With a 10:3 ratio in support for migrating the community off of lemmy.world as suggested by @[email protected] in this post, we will run a trial period of migrating the community over to [email protected]

Trial period

We will run a trial period of 2 months on lemm.ee to see how stable the instance is. Since all of you who discussed potential instances only had good things to say about the instance, expect the trial period to just be formality though.

Locking down [email protected]

The FootballAutoMod has been configured to automatically lock down any new posts and tell the posters to repost in the new community. While I could lock the community completely, I think this solution might help inform those of you who maybe don’t check in on the community regularly.

Why lemm.ee

It’s a moderately large EU-based instance with little to no drama as far as I know. The admin team seems to be competent and maintains a neutral stance with few defederations across the fediverse. The instance also got multiple recommendations amongst our users.

  • @[email protected]
    link
    fedilink
    English
    4
    edit-2
    17 hours ago

    With this script, can you get the list of subcribers ?

    I don’t think the Lemmy API exposes the subscriber list of a community, you’ll need access to the instance database.

    What I did was checking every post for the last 365 days for user activity and store every user that have interacted with the community.

    lemmy_session and lemmy_references are the same as for the LiveThreadBot (inside the src folder): https://gitlab.com/UlrikHD/lemmy-match-thread-bot

    get_posts() may be missing from the lemmy_session.py file though

        def get_posts(self, *, community: int | str | LemmyCommunity, sort: str = 'New', page: int = 1) -> dict[str, any]:
            """ Gets the posts of a community.
    
            :param community: The ID of the community to get the posts of, can also be a LemmyCommunity parseable
            string/object.
            :param sort: The sorting method of the posts, by default 'New'.
            :param page: The page number of the posts, by default 1.
            :return: The response JSON of the request as a dictionary.
            """
            if isinstance(community, LemmyCommunity) or isinstance(community, str):
                response: Final[requests.Response] = self.srv.get_posts(community_name=community, sort=sort, page=page,
                                                                        limit=50)
            else:
                response: Final[requests.Response] = self.srv.get_posts(community_id=community, sort=sort, page=page,
                                                                        limit=50)
            if response.status_code != 200:
                raise requests.exceptions.HTTPError(response.text)
            return response.json()
    

    Excuse the ugly code, it was written as a one-off

    import os
    import time
    import datetime
    from json import load, dump
    import requests
    from lemmy_references import LemmyCommunity, LemmyUser
    from lemmy_session import LemmySession
    
    
    session: LemmySession = LemmySession(website='https://lemmy.world/',
                                         username='TestUlrikHD',
                                         password='---',
                                         end_script_signal=None)
    
    posts: list[dict[[str, any]]] = []
    cutoff_date: datetime.datetime = datetime.datetime.now(datetime.timezone.utc) - datetime.timedelta(days=365)
    page_count: int = 1
    loop_break: bool = False
    while True:
        post_response: dict[str, any] = session.get_posts(community=LemmyCommunity('football', 'lemmy.world'),
                                                          page=page_count)
        page_count += 1
        for post in post_response['posts']:
            if datetime.datetime.fromisoformat(post['post']['published']) > cutoff_date:
                posts.append(post)
            else:
                loop_break = True
                break
        if loop_break:
            break
    
    user_dict: dict[str, dict[str, any]] = {}
    for post in posts:
        user_dict[str(LemmyUser(post['creator']['actor_id']))] = {'post': True, 'post_id': post['post']['id']}
        comments = session.get_post_comments(post_id=post['post']['id'])
        for comment in comments['comments']:
            user: str = str(LemmyUser(comment['creator']['actor_id']))
            if user not in user_dict:
                user_dict[user] = {'post': False, 'post_id': comment['post']['id'], 'parent_id': comment['comment']['id']}
    del user_dict[str(LemmyUser('[email protected]'))]
    del user_dict[str(LemmyUser('[email protected]'))]
    with open('user_dict', 'w', encoding='utf-8') as file:
        dump(user_dict, file, ensure_ascii=False, indent=4)
    
    
    def log_reply(usr: str) -> None:
        user_list: list[str] = []
        if os.path.isfile('reply_list.json'):
            with open('reply_list.json', 'r', encoding='utf-8') as file:
                user_list = load(file)
        user_list.append(str(usr))
        with open('reply_list.json', 'w', encoding='utf-8') as file:
            dump(user_list, file, ensure_ascii=False, indent=4)
    
    
    for username, user in user_dict.items():
        time.sleep(1)
        try:
            #if user['post']:
            #    session.reply(content='migration message', post_id=user['post_id'], parent_id=None)
            #else:
            #    session.reply(content='migration message', post_id=user['post_id'], parent_id=user['parent_id'])
            log_reply(usr=LemmyUser(username).str_link())
        except requests.HTTPError as e:
            print(f'Failed to send message to {username} - {e}')
    
    

    and this part creates txt for easy copy pasting for tagging.

    from json import load
    
    with open('reply_list.json', 'r', encoding='utf-8') as file:
        user_list: list[str] = load(file)
    
    loop_count: int = len(', '.join(user_list)) // 9500 + 1
    for i in range(loop_count):
        with open(f'reply_list_{i}.txt', 'w', encoding='utf-8') as file:
            print(len(' '.join(user_list[i * len(user_list) // loop_count:(i + 1) * len(user_list) // loop_count])))
            file.write(', '.join(user_list[i * len(user_list) // loop_count:(i + 1) * len(user_list) // loop_count]))
    
    • Snoopy
      link
      fedilink
      English
      416 hours ago

      Thank a lot, i will share on our Meta community so we can save it and setup an migration bot. :)

      I also shared this idea so maybe Piefed can integrate it into its core with a migration button where you add the new address and subcribers will receive a notification to join the new one.