lynnesbian/mstdn-ebooks

NO LONGER BEING UPDATED, NOT GUARANTEED TO WORK AT ALL

mstdn-ebooks's Issues

Code cleanup

functions.py in particular is disgraceful

The @ and instance name are not removed when someone's handle appears in a toot

I'm not quite sure why it's not doing this as I was sure this was a feature that is implemented. Everything is set up according to the guide, and I re-set up the bot again just now to make sure. Is there a glaringly obvious configuration option that I could be missing?

Running this on ubuntu server 18.04.3 LTS if that matters.

Reorganise into a package

While this project is small, I think it would be a good idea to put the current code into a proper installable package.

I can submit a PR if you like.

Posts from Mastodon servers with AUTHORIZED_MODE enabled can't be accessed

We need an oauth token but I don't know how to get one

Keep "best of" lists

As described here: https://fedi.lynnesbian.space/web/statuses/102094864542977892

The bots could check how many faves and boosts a given post gets over X hours/days, then save the "score". They could use this to keep a "best of" list.

Error when decoding Emoji

I got an error when running gen.py with --simulate. Apparently the script fails when trying to decode the "zany-face"-Emoji.

╭─lerk@comm ~/mstdn-ebooks  ‹master*› 
╰─$ python3 ./gen.py --simulate
Traceback (most recent call last):
  File "./gen.py", line 39, in <module>
    print(toot['toot'])
UnicodeEncodeError: 'ascii' codec can't encode character '\U0001f92a' in position 100: ordinal not in range(128)

Can I ignore invalid characters somehow?

Don't create infinite loop threads

There are multiple situations that will cause an ebooks bot to create an infinite reply loop:

• An ebooks bot mentions another ebooks bot
• An ebooks bot mentions a bot like OCRbot or that alpine shell bot
• Alice the Attacker creates a bot that constantly tags random ebooks bots

We need to guard against this. One possible option is to ignore a thread after already replying in it five times.

Fix compatibility with Pleroma

Turns out that snowflake IDs broke mstdn-ebooks a while ago and nobody noticed until now

extract_toot deletes the content of <p> blocks

Hello,

I've tried using this and generation wouldn't work, so I dug into the code and found out that all toots' content in the db were empty. Looking further, I've found the reason : 8274409 line 73

This commit changed the handling of p tags from "Append a line break and unwrap" to "replace with a linebreak". I'm assuming that's an unfortunate copy/paste with the above change about br tags, but either way the result is that all content inside p tags is sent into oblivion. I don't know if it's common or something about my instance but all of my toots' content are wrapped in a p tag, thus the database emptiness. Reverting that change made the code work again for me.

Add 1-button Heroku/Now.sh deployment

Hiiiiii!

Was reading through the ebooks bot guide and it sort of dawned on me that it'd be really simple to set up a button on the repo to deploy it to Now.sh or Heroku (probably the former because it supports Docker).

Docs:

• Now.sh
• Heroku

I've only really ever had luck with the latter but I haven't tried Docker on Now.sh yet, might work better.

Happy to PR if it helps! 💚

-Andi

Add option to download followers-only posts

This isn't possible with the current AP outbox method, and will require re-implementing the legacy method. It will only work with Mastodon instances.

Convey rate limit error better

Right now, rate limiting produces the generic "fucky wucky" error. We should catch the HTTP error (if any) and print something more helpful in case of HTTP 429.

Pleroma toot collection fails

here is the error output, running the latest version from git, i tried twice so it did get through some tootsmews but eventually there came a problem toot

WebFingering...
catgirl.science is a pleroma instance -- falling back to legacy toot collection method
Downloading and parsing tootsEncountered an error! Saving toots to database and exiting.

WebFingering...
catgirl.science is a pleroma instance -- falling back to legacy toot collection method
Downloading and parsing tootsEncountered an error! Saving toots to database and exiting.```

Allow people to request toot pinning

If someone the bot is following types "@.bot pin" to a post, the bot will pin the post. "@.bot unpin" would also be possible.

Add optional boost/fave "high scores"

Keep track of the top 10 people to fave/boost the bot's posts, because why not

Mastodon users with one post break the script.

I have a user with one post which breaks the script with the following error:

Using config.json as configuration file
Downloading posts for user @robin, starting from 0
WebFingering...
Using Pleroma compatibility mode
Traceback (most recent call last):
  File "main.py", line 199, in <module>
    j = j['first']
KeyError: 'first'

The issue seems to be that the JSON being returned does not contain a next field (presumably because there are no next posts to load) and so the script assumes Pleroma compatibility mode is needed, which then causes the rest of the script to fail.

Grabing posts from Pleroma fails

1. Clean run without any toots.db generated yet
2. Bot account is following this accounts (raru.me are mastodon and the other two are pleroma instances)
   • [email protected]
   • [email protected]
   • [email protected]
   • [email protected]
3. Running the bot:

Using config.json as configuration file
Downloading posts for user @[email protected], starting from 0
WebFingering...
Using standard mode
Downloading and saving posts............................................ Done!
Downloading posts for user @[email protected], starting from 0
WebFingering...
Using standard mode
Downloading and saving posts................................. Done!
Downloading posts for user @[email protected], starting from 0
WebFingering...
Using standard mode
Downloading and saving posts........... Done!
Downloading posts for user @[email protected], starting from 0
WebFingering...
Using standard mode
Traceback (most recent call last):
File "/app/main.py", line 215, in
j = r.json()
File "/usr/lib/python3.8/site-packages/requests/models.py", line 897, in json
return complexjson.loads(self.text, **kwargs)
File "/usr/lib/python3.8/json/init.py", line 357, in loads
return _default_decoder.decode(s)
File "/usr/lib/python3.8/json/decoder.py", line 337, in decode
obj, end = self.raw_decode(s, idx=_w(s, 0).end())
File "/usr/lib/python3.8/json/decoder.py", line 355, in raw_decode
raise JSONDecodeError("Expecting value", s, err.value) from None
json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)

4. As it can be seen on the fedi.absturztau.me fails to grab posts.
5. Re-run the sync:

Using config.json as configuration file
Downloading posts for user @[email protected], starting from 40dd49ac-3b6e-4ea6-9f99-7415a57bed3c
WebFingering...
Using standard mode
Traceback (most recent call last):
File "/app/main.py", line 215, in
j = r.json()
File "/usr/lib/python3.8/site-packages/requests/models.py", line 897, in json
return complexjson.loads(self.text, **kwargs)
File "/usr/lib/python3.8/json/init.py", line 357, in loads
return _default_decoder.decode(s)
File "/usr/lib/python3.8/json/decoder.py", line 337, in decode
obj, end = self.raw_decode(s, idx=_w(s, 0).end())
File "/usr/lib/python3.8/json/decoder.py", line 355, in raw_decode
raise JSONDecodeError("Expecting value", s, err.value) from None
json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)

6. And now the other pleroma instance (kitsune.cafe) starts failing too. I tried waiting for half an hour in case it was a rate-limit but it still kept failing

TODO: make it so that it removes the @ and instance but keeps the name

TODO on

mstdn-ebooks/main.py

Line 57 in dd7a4ef

# TODO: make it so that it removes the @ and instance but keeps the name

Not sure if this is still valid but still listed as a TODO. @Lynnesbian can you confirm if this is still broken as it seems to be working on live bots?

Filter out words from learning

Certain terms keep coming up (names of instances, names of people) that I don't want. Allowing a filter to remove certain words from being used in posts by filtering them from being learned from would make the Bot Experience Better

Add option to strip quotes and parens

This would prevent unclosed quotes and parentheses while mostly leaving things intact.

Allow for customisation of mentions

mstdn-ebooks shouldn't make posts that start with mentions

Fix multithreading

Multithreading doesn't actually work properly, I need to fix it

Dependabot couldn't authenticate with https://pypi.python.org/simple/

Dependabot couldn't authenticate with https://pypi.python.org/simple/.

You can provide authentication details in your Dependabot dashboard by clicking into the account menu (in the top right) and selecting 'Config variables'.

View the update logs.