Calling mutation from python script on macOS

[BBArrosDias]

I am trying to call some mutation tools from a python script. As a simple example, I am trying to use mppequil in the following way:

os.system('mppequil -h')

and I get the following issue:

dyld: Library not loaded: @rpath/libmutation++.dylib

In the terminal, if I do echo $DYLD_LIBRARY_PATH I see the path to libmutation++.dylib, but inside the script, if I do the same, the answer is empty.

I am using the latest macOS Catalina, and I know that doing the same in Linux OS it works.
I dig up a bit in google, and it seems that problem is related to System Integrity Protect (SIP) and in this website, they recommend to disable it (which is not secure):

https://help.mulesoft.com/s/article/Variables-LD-LIBRARY-PATH-DYLD-LIBRARY-PATH-are-ignored-on-MAC-OS-if-System-Integrity-Protect-SIP-is-enable

If I create a bash file and try to do the same I end up the same problem, but if inside the file I include the export DYLD_LIBRARY_PATH=.... it works.

Any idea???

[rdbisme]

Let's say that calling the binaries from python is not so cool, because then you need to parse the output. Moreover you should use subprocess instead of os.system.

In any case those are not fixes to your problem I guess. I'll try next week on my Mac, but I guess that the real solution here would be to provide python binding to the C++ code.

[jbscoggi]

@rubendibattista, I agree that calling M++ binaries from python is not the best and it would be really nice to have the Python wrapper working. In this case though, I think Bruno is using python as a scripting language to run another code which is linked with M++, so that would not really help for this situation.

[BBArrosDias]

@jbscoggi is completely right. I am using python to run other codes that are linked to m++. I have the same problem with bash scripts which are fixed by export the DYLB_library_path on the script.

[rdbisme]

Ok, I tested, and it works on my side.

1. Compile and install wherever you want. Let's call your install path MUTATIONPP_INSTALL_PREFIX. 2.export DYLD_LIBRARY_PATH=$(realpath $MUTATIONPP_INSTALL_PREFIX/install/lib):$DYLD_LIBRARY_PATH (realpath is not available on macOs natively, you need to install it using brew or macports. In any case you need to provide the absolute path...)
2. Run this python script:

   
   import subprocess

subprocess.call(["/install/bin/mppequil", "-h"])

Returns this: 

Usage: ./install/bin/mppequil [OPTIONS] mixture Compute equilibrium properties for mixture over a set of temperatures and pressures using the Mutation++ library.

-h, --help          prints this help message
    --no-header     no table header will be printed
-T                  temperature range in K "T1:dT:T2" or simply T (default = 300:100:15000 K)
-P                  pressure range in Pa "P1:dP:P2" or simply P (default = 1 atm)
-B                  magnitude of the magnetic field in teslas (default = 0 T)
-m                  list of mixture values to output (see below)
-s                  list of species values to output (see below)
-r                  list of reaction values to output (see below)
-o                  list of other values to output (see below)
    --species-list  instead of mixture name, use this to list species in mixture
    --elem-x        set elemental mole fractions (ex: N:0.8,O:0.2)
    --elem-comp     set elemental composition with a name from the mixture file
    --thermo-db     overrides thermodynamic database type (NASA-7, NASA-9, RRHO)
    --scientific    outputs in scientific format with given precision

Mixture values (example format: "1-3,7,9-11"): 0 : Th [K] heavy particle temperature 1 : P [Pa] pressure 2 : B [T] magnitude of magnetic field 3 : rho [kg/m^3] density 4 : nd [1/m^3] number density 5 : Mw [kg/mol] molecular weight 6 : Cp_eq [J/mol-K] equilibrium specific heat at constant pressure 7 : H [J/mol] mixture enthalpy 8 : S [J/mol-K] entropy 9 : Cp_eq [J/kg-K] equilibrium specific heat at constant pressure 10: H [J/kg] mixture enthalpy 11: H-H0 [J/kg] mixture enthalpy minus the enthalpy at 0K 12: S [J/kg-K] entropy 13: Cv_eq [J/kg-K] equilibrium specific heat at constant volume 14: Cp [J/mol-K] frozen specific heat at constant pressure 15: Cv [J/mol-K] frozen specific heat at constant volume 16: Cp [J/kg-K] frozen specific heat at constant pressure 17: Cv [J/kg-K] frozen specific heat at constant volume 18: gam_eq [-] equilibrium ratio of specific heats 19: gamma [-] frozen ratio of specific heat 20: Ht [J/mol] translational enthalpy 21: Hr [J/mol] rotational enthalpy 22: Hv [J/mol] vibrational enthalpy 23: Hel [J/mol] electronic enthalpy 24: Hf [J/mol] formation enthalpy 25: Ht [J/kg] translational enthalpy 26: Hr [J/kg] rotational enthalpy 27: Hv [J/kg] vibrational enthalpy 28: Hel [J/kg] electronic enthalpy 29: Hf [J/kg] formation enthalpy 30: e [J/mol] mixture energy 31: e [J/kg] mixture energy 32: mu [Pa-s] dynamic viscosity 33: lambda [W/m-K] mixture equilibrium thermal conductivity 34: lam_reac [W/m-K] reactive thermal conductivity 35: lam_bb [W/m-K] Butler-Brokaw reactive thermal conductivity 36: lam_soret [W/m-K] Soret thermal conductivity 37: lam_int [W/m-K] internal energy thermal conductivity 38: lam_h [W/m-K] heavy particle translational thermal conductivity 39: lam_e [W/m-K] electron translational thermal conductivity 40: sigma [S/m] electric conductivity (B=0) 41: a_f [m/s] frozen speed of sound 42: a_eq [m/s] equilibrium speed of sound 43: Eam [V/K] ambipolar electric field (SM Ramshaw) 44: drho/dP [kg/J] equilibrium density derivative w.r.t pressure

Species values (same format as mixture values): 0 : X [-] mole fractions 1 : dX/dT [1/K] partial of mole fraction w.r.t. temperature 2 : Y [-] mass fractions 3 : rho [kg/m^3] mass densities 4 : conc [mol/m^3] molar concentrations 5 : Cp [J/mol-K] specific heats at constant pressure 6 : H [J/mol] enthalpies 7 : S [J/mol-K] entropies 8 : G [J/mol] Gibbs free energies 9 : Cp [J/kg-K] specific heats at constant pressure 10: H [J/kg] enthalpies 11: S [J/kg-K] entropies 12: G [J/kg] Gibbs free energies 13: J [kg/m^2-s] Species diffusion fluxes (SM Ramshaw) 14: omega [kg/m^3-s] production rates due to reactions 15: Omega11 [m^2] (1,1) pure species collision integrals 16: Omega22 [m^2] (2,2) pure species collision integrals 17: Chi^h [-] heavy thermal diffusion ratios 18: Dm [m^2/s] mixture averaged diffusion coefficients

Reaction values (same format as mixture values): 0 : kf [mol,m,s,K] forward reaction rate coefficients 1 : kb [mol,m,s,K] backward reaction rate coefficients

Other values (same format as mixture values): 0 : Dij [m^2/s] multicomponent diffusion coefficients 1 : pi_i [-] element potentials 2 : N_p [mol] phase moles 3 : iters [-] number of continuation step iterations 4 : newts [-] total number of newton iterations 5 : Fp_k [kg/m-Pa-s] elemental diffusion fluxes per pressure gradient 6 : Ft_k [kg/m-K-s] elemental diffusion fluxes per temperature gradient 7 : Fz_k [kg/m-s] elemental diffusion fluxes per element mole fraction gradient 8 : sigmaB [S/m] anisotropic electric conductivity 9 : lamB_e [W/m-K] anisotropic electron thermal conductivity

Example: ./install/bin/mppequil -T 300:100:15000 -P 101325 -m 1-3,8 air11

So it looks like it works as expected. 

[BBArrosDias]

@rubendibattista It works like a charm. Thanks a lot. You, sir, deserve a free beer.