MQTPredictor update

[flowerthrower]

Mostly bug fixes that became clear when trying to update MQTPredictor based on the new device class.

[codecov]

Codecov Report

Attention: Patch coverage is 95.55556% with 2 lines in your changes are missing coverage. Please review.

Project coverage is 91.9%. Comparing base (8f2db01) to head (9dc6b0c).

Files	Patch %	Lines
src/mqt/bench/utils.py	66.6%	2 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##            main    #299     +/-   ##
=======================================
- Coverage   93.2%   91.9%   -1.3%     
=======================================
  Files         46      46             
  Lines       2407    2427     +20     
=======================================
- Hits        2244    2232     -12     
- Misses       163     195     +32     

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

[nquetschlich]

Thanks a lot for those changes. I just have a few questions to better understand them.

Additionally, I have one larger comment that probably requires a moderate effort: We should substitute the M3 for the M2 instead of keeping both devices. This means, that all files must be clear of the M2 device including the configuration file.

For the benchviewer, we currently stick to the M2 device and, therefore, those files must not be adjusted. The rationale is, that we will soon update to Qiskit v1.0.0 and then re-create all benchmark files and do not want to do it now additionally since it comes with a significant effort.

[nquetschlich]

resolves #279

[nquetschlich]

As soon as the code comments are clarified and the currently failing test is fixed, this PR is ready to be merged.