Fix Chain Burn-in Calculation in Thermodynamic Integration

[arrjon]

Several weeks ago, I introduced a function to calculate the log evidence from a parallel tempering run using thermodynamic integration. Initially, only the first chain was assessed for convergence and the burn-in was determined. Now, I've enhanced the process by computing the burn-in individually for each chain, enhancing the validity and accuracy of the approximation.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 84.46%. Comparing base (3163687) to head (40d3574).

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #1361      +/-   ##
===========================================
+ Coverage    84.43%   84.46%   +0.02%     
===========================================
  Files          157      157              
  Lines        12881    12899      +18     
===========================================
+ Hits         10876    10895      +19     
+ Misses        2005     2004       -1     

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

[vwiela]

As discussed, adding a option whether to take the maximum of the burn-ins over all chains or each chain individually, would be great. But for the moment fine as it is. Thanks

[vwiela]

Thanks for the warnings and option implementation.