fixing multi-stream decoding with several identical programs

[Natooz]

Following #74, this PR now handles multi-stream decoding with several identical programs (for all tokenizers).
Previously, identical programs resulted in merged tracks.

[codecov]

Codecov Report

Attention: 25 lines in your changes are missing coverage. Please review.

Comparison is base (01cdac0) 90.29% compared to head (b1a37dc) 90.25%.
Report is 1 commits behind head on main.

❗ Current head b1a37dc differs from pull request most recent head 9ef5c23. Consider uploading reports for the commit 9ef5c23 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #81      +/-   ##
==========================================
- Coverage   90.29%   90.25%   -0.04%     
==========================================
  Files          31       31              
  Lines        4305     4425     +120     
==========================================
+ Hits         3887     3994     +107     
- Misses        418      431      +13     

Files	Coverage Δ
miditok/midi_tokenizer.py	90.51% <100.00%> (-0.12%)	⬇️
miditok/tokenizations/structured.py	97.03% <100.00%> (+0.26%)	⬆️
tests/tests_utils.py	77.01% <50.00%> (-0.34%)	⬇️
miditok/tokenizations/cp_word.py	95.07% <96.49%> (+0.16%)	⬆️
miditok/tokenizations/tsd.py	97.99% <93.75%> (-0.30%)	⬇️
tests/test_one_track.py	79.16% <75.00%> (+1.89%)	⬆️
miditok/tokenizations/octuple.py	87.93% <84.00%> (+0.70%)	⬆️
miditok/tokenizations/remi.py	95.93% <90.90%> (-0.71%)	⬇️
miditok/tokenizations/midi_like.py	92.96% <78.57%> (-1.45%)	⬇️

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