Add BEETL data to MOABB for benchmarking

[Samuel-Boehm]

Hi MOABB team,

I'm a PhD student working on BCI benchmarking and would like to contribute code to integrate the BEETL dataset into MOABB. Before proceeding with the review, I want to be transparent that I am not the owner or license holder of the original BEETL dataset - I'm simply contributing code to make this publicly available dataset accessible through MOABB for benchmarking purposes.

The dataset is publicly available through Figshare and was published as part of a BCI competition. I've implemented the loader following MOABB's conventions and patterns as good as I can. I basically followed what is done for Stieger2021 as I saw that they also uploaded the data to Figshare.

Key implementation details:

• Dataset split into leaderboard and final evaluation phases (This might be confusing first but I wanted to stick to the original split as it was during the competition.
• Handles two different data formats (Dataset A and B) with different sampling rates and channel configurations
• Proper handling of labels and trial information
• After the competition was done a .txt with the hidden labels was released, this is also downloaded and added to the data here.

Important Implementation Note:
In the current implementation, I concatenate all runs/races for each subject into a single continuous Raw object. This differs from the original data structure where trials were stored separately. I made this choice to align with MOABB's typical usage patterns, but I'm open to discussion if this should be handled differently.

I would greatly appreciate:

1. Your review of the code quality and adherence to MOABB standards
2. Guidance on whether it's appropriate to include this dataset given I'm not the original data holder

Greetings from Germany, Merry Christmas and a Happy New Year!
Samuel