twoot/twoot.py

#! /usr/bin/env python3
# -*- coding: utf-8 -*-

"""
    Copyright (C) 2019-2022  Jean-Christophe Francois

    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation, either version 3 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <http://www.gnu.org/licenses/>.
"""

import sys
import logging
import argparse
import os
import random
import requests
from bs4 import BeautifulSoup, element
import sqlite3
import datetime
import time
import re
from pathlib import Path
from mastodon import Mastodon, MastodonError, MastodonAPIError, MastodonIllegalArgumentError
import subprocess
import shutil

NITTER_URLS = [
    'https://nitter.42l.fr',
    'https://nitter.pussthecat.org',
    'https://nitter.fdn.fr',
    'https://nitter.eu',
    'https://nitter.namazso.eu',
    'https://n.actionsack.com',
    'https://nitter.moomoo.me',
    'https://n.ramle.be',
    ]

# Update from https://www.whatismybrowser.com/guides/the-latest-user-agent/
USER_AGENTS = [
    'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/104.0.0.0 Safari/537.36',
    'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:103.0) Gecko/20100101 Firefox/103.0',
    'Mozilla/5.0 (Macintosh; Intel Mac OS X 12_5_1) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/15.6 Safari/605.1.15',
    'Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/104.0.0.0 Safari/537.36 Vivaldi/5.4.2753.37',
    ]


def process_media_body(tt_iter):
    """
    Receives an iterator over all the elements contained in the tweet-text container.
    Processes them to make them suitable for posting on Mastodon
    :param tt_iter: iterator over the HTML elements in the text of the tweet
    :return:        cleaned up text of the tweet
    """
    tweet_text = ''
    # Iterate elements
    for tag in tt_iter:
        # If element is plain text, copy it verbatim
        if isinstance(tag, element.NavigableString):
            tweet_text += tag.string

        # If it is an 'a' html tag
        elif tag.name == 'a':
            tag_text = tag.get_text()
            if tag_text.startswith('@'):
                # Only keep user name
                tweet_text += tag_text
            elif tag_text.startswith('#'):
                # Only keep hashtag text
                tweet_text += tag_text
            else:
                # This is a real link, keep url
                tweet_text += tag.get('href')
        else:
            logging.warning("No handler for tag in twitter text: " + tag.prettify())

    return tweet_text


def process_card(nitter_url, card_container):
    """
    Extract image from card in case mastodon does not do it
    :param card_container: soup of 'a' tag containing card markup
    :return: list with url of image
    """
    list = []

    img = card_container.div.div.img
    if img is not None:
        image_url = nitter_url + img.get('src')
        list.append(image_url)
        logging.debug('Extracted image from card')

    return list


def process_attachments(nitter_url, attachments_container, get_vids, twit_account, status_id, author_account):
    """
    Extract images or video from attachments. Videos are downloaded on the file system.
    :param nitter_url: url of nitter mirror
    :param attachments_container: soup of 'div' tag containing attachments markup
    :param get_vids: whether to download videos or not
    :param twit_account: name of twitter account
    :param status_id: id of tweet being processed
    :param author_account: author of tweet with video attachment
    :return: list with url of images
    """
    # Collect url of images
    pics = []
    images = attachments_container.find_all('a', class_='still-image')
    for image in images:
        pics.append(nitter_url + image.get('href'))

    logging.debug('collected ' + str(len(pics)) + ' images from attachments')

    # Download nitter video (converted animated GIF)
    gif_class = attachments_container.find('video', class_='gif')
    if gif_class is not None:
        gif_video_file = nitter_url + gif_class.source.get('src')

        video_path = os.path.join('output', twit_account, status_id, author_account, status_id)
        os.makedirs(video_path, exist_ok=True)

        # Open directory for writing file
        orig_dir = os.getcwd()
        os.chdir(video_path)
        with requests.get(gif_video_file, stream=True) as r:
            r.raise_for_status()
            # Download chunks and write them to file
            with open('gif_video.mp4', 'wb') as f:
                for chunk in r.iter_content(chunk_size=16*1024):
                    f.write(chunk)

        logging.debug('downloaded video of GIF animation from attachments')

        # Close directory
        os.chdir(orig_dir)

    # Download twitter video
    vid_in_tweet = False
    vid_class = attachments_container.find('div', class_='video-container')
    if vid_class is not None:
        video_file = os.path.join('https://twitter.com', author_account, 'status', status_id)
        if get_vids:
            # Download video from twitter and store in filesystem. Running as subprocess to avoid
            # requirement to install ffmpeg and ffmpeg-python for those that do not want to post videos
            try:
                # Set output location to ./output/twit_account/status_id
                dl_feedback = subprocess.run(
                    ["./twitterdl.py", video_file, "-ooutput/" + twit_account + "/" + status_id, "-w 500"],
                    capture_output=True,
                    timeout=300
                )
                if dl_feedback.returncode != 0:
                    logging.warning('Video in tweet ' + status_id + ' from ' + twit_account + ' failed to download')
                    vid_in_tweet = True
                else:
                    logging.debug('downloaded twitter video from attachments')

            except OSError:
                logging.fatal("Could not execute twitterdl.py (is it there? Is it set as executable?)")
                sys.exit(-1)
        else:
            vid_in_tweet = True

    return pics, vid_in_tweet


def contains_class(body_classes, some_class):
    """
    :param body_classes: list of classes to search
    :param some_class: class that we are interested in
    :return: True if found, false otherwise
    """
    found = False
    for body_class in body_classes:
        if body_class == some_class:
            found = True

    return found

def is_time_valid(timestamp, max_age, min_delay):
    ret = True
    # Check that the tweet is not too young (might be deleted) or too old
    age_in_hours = (time.time() - float(timestamp)) / 3600.0
    min_delay_in_hours = min_delay / 60.0
    max_age_in_hours = max_age * 24.0

    if age_in_hours < min_delay_in_hours or age_in_hours > max_age_in_hours:
        ret = False

    return ret


def login(instance, account, password):
    # Create Mastodon application if it does not exist yet
    if not os.path.isfile(instance + '.secret'):
        try:
            Mastodon.create_app(
                'twoot',
                api_base_url='https://' + instance,
                to_file=instance + '.secret'
            )

        except MastodonError as me:
            logging.fatal('failed to create app on ' + instance)
            logging.fatal(me)
            sys.exit(-1)

    # Log in to Mastodon instance
    try:
        mastodon = Mastodon(
            client_id=instance + '.secret',
            api_base_url='https://' + instance
        )

        mastodon.log_in(
            username=account,
            password=password,
            to_file=account + ".secret"
        )
        logging.info('Logging in to ' + instance)

    except MastodonError as me:
        logging.fatal('ERROR: Login to ' + instance + ' Failed')
        logging.fatal(me)
        sys.exit(-1)

    # Check ratelimit status
    logging.info('Ratelimit allowed requests: ' + str(mastodon.ratelimit_limit))
    logging.info('Ratelimit remaining requests: ' + str(mastodon.ratelimit_remaining))
    logging.info('Ratelimit reset time: ' + time.asctime(time.localtime(mastodon.ratelimit_reset)))
    logging.info('Ratelimit last call: ' + time.asctime(time.localtime(mastodon.ratelimit_lastcall)))

    return mastodon


def main(argv):
    # Start stopwatch
    start_time = time.time()

    # Build parser for command line arguments
    parser = argparse.ArgumentParser(description='toot tweets.')
    parser.add_argument('-t', metavar='<twitter account>', action='store', required=True)
    parser.add_argument('-i', metavar='<mastodon instance>', action='store', required=True)
    parser.add_argument('-m', metavar='<mastodon account>', action='store', required=True)
    parser.add_argument('-p', metavar='<mastodon password>', action='store', required=True)
    parser.add_argument('-r', action='store_true', help='Also post replies to other tweets')
    parser.add_argument('-v', action='store_true', help='Ingest twitter videos and upload to Mastodon instance')
    parser.add_argument('-a', metavar='<max age (in days)>', action='store', type=float, default=1)
    parser.add_argument('-d', metavar='<min delay (in mins)>', action='store', type=float, default=0)
    parser.add_argument('-c', metavar='<max # of toots to post>', action='store', type=int, default=0)

    # Parse command line
    args = vars(parser.parse_args())

    twit_account = args['t']
    mast_instance = args['i']
    mast_account = args['m']
    mast_password = args['p']
    tweets_and_replies = args['r']
    get_vids = args['v']
    max_age = float(args['a'])
    min_delay = float(args['d'])
    cap = int(args['c'])

    # Remove previous log file
    #try:
    #    os.remove(twit_account + '.log')
    #except FileNotFoundError:
    #    pass

    # Setup logging to file
    logging.basicConfig(
        filename=twit_account + '.log',
        level=logging.INFO,
        format='%(asctime)s %(levelname)-8s %(message)s',
        datefmt='%Y-%m-%d %H:%M:%S',
    )

    logging.info('Running with the following parameters:')
    logging.info('    -t ' + twit_account)
    logging.info('    -i ' + mast_instance)
    logging.info('    -m ' + mast_account)
    logging.info('    -r ' + str(tweets_and_replies))
    logging.info('    -v ' + str(get_vids))
    logging.info('    -a ' + str(max_age))
    logging.info('    -d ' + str(min_delay))
    logging.info('    -c ' + str(cap))

    # Try to open database. If it does not exist, create it
    sql = sqlite3.connect('twoot.db')
    db = sql.cursor()
    db.execute('''CREATE TABLE IF NOT EXISTS toots (twitter_account TEXT, mastodon_instance TEXT,
               mastodon_account TEXT, tweet_id TEXT, toot_id TEXT)''')
    db.execute('''CREATE INDEX IF NOT EXISTS main_index ON toots (twitter_account,
               mastodon_instance, mastodon_account, tweet_id)''')

    # Select random nitter instance to fetch updates from
    nitter_url = NITTER_URLS[random.randint(0, len(NITTER_URLS) - 1)]

    # **********************************************************
    # Load twitter page of user. Process all tweets and generate
    # list of dictionaries ready to be posted on Mastodon
    # **********************************************************
    # To store content of all tweets from this user
    tweets = []

    # Initiate session
    session = requests.Session()

    # Get a copy of the default headers that requests would use
    headers = requests.utils.default_headers()

    # Update default headers with randomly selected user agent
    headers.update(
        {
            'User-Agent': USER_AGENTS[random.randint(0, len(USER_AGENTS)-1)],
            'Cookie': 'replaceTwitter=; replaceYouTube=; hlsPlayback=on; proxyVideos=',
        }
    )

    url = nitter_url + '/' + twit_account
    # Use different page if we need to handle replies
    if tweets_and_replies:
        url += '/with_replies'

    # Download twitter page of user.
    try:
        twit_account_page = session.get(url, headers=headers)
    except requests.exceptions.ConnectionError:
        logging.fatal('Host did not respond when trying to download ' + url)
        exit(-1)

    # Verify that download worked
    if twit_account_page.status_code != 200:
        logging.fatal('The Nitter page did not download correctly from ' + url + ' (' + str(twit_account_page.status_code) + '). Aborting')
        exit(-1)

    logging.info('Nitter page downloaded successfully from ' + url)

    # DEBUG: Save page to file
    # of = open(twit_account + '.html', 'w')
    # of.write(twit_account_page.text)
    # of.close()

    # Make soup
    soup = BeautifulSoup(twit_account_page.text, 'html.parser')

    # Replace twit_account with version with correct capitalization
    ta = soup.find('meta', property='og:title').get('content')
    ta_match = re.search(r'\(@(.+)\)', ta)
    if ta_match is not None:
        twit_account = ta_match.group(1)

    # Extract twitter timeline
    timeline = soup.find_all('div', class_='timeline-item')

    logging.info('Processing ' + str(len(timeline)) + ' tweets found in timeline')

    # **********************************************************
    # Process each tweets and generate dictionary
    # with data ready to be posted on Mastodon
    # **********************************************************
    out_date_cnt = 0
    in_db_cnt = 0
    for status in timeline:
        # Extract tweet ID and status ID
        tweet_id = status.find('a', class_='tweet-link').get('href').strip('#m')
        status_id = tweet_id.split('/')[3]

        logging.debug('processing tweet %s', tweet_id)

        # Extract time stamp
        time_string = status.find('span', class_='tweet-date').a.get('title')
        try:
            timestamp = datetime.datetime.strptime(time_string, '%d/%m/%Y, %H:%M:%S').timestamp()
        except:
            # Dec 21, 2021 · 12:00 PM UTC
            timestamp = datetime.datetime.strptime(time_string, '%b %d, %Y · %I:%M %p %Z').timestamp()

        # Check if time is within acceptable range
        if not is_time_valid(timestamp, max_age, min_delay):
            out_date_cnt += 1
            logging.debug("Tweet outside valid time range, skipping")
            continue

        # Check in database if tweet has already been posted
        db.execute("SELECT * FROM toots WHERE twitter_account=? AND mastodon_instance=? AND mastodon_account=? AND tweet_id=?",
                   (twit_account, mast_instance, mast_account, tweet_id))
        tweet_in_db = db.fetchone()

        if tweet_in_db is not None:
            in_db_cnt += 1
            logging.debug("Tweet %s already in database", tweet_id)
            # Skip to next tweet
            continue
        else:
            logging.debug('Tweet %s not found in database', tweet_id)

        # extract author
        author = status.find('a', class_='fullname').get('title')

        # Extract user name
        author_account = status.find('a', class_='username').get('title').lstrip('@')

        # Extract URL of full status page (for video download)
        full_status_url = 'https://twitter.com' + tweet_id

        # Initialize containers
        tweet_text = ''
        photos = []

        # Add prefix if the tweet is a reply-to
        # Only consider item of class 'replying-to' that is a direct child
        # of class 'tweet-body' in status. Others can be in a quoted tweet.
        replying_to_class = status.select("div.tweet-body > div.replying-to")
        if len(replying_to_class) != 0:
            tweet_text += 'Replying to ' + replying_to_class[0].a.get_text() + '\n\n'

        # Check it the tweet is a retweet from somebody else
        if author_account.lower() != twit_account.lower():
            tweet_text = 'RT from ' + author + ' (@' + author_account + ')\n\n'

        # extract iterator over tweet text contents
        tt_iter = status.find('div', class_='tweet-content media-body').children

        # Process text of tweet
        tweet_text += process_media_body(tt_iter)

        # Process quote: append link to tweet_text
        quote_div = status.find('a', class_='quote-link')
        if quote_div is not None:
            tweet_text += '\n\nhttps://twitter.com' + quote_div.get('href').strip('#m')

        # Process card : extract image if necessary
        card_class = status.find('a', class_='card-container')
        if card_class is not None:
            photos.extend(process_card(nitter_url, card_class))

        # Process attachment: capture image or .mp4 url or download twitter video
        attachments_class = status.find('div', class_='attachments')
        if attachments_class is not None:
            pics, vid_in_tweet = process_attachments(nitter_url, attachments_class, get_vids, twit_account, status_id, author_account)
            photos.extend(pics)
            if vid_in_tweet:
                tweet_text += '\n\n[Video embedded in original tweet]'

        # Add footer with link to original tweet
        tweet_text += '\n\nOriginal tweet : ' + full_status_url

        # If no media was specifically added in the tweet, try to get the first picture
        # with "twitter:image" meta tag in first linked page in tweet text
        if not photos:
            m = re.search(r"http[^ \n\xa0]*", tweet_text)
            if m is not None:
                link_url = m.group(0)
                if link_url.endswith(".html"):  # Only process a web page
                    try:
                        r = requests.get(link_url, timeout=10)
                        if r.status_code == 200:
                            # Matches the first instance of either twitter:image or twitter:image:src meta tag
                            match = re.search(r'<meta name="twitter:image(?:|:src)" content="(.+?)".*?>', r.text)
                            if match is not None:
                                url = match.group(1).replace('&amp;', '&')  # Remove HTML-safe encoding from URL if any
                                photos.append(url)
                    # Give up if anything goes wrong
                    except (requests.exceptions.ConnectionError,
                            requests.exceptions.Timeout,
                            requests.exceptions.ContentDecodingError,
                            requests.exceptions.TooManyRedirects,
                            requests.exceptions.MissingSchema):
                        pass
                    else:
                        logging.debug("downloaded twitter:image from linked page")

        # Check if video was downloaded
        video_file = None

        video_path = Path('./output') / twit_account / status_id
        if video_path.exists():
            # Take the first subdirectory of video path (named after original poster of video)
            video_path = [p for p in video_path.iterdir() if p.is_dir()][0]
            # Take again the first subdirectory of video path (named after status id of original post where video is attached)
            video_path = [p for p in video_path.iterdir() if p.is_dir()][0]
            # list video files
            video_file_list = list(video_path.glob('*.mp4'))
            if len(video_file_list) != 0:
                # Extract posix path of first video file in list
                video_file = video_file_list[0].absolute().as_posix()

        # Add dictionary with content of tweet to list
        tweet = {
            "author": author,
            "author_account": author_account,
            "timestamp": timestamp,
            "tweet_id": tweet_id,
            "tweet_text": tweet_text,
            "video": video_file,
            "photos": photos,
        }
        tweets.append(tweet)

        logging.debug('Tweet %s added to list of toots to upload', tweet_id)

    # Log summary stats
    logging.info(str(out_date_cnt) + ' tweets outside of valid time range')
    logging.info(str(in_db_cnt) + ' tweets already in database')

    # DEBUG: Print extracted tweets
    # for t in tweets:
    # print(t)

    # Login to account on maston instance
    mastodon = None
    if len(tweets) != 0:
        mastodon = login(mast_instance, mast_account, mast_password)

    # **********************************************************
    # Iterate tweets in list.
    # post each on Mastodon and record it in database
    # **********************************************************

    posted_cnt = 0
    for tweet in reversed(tweets):
        # Check if we have reached the cap on the number of toots to post
        if cap != 0 and posted_cnt >= cap:
            logging.info('%d toots not posted due to configured cap', len(tweets) - cap)
            break

        logging.debug('Uploading Tweet %s', tweet["tweet_id"])

        media_ids = []

        # Upload video if there is one
        if tweet['video'] is not None:
            try:
                logging.debug("Uploading video to Mastodon")
                media_posted = mastodon.media_post(tweet['video'])
                media_ids.append(media_posted['id'])
            except (MastodonAPIError, MastodonIllegalArgumentError, TypeError):  # Media cannot be uploaded (invalid format, dead link, etc.)
                logging.debug("Uploading video failed")
                pass

        else:  # Only upload pic if no video was uploaded
            # Upload photos
            for photo in tweet['photos']:
                media = False
                # Download picture
                try:
                    logging.debug('downloading picture')
                    media = requests.get(photo)
                except:  # Picture cannot be downloaded for any reason
                    pass

                # Upload picture to Mastodon instance
                if media:
                    try:
                        logging.debug('uploading picture to Mastodon')
                        media_posted = mastodon.media_post(media.content, mime_type=media.headers['content-type'])
                        media_ids.append(media_posted['id'])
                    except (MastodonAPIError, MastodonIllegalArgumentError, TypeError):  # Media cannot be uploaded (invalid format, dead link, etc.)
                        pass

        # Post toot
        try:
            mastodon = Mastodon(
                access_token=mast_account + '.secret',
                api_base_url='https://' + mast_instance
            )

            if len(media_ids) == 0:
                toot = mastodon.status_post(tweet['tweet_text'], visibility='public')
            else:
                toot = mastodon.status_post(tweet['tweet_text'], media_ids=media_ids, visibility='public')

        except MastodonError as me:
            logging.error('posting ' + tweet['tweet_text'] + ' to ' + mast_instance + ' Failed')
            logging.error(me)

        else:
            posted_cnt += 1
            logging.debug('Tweet %s posted on %s', tweet['tweet_id'], mast_account)

        # Insert toot id into database
        if 'id' in toot:
            db.execute("INSERT INTO toots VALUES ( ? , ? , ? , ? , ? )",
                       (twit_account, mast_instance, mast_account, tweet['tweet_id'], toot['id']))
            sql.commit()

    logging.info(str(posted_cnt) + ' Tweets posted to Mastodon')

    # Cleanup downloaded video files
    try:
        shutil.rmtree('./output/' + twit_account)
    except FileNotFoundError:  # The directory does not exist
        pass

    logging.info('Run time : %2.1f seconds' % (time.time() - start_time))
    logging.info('_____________________________________________________________________________________')


if __name__ == "__main__":
    main(sys.argv)