Fix energy_bias_resolution_from_energy_dispersion

[maxnoe]

This function still assumed the wrong normalization of the energy dispersion fixed in pyirf 0.10.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (dbacc33) 95.35% compared to head (9a4baa1) 95.36%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #268      +/-   ##
==========================================
+ Coverage   95.35%   95.36%   +0.01%     
==========================================
  Files          60       60              
  Lines        3097     3109      +12     
==========================================
+ Hits         2953     2965      +12     
  Misses        144      144              

Files	Coverage Δ
pyirf/benchmarks/energy_bias_resolution.py	76.00% <100.00%> (+1.53%)	⬆️
pyirf/benchmarks/tests/test_bias_resolution.py	71.21% <100.00%> (+4.54%)	⬆️

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

[maxnoe]

@Tobychev the mistake was obvious (the function was overlooked in the update), but could you anyway maybe check with your plot in #267 ?

[Tobychev]

In case there is nothing in an energy bin energy_bias_resolution_from_energy_dispersion will return the upper limit of the dispersion axis. It is probably better to return NaN