vasprun.xml to json file

将vasprun.xml 的configuration 转为 json file

主要用于FitSNAP程序的前期数据集的获取

import json
import numpy as np

#read energy lattice pos force stress from vasprun
i=0
j=0
k=0
num_atom=0
force=[]
s_step=1#int(input("starting step:\n"))
d_step=1#int(input("interval of steps:\n"))

#Dataset-->struc_inf-->data
Dataset={}
struc_inf={}
struc_data={}

struc_inf["Label:"]="SNAP dataset"
struc_inf["LatticeStyle"]="angstrom"
struc_inf["EnergyStyle"]="electronvolt"
struc_inf["StressStyle"]="bar"
struc_inf["AtomTypeStyle"]="chemicalsymbol"
struc_inf["PositionsStyle"]="bar"
struc_inf["ForcesStyle"]="electronvoltperangstrom"

with open('PPOSCAR','r') as pposcar:
    text=pposcar.readlines()
    atom_list=[]
    atom_dict=[]
    one_list=['one']
    if text[5].strip() < text[6].strip():
        print('please check the input format of PPOSCAR')
        # for n in range(num_atom):
            # position0.append(text[n+7].split())
    else:
        num_class=text[5].strip().split(' ')
        print(num_class)
        for c in range(len(num_class)):
            num_atom=num_atom+int(text[6].strip().split(' ')[c])
            print(int(text[6].strip().split(' ')[c]))
            one_list[0]=num_class[c]
            atom_dict=one_list*(int(text[6].strip().split(' ')[c]))
            atom_list=atom_list+atom_dict
        # for n in range(num_atom):
            # position0.append(text[n+8].split())
    # print(num_atom)


with open('vasprun.xml','r') as    vasprun:
    text=vasprun.readlines()
    num_lines=len(text)
    for n in range(num_lines):

        # the original positions of atoms
        # if text[n] == "  <varray name=\"positions\" >\n":
            # for m in range(num_atom):
                # position0.append(text[n+m+1].split())
        if text[n] == "  <varray name=\"forces\" >\n":
            lattice=np.mat(np.zeros((3,3)))
            position=np.mat(np.zeros((num_atom,3)))
            force=np.mat(np.zeros((num_atom,3)))
            stress=np.mat(np.zeros((3,3)))

            i=i+1
            if i > s_step-1:
                j=j+1
                if ((j-1)-d_step*((j-1)//d_step))==0:
                    k=k+1
                    #force.append(i)
                    for lindex in np.arange(0,3,1):
                        lattice_str=[float(lat) for lat in text[n-num_atom-14+lindex].split()[1:4]]
                        lattice[lindex,:]=np.array(lattice_str)
                        stress_str=[float(stre) for stre in text[n+num_atom+3+lindex].split()[1:4]]
                        #print(stress_str)
                        stress[lindex,:]=np.array(stress_str)*1000

                    for m in range(num_atom):
                        force_str=[float(f) for f in text[n+m+1].split()[1:4]]
                        force[m,:]=np.array(force_str)

                        position_str=[float(postr) for postr in text[n+m-num_atom-2].split()[1:4]]
                        position_frac=np.array(position_str)
                        position_cart=np.dot(lattice,position_frac.T)
                        position[m,:]=position_cart

                    struc_data["NumAtoms"]=num_atom
                    struc_data["Lattice"]=lattice.tolist()
                    struc_data["Energy"]=float(text[n+num_atom+9].split()[2])
                    struc_data["Stress"]=stress.tolist()
                    struc_data["AtomTypes"]=atom_list
                    struc_data["Positions"]=position.tolist()
                    struc_data["Forces"]=force.tolist()

                    struc_inf["Data"]=[struc_data]
                    Dataset["Dataset"]=struc_inf

                    config=json.dumps(Dataset)
                    config_json=open("MD_{}.json".format(i),mode="w")
                    print('# SNAP datasets',file=config_json)
                    print(config, file=config_json)
print("total_steps=",i)
#print(j)
print("output_steps=",k)