PERF: Use direct bitshift logic instead of _extract_bits function

[greglucas]

int.from_bytes() is about 2x slower than a direct bitshift of two values, and we avoid some other checks as well.

python -m timeit "b=b'\x01\x02\x03\x04\x05\x06'*100; int.from_bytes(b[-2:], 'big')" 
5000000 loops, best of 5: 80.2 nsec per loop

python -m timeit "b=b'\x01\x02\x03\x04\x05\x06'*100; (b[-2] << 8) | b[-1]"
5000000 loops, best of 5: 47.8 nsec per loop

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.86%. Comparing base (5028734) to head (92da56a).
Report is 10 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #150      +/-   ##
==========================================
- Coverage   91.86%   91.86%   -0.01%     
==========================================
  Files          36       36              
  Lines        2557     2556       -1     
==========================================
- Hits         2349     2348       -1     
  Misses        208      208              

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.