Fixing time signature encoding/decoding

[Natooz]

This PR fixes a bug on the encoding and decoding or TimeSignature tokens, especially for Bar/Position-based tokenizers. (#74)
REMI, CPWord, Octuple and MMM are affected.

[codecov]

Codecov Report

Attention: 11 lines in your changes are missing coverage. Please review.

Comparison is base (bd408a9) 89.68% compared to head (8b20a28) 90.23%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #76      +/-   ##
==========================================
+ Coverage   89.68%   90.23%   +0.55%     
==========================================
  Files          31       31              
  Lines        4275     4303      +28     
==========================================
+ Hits         3834     3883      +49     
+ Misses        441      420      -21     

Files	Coverage Δ
miditok/midi_tokenizer.py	90.40% <100.00%> (+1.33%)	⬆️
miditok/tokenizations/midi_like.py	94.40% <100.00%> (+0.65%)	⬆️
miditok/tokenizations/mmm.py	95.16% <100.00%> (+2.82%)	⬆️
miditok/tokenizations/tsd.py	98.29% <100.00%> (+0.85%)	⬆️
tests/test_bpe.py	90.10% <ø> (ø)
tests/test_io_formats.py	90.32% <100.00%> (+0.32%)	⬆️
tests/test_multitrack.py	84.12% <100.00%> (+0.52%)	⬆️
tests/test_saving_loading_config.py	92.00% <ø> (ø)
miditok/tokenizations/cp_word.py	94.91% <95.74%> (+1.63%)	⬆️
miditok/tokenizations/remi.py	96.64% <94.87%> (+1.66%)	⬆️
... and 2 more

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.