Select: support user-provided calibration results

[dilpath]

Currently, the design is that:

• a user can provide previously-calibrated models as user_calibrated_models
• PEtab Select will put these models into the candidate space, instead of the uncalibrated versions
• pyPESTO still needs to skip these models

The purpose is that the path through model space is still reproducible, because models are not skipped. TODO: fully handle this in PEtab Select, so pyPESTO never sees already-calibrated models, except when it asks PEtab Select for the final results.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 66.66667% with 2 lines in your changes are missing coverage. Please review.

Project coverage is 84.26%. Comparing base (8016726) to head (3b51af9).

Files	Patch %	Lines
pypesto/select/method.py	66.66%	2 Missing :warning:

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

Additional details and impacted files

```diff @@ Coverage Diff @@ ## develop #1338 +/- ## =========================================== - Coverage 84.28% 84.26% -0.03% =========================================== Files 152 152 Lines 12441 12446 +5 =========================================== + Hits 10486 10487 +1 - Misses 1955 1959 +4 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.