fixes for a couple of things in leaderboard cog

[b9r5]

Description

siro requested a change in leaderboard_cog.py where to use interaction.response.send_message for initial validations and thereafter use interaction.followup.send.

I noticed a problem where the submission content was not being decoded. As a result, we get rows in the submission table like this (note the binary-looking data in the code column):

clusterdev=# select * from leaderboard.submission where id=8;
-[ RECORD 1 ]---
id              | 8
problem_id      | 1
name            | train.py
user_id         | 704502776683692072
code            | \x696d706f727420746f7263680a0a64656620637573746f6d5f736f66746d617828783a20746f7263682e54656e736f722c2064696d3a20696e74203d202d3129202d3e20746f7263682e54656e736f723a0a2020202072657475726e20746f7263682e6e6e2e66756e6374696f6e616c2e736f66746d617828782c2064696d3d64696d290a
submission_time | 2024-12-12 18:24:07.769627-08
score           | 0.012549400329589844

After this change, we get rows in the submission table like this (note that the code column is legible after the change):

clusterdev=# select * from leaderboard.submission where id=9;
 id | problem_id |   name   |      user_id       |                                code                                 |        submission_time        |        score         
----+------------+----------+--------------------+---------------------------------------------------------------------+-------------------------------+----------------------
  9 |          1 | train.py | 704502776683692072 | import torch                                                       +| 2024-12-12 18:40:29.929962-08 | 0.011950969696044922
    |            |          |                    |                                                                    +|                               | 
    |            |          |                    | def custom_softmax(x: torch.Tensor, dim: int = -1) -> torch.Tensor:+|                               | 
    |            |          |                    |     return torch.nn.functional.softmax(x, dim=dim)                 +|                               | 
    |            |          |                    |                                                                     |                               | 
(1 row)

A couple of screenshots of tests:

1. Submitting to a leaderboard that does not exist:

2. Submitting a file that is not UTF-8 encoded:

Checklist

Before submitting this PR, ensure the following steps have been completed:

I manually verified that /run modal works, and that /run github works for Nvidia.
For some reason, I am not able to execute /run github for AMD at the moment.

• Run the slash command /verifyruns on your own server.
  • Run the cluster bot on your server:

    python discord-bot.py

  • Start training runs with the slash command /verifyruns.
  • Verify that the bot eventually responds with:

    ✅ All runs completed successfully!
    

    (It may take a few minutes for all runs to finish. In particular, the GitHub
    runs may take a little longer. The Modal run is typically quick.)
    For more information on running a cluster bot on your own server, see
    README.md.