Experiment/compare performance (#12)

* feat(scripts): add Slurm scheduler support for model comparisons

Introduces scheduler integration to scripts/compare_models.py to allow running baseline and SpatialTranscriptFormer model training in parallel on Slurm HPC clusters.

- Add --slurm execution mode to generate Slurm scripts with POSIX-compatible paths

- Add --submit option to submit via sbatch and register collection dependency

- Add --collect-only mode to aggregate reports/plots directly from outputs

- Add Slurm resource customisation CLI flags

- Use dynamic SLURM_SUBMIT_DIR directory to prevent hardcoded Windows paths

* style: improve code readability with consistent formatting adjustments

- Applied consistent line breaks in multi-line function calls.
- Aligned and formatted CLI argument definitions for clarity.
- Standardised string formatting with proper quotation mark usage.
- Enhanced indentation for better structure and readability.

Signed-off-by: BenjaminIsaac0111 <12176376+BenjaminIsaac0111@users.noreply.github.com>

---------

Signed-off-by: BenjaminIsaac0111 <12176376+BenjaminIsaac0111@users.noreply.github.com>

Author: BenjaminIsaac0111