Fixes steadystate guesses

[FFroehlich]

• reinitialize objective during startpoint resampling to avoid using steadystate guesses.
• fix shape bug introduced via construct condition specific plist in parameter mapping AMICI-dev/AMICI#1487 and account for plist in edata when slicing amici results #621
• fix check for elementwise changes in linear updates

[yannikschaelte]

looks good to me if tests pass. re-initializing in startpoint method certainly makes sense.

[codecov-commenter]

Codecov Report

Merging #715 (b742c0d) into develop (b53e6f2) will increase coverage by 41.69%.
The diff coverage is 66.66%.

@@             Coverage Diff              @@
##           develop     #715       +/-   ##
============================================
+ Coverage    46.38%   88.08%   +41.69%     
============================================
  Files           96       96               
  Lines         6521     6523        +2     
============================================
+ Hits          3025     5746     +2721     
+ Misses        3496      777     -2719     

Impacted Files	Coverage Δ
pypesto/startpoint/util.py	85.71% <50.00%> (-1.79%)	⬇️
pypesto/objective/amici.py	82.69% <100.00%> (+17.78%)	⬆️
pypesto/petab/importer.py	64.08% <0.00%> (-21.00%)	⬇️
pypesto/predict/result.py	57.26% <0.00%> (-17.10%)	⬇️
pypesto/predict/amici_predictor.py	82.71% <0.00%> (-14.82%)	⬇️
pypesto/objective/amici_util.py	75.43% <0.00%> (-8.78%)	⬇️
pypesto/petab/pysb_importer.py	73.33% <0.00%> (-6.67%)	⬇️
pypesto/objective/amici_calculator.py	90.00% <0.00%> (-3.34%)	⬇️
pypesto/objective/base.py	81.63% <0.00%> (+1.53%)	⬆️
pypesto/optimize/optimizer.py	90.09% <0.00%> (+3.45%)	⬆️
... and 60 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b53e6f2...b742c0d. Read the comment docs.