udate min fval

[FFroehlich]

This corrects the computation of min_val in the waterfall plot. This is a breaking change, since now log-plots yield use the standard offset_y=1-minval by default instead of an offst_y=0 by default. However, I believe this makes sense, since the previous implementation made it impossible to use the offset_y=1-minval, since this would always erroneously trigger

pyPESTO/pypesto/visualize/misc.py

Line 115 in cb8bffc

if (scale_y == 'log10') and (min_val + offset_y <= 0.):

. This implementation corrects this and the previous can still be restored by specifying offset_y=0.

[codecov-io]

Codecov Report

Merging #593 (4b04495) into develop (160c2a8) will decrease coverage by 53.30%.
The diff coverage is 27.95%.

@@             Coverage Diff              @@
##           develop     #593       +/-   ##
============================================
- Coverage    88.16%   34.85%   -53.31%     
============================================
  Files           79       91       +12     
  Lines         5257     5726      +469     
============================================
- Hits          4635     1996     -2639     
- Misses         622     3730     +3108     

Impacted Files	Coverage Δ
pypesto/ensemble/covariance_analysis.py	18.36% <0.00%> (-0.39%)	⬇️
pypesto/objective/aggregated.py	25.49% <0.00%> (-74.51%)	⬇️
pypesto/petab/importer.py	0.00% <0.00%> (-85.90%)	⬇️
pypesto/visualize/waterfall.py	14.63% <0.00%> (-81.67%)	⬇️
pypesto/store/read_from_hdf5.py	51.32% <5.55%> (-6.57%)	⬇️
pypesto/store/save_to_hdf5.py	18.96% <7.69%> (-43.18%)	⬇️
pypesto/objective/amici.py	20.32% <10.00%> (-69.51%)	⬇️
pypesto/visualize/misc.py	13.18% <11.11%> (-71.75%)	⬇️
pypesto/visualize/sampling.py	10.89% <12.42%> (-67.74%)	⬇️
pypesto/ensemble/ensemble.py	12.90% <22.22%> (-67.87%)	⬇️
... and 92 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 bc73d34...4b04495. Read the comment docs.

[yannikschaelte]

sounds fine for me. @paulstapor @DantongWang any concerns?