manimcommunity / discordmanimator Goto Github PK

Would be possible with this script.

import re
from textwrap import indent
inputstring= """
class EllipseExample(Scene):
    def construct(self):
        ellipse_1 = Ellipse(width=2.0, height=4.0, color=BLUE_B)
        self.add(ellipse_1)
"""
inputstring= indent(inputstring, '    ')
template="""
\"""
Examples
--------

.. manim:: CLASSNAMEHERE
    :save_last_frame:

    CODEHERE
\"""
"""

outputstring=template.replace("CODEHERE", inputstring)
#possible_scene_names = ["Scene", "GraphScene"]
first_word = "class "
scene_name = "Scene" # does not support e.g. GraphScene yet. Maybe with a for loop ?
classname=re.search(fr'{first_word}(.*?).({scene_name})', inputstring).group(1)
outputstring= outputstring.replace("CLASSNAMEHERE", classname)
print(outputstring)

"""

outputstring=template.replace("CODEHERE", inputstring)
first_word = "class "
scene_name = "Scene" # does not support e.g. GraphScene yet. Maybe with a for loop ? #possible_scene_names = ["Scene", "GraphScene"]
classname=re.search(fr'{first_word}(.*?).({scene_name})', inputstring).group(1)
outputstring= outputstring.replace("CLASSNAMEHERE", classname)
print(outputstring)

I'm not sure if this command is staying after the hackathon, but if it is having black be applied would make code look a lot cleaner.

It would be nice if the bot can detect when the code in a previous manimate command was changed, and have it delete the old rendering and rerender the animation.

Making !mhelp more descriptive by adding info about !mdocstring, !mdoc, !msearch would be nice, as beginners can know how to use DiscordManimator bot efficiently in discord.

Currently, !mhelp is returning info only about !manimate. It is much better if we include info about !mdocstring, !mdoc, !msearch as well.

you could use glitch.com to host the discord bot

It would be nice if you could specify the manim version you want !mdoc for for users troubleshooting a problem on old versions. Something like !mdoc 0.4.0 LaggedStart

Currently the bot resets the cli flags when pressing the rerender button and automagically rerender when pressing the change settings button which makes it pretty difficult to test scenes which actually use command line flags because you have to retype them everytime you want to render something.

          It seems as if long (>30 sec?) render queries block the heartbeat of the bot; the communication with docker should be made asynchronous if possible. (There is https://aiodocker.readthedocs.io/en/latest/index.html which could be a suitable replacement for the docker package.)

Originally posted by @behackl in #38 (comment)

If you msearch 2 or more queries quickly, and react with an emote to one of them, the other results will also be affected. This should definitely not be the case, reactions on a message should affect that message only.

Currently, the bot does not send anything if attempting to manimate something results in a traceback that is too long.