later calls
    accdata['cookie'] = dict(r.cookies)
    db_set('accounts', acc_id, accdata)

def pixelfed_token_url(url='', _cookies=None):
    r = requests.get( 'https://'+config()['instance']+url, cookies=_cookies )
    _token = re.search(r'name="_token".+value="([^"]+)"', r.text).group(1)
    return r.cookies, _token

def pixelfed_dlposts(acc_id, data):
    ts = []
    for edge in data['edge_owner_to_timeline_media']['edges']:
        ts.append(edge['node']['taken_at_timestamp'])
    for edge in data['edge_felix_video_timeline']['edges']:
        ts.append(edge['node']['taken_at_timestamp'])
    ts = sorted(ts)
    items = []
    for t in ts:
        brkit = False
        for edge in data['edge_owner_to_timeline_media']['edges']:
            if edge['node']['taken_at_timestamp'] == t:
                items.append(edge['node'])
                brkit = True
                break
        if brkit:
            continue
        for edge in data['edge_felix_video_timeline']['edges']:
            if edge['node']['taken_at_timestamp'] == t:
                items.append(edge['node'])
                break

    # mirror posts from the account (only the last N, without loading more),
    #   but only the ones that has not already been imported
    accposts = db_get('posts', acc_id)
    accdata = db_get('accounts', acc_id)
    for item in items:
        if item['is_video']:
            continue

        if item['shortcode'] in accposts:
            print('I| skipping IG post {}:{}. Already added'.format(acc_id, item['shortcode']))
            continue

        print('I| processing IG post {}:{}'.format(acc_id, item['shortcode']))
        ig_url = 'https://www.instagram.com/p/{}/'.format(item['shortcode'])
        title = item['title'] if 'title' in item else None

        caption = item['edge_media_to_caption']['edges'][0]['node']['text'] \
            if len(item['edge_media_to_caption']['edges']) > 0 else ''
        altcaption = item['accessibility_caption'] if 'accessibility_caption' in item else None
        full_caption    = caption
        full_altcaption = altcaption

        # add support for posts with multiple images
        #  get the data from the post URL. (we need all images, as IG can have not only 1 image in the post)
        postdata = json.loads(instagram_get('/p/{}/?__a=1'.format(item['shortcode'])))
        multiple = False
        multmedia = None
        if len(postdata.keys()) > 0 and 'edge_sidecar_to_children' in postdata['graphql']['shortcode_media']:
            multiple = True
            multmedia = postdata['graphql']['shortcode_media']['edge_sidecar_to_children']['edges']

        # for now, we only support images (not videos :( )
        if not item['is_video']:
            jsdata_items = []
            _token = None
            failed = False

            # we add support to multiple media here
            media2iterate = [a['node']['display_url'] for a in multmedia] if multiple else [item['display_url']]
            for media in media2iterate:
                _token, jsdata = pixelfed_postimage(acc_id, media, accdata)
                if not jsdata:
                    print('E| Could not upload media for {}:{}'.format(acc_id, item['shortcode']))
                    failed = True
                    break
                jsdata_items.append(jsdata)
            if failed:
                continue

            # add the accesibility captions
            caption = caption[0:136]+'...' if len(caption) > 140 else caption
            i = 0
            for jsdata in jsdata_items:
                jsdata['description'] = ig_url
                jsdata['cw'] = False
                jsdata['alt'] = ''
                if multiple:
                    mnode = multmedia[i]['node']
                    if 'accessibility_caption' in mnode and not mnode['accessibility_caption'] is None:
                        jsdata['alt'] = mnode['accessibility_caption'][0:136]+'...' \
                            if len(mnode['accessibility_caption']) > 140 else mnode['accessibility_caption']
                else:
                    jsdata['alt'] = altcaption[0:136]+'...' if len(altcaption) > 140 else altcaption
                i += 1

            # publish the post using Pixelfed API
            r = requests.post('https://'+config()['instance']+'/api/compose/v0/publish',\
                json={"media": jsdata_items, "caption": caption, "visibility": "public", "cw": False,\
                    "comments_disabled": False, "place": False, "tagged": [],"optimize_media": True},\
                cookies=accdata['cookie'],\
                headers={
                    'Content-Type': 'application/json',
                    'X-CSRF-TOKEN': _token,
                    'X-Requested-With': 'XMLHttpRequest',
                    'X-XSRF-TOKEN': accdata['cookie']['XSRF-TOKEN']
                }
            )
            accposts.append(item['shortcode'])
            print('I| uploaded media for {}:{} : {}'.format(acc_id, item['shortcode'], r.status_code))

    print('I| done uploading media for {}'.format(acc_id))
    db_set('posts', acc_id, accposts)

# upload media and return data
def pixelfed_postimage(acc_id, image_url, accdata=None):
    if accdata is None:
        accdata = db_get('accounts', acc_id)
    cachef = pixelfed_cacheimg(image_url)
    _, _token = pixelfed_token_url('', accdata['cookie'])
    r = requests.post( 'https://'+config()['instance']+'/api/compose/v0/media/upload',\
        files={'file': open(cachef, 'rb')}, cookies=accdata['cookie'],\
        headers={
            'X-CSRF-TOKEN': _token,
            'X-Requested-With': 'XMLHttpRequest',
            'X-XSRF-TOKEN': accdata['cookie']['XSRF-TOKEN']
        }
    )
    if r.status_code == 200:
        return _token, json.loads(r.text)
    return None, False


# get the image by URL but cache it forever, as if the profile changes the pic
#  the url will be different, and therefore, the sum will also be different
def pixelfed_cacheimg(image_url):
    cachef = './cache/{}.jpg'.format(md5sum(image_url))
    if not os.path.exists(cachef):
        r = requests.get(image_url)
        w = open(cachef, 'wb')
        w.write(r.content)
        w.close()
    return cachef

def pixelfed_setpic(acc_id, pic_url, count=0):
    count += 1
    pixelfed_login(acc_id)

    cachef = pixelfed_cacheimg(pic_url)
    accdata = db_get('accounts', acc_id)
    print('I| setting avatar for "{}" '.format(acc_id), end="")
    _, _token = pixelfed_token_url('/settings/home', accdata['cookie'])
    r = requests.post( 'https://'+config()['instance']+'/settings/avatar',\
        data={'_token': _token}, cookies=accdata['cookie'], files={'avatar': open(cachef, 'rb')}
    )

    # try to login if the upload failed
    if r.status_code == 419 and count < 3:
        print('err (login required)')
        pixelfed_login(acc_id, True)
        return pixelfed_setpic(acc_id, pic_url, count)

    print('ok')

    return True

def pixelfed_setinfo(acc_id, bio, website, count=0):
    accdata = db_get('accounts', acc_id)

    name = accdata['name'] + ' [Mirror]'
    if count == 0:
        bio = 'Mirrored from Instagram: instagram.com/{} | {}'.format(acc_id, bio)

    count += 1
    pixelfed_login(acc_id)

    print('I| setting account-info for "{}" '.format(acc_id), end="")
    _, _token = pixelfed_token_url('/settings/home', accdata['cookie'])
    r = requests.post( 'https://'+config()['instance']+'/settings/home',\
        data={
            '_token': _token, 'name': name,
            'website': website, 'bio': bio, 'language': 'en'
        },
        cookies=accdata['cookie']
    )

    # try to login if the upload failed
    if r.status_code == 419 and count < 3:
        print('err (login required)')
        pixelfed_login(acc_id, True)
        return pixelfed_setinfo(acc_id, bio, website, count)

    print('ok')

    return True


def random_string(count=32):
    return ''.join(random.choices(string.ascii_uppercase + string.ascii_lowercase + string.digits, k=count))

def md5sum(_str):
    return hashlib.md5(_str.encode()).hexdigest()

# get all profile data from user:
#  - display name
#  - bio description
#  - shared posts (images/videos)
#  - much more info...
def getig_user_data(acc_id):
    return json.loads(
        instagram_get('/{}/?__a=1'.format(acc_id))
    )

# runs a basic GET request emulating Tor Browser
def instagram_get(url, CACHE_SECS=600):
    headers = get_random_headers()
    default_headers = {
        'Accept':		'*/*',
        'Accept-Language':	'en-US,en;q=0.5',
        'Accept-Encoding':	'gzip, deflate, br',
        'Connection':		'keep-alive',
        'Cache-Control':	'no-cache',
        'Pragma':		'no-cache',
        'Host':			'www.instagram.com',
        'Referer':		'https://www.instagram.com/',
        'User-Agent':		'Mozilla/5.0 (Windows NT 10.0; rv:78.0) Gecko/20100101 Firefox/78.0'
    }
    for key in default_headers.keys():
        if not key in headers:
            headers[key] = default_headers[key]
    url = 'https://www.instagram.com{}'.format(url)
    cachef = './cache/'+md5sum(url)
    now = str(time.time())
    now = int(now[:now.index('.')])
    if os.path.exists(cachef):
        cache = readf(cachef).splitlines()
        ctime = int(cache[0])
        if now < ctime:
            return '\n'.join(cache[1:])

    r = requests.get(url, headers=headers)
    resp = r.text
    w = open(cachef, 'w')
    w.write(str(now+CACHE_SECS) + '\n')
    w.write(resp)
    return resp

def get_random_headers():
    a = os.listdir('./headers')
    rin = 0
    if len(a)-1 > 0:
        rin = random.randint(0, len(a)-1)
    lines = readf('./headers/{}'.format(a[rin])).splitlines()
    headers = {}
    for line in lines:
        reg = re.search('(^[^:]+):(.*)', line)
        headers[reg.group(1).strip()] = reg.group(2).strip()
    return headers

def db_set(table, acc_id, accdata):
    w = open('./db/{}/{}'.format(table, acc_id), 'w')
    w.write(json.dumps(accdata))
    w.close()

def db_get(table, acc_id):
    return json.loads(readf('./db/{}/{}'.format(table, acc_id)))

def config():
    global CONFIG
    if len(CONFIG.keys()) == 0:
        CONFIG = json.loads(readf('./config.json'))
    return CONFIG

def readf(f):
    r = open(f,'r')
    c = r.read().strip()
    r.close()
    return c

if __name__ == '__main__':
    main()