recalc with updated SH12A

[nbassler]

Recalculation and testing of fix for https://gitlab.au.dk/shieldhit/shieldhit/-/issues/439

[nbassler]

v1.0.1-45-g9c9cc584 introduces new issue fd04f1b#diff-6c2d8efdda9ba26991b3ebbf0d0655045dc879e3a3f1b116cd4cf00980607e77
NANs in water scorer.

data/sh12a/results/plan2/NB_Z_narrow_dose_water__p2.png

[nbassler]

debugging

     Primary particle no.      281000      ETR:   0 hours 29 minutes 13 seconds
*** Warning:  SH_MDEDX_WRAP: z > 20 unavailable. (z,a):          21          39
*** Warning:  SH_MDEDX_WRAP: z > 20 unavailable. (z,a):          21          39
---------------------------------------------------Print particle:
  jpart  = 25
  PDG    = 1000210390
  id     = 1
  Z      = 21
  A      = 39
  amass  =  36314.09731 MeV/c**2
  amu    =     38.98479 u
  gen    = 1
  nprim  = 281484
  weight = 1.000000
POINT Override settings: 4 1.000000 -nan
POINT val = energy * volinv / rho_med * mstpr = 1.230358 * 2.531646 / 1.032000 * -nan =  -nan
POINT NAN in scoring

The nans are due to missing stopping power tables for fragments with z > 21. (solid water medium has some heavier elements added)

[nbassler]

fixed in SH12A v1.0.1-49-gb2876ad9

[nbassler]

Issue #3 primaries still weird, but uncertainties are probably quite ok. @mdsainth maybe take a quick look, you can average the 20 bin target bins if it helps.

If you agree that the test plan looks OK, lets recalculate all plans, but as you suggested for a single target bin, and then merge this branch into master.

filed here: DataMedSci/pymchelper#718

[nbassler]

Is queued for recalculation, but will be closed when #6 is fixed too.

[nbassler]

FYI: average of the same 200 jobs with a new test version of pymchelper

DataMedSci/pymchelper#723

now primaries reported should work and be consistent.

Error bars are however missing now, still TODO.

[nbassler]

@mdsainth errorbars there now ... tell us if you see something weird

[nbassler]

DataMedSci/pymchelper#723
commit DataMedSci/pymchelper@d6c5ac3

[nbassler]

This looks OK to me