running rhyolite melts with alphamelts front end

[jhaddad1]

Is it possible to run rhyolite melts using the alpha melts front end? If not, is there any plan to make a scriptable front end for rhyolite melts?

[Paula]

Hi,

Sorry for the delay in replying. There are various options, especially if you use Python... which I seem to remember you do.

There is a beta of alphaMELTS 2 that includes the rhyoliteMELTS thermodynamic models in the virtual machine alphaMELTS, but I wouldn't really recommend trying it. It's really more alpha than beta, and it predates MELTS having gone open source in April of this year:

https://gitlab.com/ENKI-portal/xMELTS/

Despite the name, the xMELTS repository on GitLab is actually for RhyoliteMELTS 1.02, 1.10, and 1.20 and the pMELTS model. There are various packages you can build, which would provide a scriptable interfaces to RhyoliteMELTS; you can see the GitLab site for more details.  There is also a simple PyMELTS package, which has its own README for build instructions. If you check out the linux_pymelts branch (which isn't really Linux-specific) then I've added the GUI's "Find Liquidus" functionality and isentropic calculations to PyMELTS, as well as accumulated minor bug fixes to the code as a whole. There's a Jupyter Notebook example of the 'tutorial' example that is featured on the General Usage board here, and which we started with in each of the alphaMELTS workshops.

A version of alphaMELTS 2 / libalphaMELTS that incorporates the public MELTS code linked above, plus some features of pHMELTS etc. is close to ready. I've been busy with other things the last few months but hope to get a 'proper' beta out soon. I have included the existing PyMELTS wrappers, and intend to add ones specific to alphaMELTS. So, if you get some calculations working on PyMELTS from ENKI you will be able to transfer it to the more fully featured alphaMELTS engine with few, if any, changes.

Within the context of ENKI (enki-portal.org), PyMELTS has been superseded by the ThermoEngine (formerly the PhaseObjC thermoengine). The equilibration routines use the MELTS family of models but you can also perform calculations using the Berman, Holland and Powell, or Stixrude-Lithgow-Bertelloni endmember databases. You can request an account on the ENKI-portal hub from Mark Ghiorso to try it out. In the meantime you can browse the documentation and Notebook examples under Documentation->ThermoEngine. The nice thing about the hub is that it can be accessed from any web browser so is platform independent.

The libpHMELTS library will be written in C++ and loosely based on / ported from ThermoEngine, so will also include RhyoliteMELTS etc. We will try to make the Python interface to libpHMELTS similar to the ThermoEngine one, but we'll have to see how ThermoEngine evolves over time. At the moment it uses Rubicon to interface with the underlying Objective C, which means offline access is limited for now (currently Mac only). Either way, it is unlikely that the ThermoEngine and libpHMELTS interfaces will be as close to each other as the corresponding PyMELTS and libalphaMELTS ones.

Hope that helps. If you are interested in getting going on PyMELTS, or trying out the Python<->libpHMELTS interface when I get to it, then let me know.

Paula

[jhaddad1]

Paula-

Thanks for the very thorough response. You are correct that I use a python wrapper to run MELTS. I was inquiring for a future project that I'm looking to get into, once that is approaching in a few months, I'd love to check out the python specific MELTS packages you're developing.

-James

[Paula]

Hi James,

Great! I'll keep you posted. Are you on the same email address as your profile shows?

Paula

[jhaddad1]

Just saw this. Yes, I am using that email address, and please do keep me posted. As always, thanks!

[Paula]

Will do. I'm back working on this so will hopefully have something available shortly.

Paula