search

prisms-center / CASMcode

First-principles statistical mechanical software for the study of multi-component crystalline solids
Other
105 stars 69 forks source link

1) Pure states are not in the hull data. 2) how can i get the ECI/multiplicity plot? #319

Open darjaved opened 10 months ago

darjaved commented 10 months ago

from casm.project import Project, Selection, write_eci from casm.learn import open_halloffame, open_input, checkhull, to_json import pandas import matplotlib.pyplot as plt

casm_learn_input = 'fit.json' selection = 'train'

input = open_input(casm_learn_input)

hall = open_halloffame(input["halloffame_filename"])

indiv_i = 0 checkhull(input, hall, indices=[indiv_i])

proj = Project()

indiv = hall[indiv_i] write_eci(proj, hall[0].eci, to_json(indiv_i, hall[indiv_i]))

sel = Selection(proj, selection, all=False)

comp_a = 'comp(a)' is_calculated = 'is_calculated' dft_Ef = 'formation_energy' clex_Ef = 'clex(formation_energy)' sel.query([comp_a, is_calculated, dft_Ef, clex_Ef])

df = sel.data df_calc = df[df.loc[:,is_calculated] == 1].apply(pandas.to_numeric, errors='ignore') h0 = plt.scatter(df_calc[comp_a], df_calc[dft_Ef], facecolors='none', edgecolors='b', label='dft') h0_hull = plt.plot(indiv.dft_gs[comp_a], indiv.dft_gs[dft_Ef], 'bo-', label='nolegend') h1 = plt.scatter(df_calc[comp_a], df_calc[clex_Ef], color='r', marker='.', label='clex') h1_hull = plt.plot(indiv.clex_gs[comp_a], indiv.clex_gs[clex_Ef], 'r.-', label='nolegend')

plt.xlabel(comp_a) plt.ylabel('Formation energy (eV/unit cell)')

plt.legend() plt.show() plt.savefig("hull.pdf")

bpuchala commented 10 months ago

It could be that the pure states are not selected in the "train" file or not calculated.

Multiplicities can read from the "basis.json" file.

darjaved commented 10 months ago

I thought setting the references will add the pure states. I have done calculations not within the CASM, So, i have to calculate for the supercell of the pure states also? If i have a supercell of pure state, will that match with my prim file which has two kinds of atoms?

bpuchala commented 10 months ago

Maybe I misunderstood what you mean by "pure states". I expected you meant a configuration that is an extreme composition, like if you had a binary A-B FCC prim and it was pure A or pure B. If the "pure states" that you want to use as a reference are not configurations of the prim, then you would have to add them manually in the script if you want them included in the plot.

darjaved commented 10 months ago

you understood right, i mean the pure A and pure B. i have given the references as casm ref --set [x,y]. but when i try to construct the convex hull, its not considering the reference states (which are pure states here).

bpuchala commented 10 months ago

The reference states are not required to be a particular configuration. They can be any composition-energy pair, and so they are not included in the "train" file which that script is plotting the hull of.

You can include the reference states manually if you want, by adding them as additional data in the pandas dataframe "df_calc".