Updated the default version of COSMIC and introduced a new flag

[kulvi90]

Changed the default COSMIC version to 8451e3134846b577fd661ac3b96

Introduced a new cosmic flag : rtmsflag

* For low metallicities (Z < 0.0008), calculated radius at main
  sequence turnoff (rtms) for massive stars (M > 200 Msun) using default
  COSMIC prescriptions (SSE interpolation and extrapolation formulae)
  can become non-physical (i.e. rtms <= 0).

* rtmsflag controls which prescriptions are used to calculate the
  rtms for stars. (for details see the zfuncs.f file in COSMIC)

* rtmsflag = 0 : uses the SSE rtms for stars EXCEPT if M>200Msun
  and Z<0.0008(~0.04Zsun). For the exception, an extrapolation
  is used (check the code in function rtmsf for details). This
  ad hoc extrapolation works well for stars with M < 4000Msun
  for Z >= 0.01Zsun. For lower metallicities, you may still run
  into issues of rtms<0 for very massive stars (i.e. for Z<0.01Zsun
  use at your own risk)

* rtmsflag > 0 uses rtms from 1D stellar evolution code simulations.
  rtmsflag = 1 : BoOST tracks
  rtmsflag = 2 : BPASS tracks
  ONLY common metallicity cases between BoOST & BPASS are coded.

* BoOST metallicities = [1.1e-4, 2.1e-4, 1e-3, 2e-3] corresponding
  to [dwarfD, IZw18, dwarfA, SMC] models (Szecsi et al. (2022))

* BPASS metallicities = [1e-4, *2e-4*, 1e-3, 2e-3].
  NOTE : For BPASS, we used a power law to fit the rtms v/s mzams
  values for each metallicity. We have coded the best fit power-laws
  for the above BPASS metallicities here.Since, Z = 2e-4 model is
  not available in the BPASS tracks, we assume the same rtms v/s
  mzams power law as Z = 1e-4 for Z = 2e-4.