simsopt.mhd.Vmec silently accepts bogus inputs -> Vmec.get_input returns bogus data

[eguiraud-pf]

Reproducer:

In [4]: !echo "not a valid vmec input file" > /tmp/input.invalid

In [5]: !cat /tmp/input.invalid
not a valid vmec input file

In [6]: simsopt.mhd.Vmec("/tmp/input.invalid").get_input()
 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
  SEQ =    1 TIME SLICE  0.0000E+00
  PROCESSING INPUT.invalid
  THIS IS PARVMEC (PARALLEL VMEC), VERSION 9.0
  Lambda: Full Radial Mesh. L-Force: hybrid full/half.

  COMPUTER:    OS:    RELEASE:   DATE = Jul 25,2024  TIME = 22:28:32
Out[6]: '&INDATA\n! This file created by simsopt on July 25 2024, 22:28:32\n\n! ---- Geometric parameters ----\nNFP = 1\nLASYM = F\n\n! ---- Resolution parameters ----\nMPOL = 6\nNTOR = 0\nNTHETA = 18\nNZETA = 1\nNS_ARRAY    =     31\nNITER_ARRAY =    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100    100\nFTOL_ARRAY  =  1e-10\n\n! ---- Boundary toroidal flux ----\nPHIEDGE = 1.0\n\n! ---- Pressure profile specification ----\nPMASS_TYPE = "power_series"\nAM = 0.0 \nPRES_SCALE = 1.0\n\n! ---- Profile specification of iota or current ----\nNCURR = 0\nPIOTA_TYPE = "power_series"\nAI = 0.0 \n\n! ---- Other numerical parameters ----\nDELT = 1.0\nNSTEP = 10\n\n! ---- Boundary shape. Array index order is (n, m) ----\n/\n\n'

Not sure what the best way to validate the input would be without actually trying to run the underlying Fortran. Maybe the Python side could at least check that the first line of the file is &INDATA.