parseLammpsLogMdext.py

# from Tanooj

import numpy as np
import pandas as pd
import argparse
import glob
import os
import sys

curr_dir = os.getcwd()

parser = argparse.ArgumentParser(description="write lammps thermo output to a csv file")
parser.add_argument('-i', action="store", dest="input_file")
args = parser.parse_args()

infile = args.input_file

parse = True

if parse:


    # purge all empty strings from token lists - and return list
    def purge(tokens):
            return [t for t in tokens if len(t) >= 1]

    def is_num(s):
        try:
            float(s)
        except ValueError:
            return False
        else:
            return True

    #    Step         PotEng         Volume         Press           Temp  

    # Step Temp PotEng TotEng Press Volume lambda lj dlj 1 gT qNa qCl
    # Step Temp PotEng TotEng Press Volume lambda lj dlj 1

    # Step Temp PotEng TotEng Press Volume lambda lj dlj 1 qNa qCl

    # ======== READ INPUT FILE ===========

    

    with open(infile,'r') as f:
            lines = f.readlines()

    lines = [l.strip('\n') for l in lines]

    datastart_idxs = []
    dataend_idxs = []

    linecounter = 0
    for line in lines:
            # if line.startswith('Step'):
            if line.startswith('   Step'):
                    datastart_idxs.append(linecounter)
            linecounter += 1
    # print(f'{datastart_idxs=}')

    # if truncated file - need to remove one of the starts
    del datastart_idxs[-1]

    linecounter = 0
    for line in lines:
            # loop of time is end
            if (line.startswith('Loop time of')) and (lines.index(line) > np.min(datastart_idxs)):
                    dataend_idxs.append(linecounter)
            linecounter += 1

    print(datastart_idxs)
    print(dataend_idxs)

    # may not need with the reset with mdext.... 
    # print(purge(lines[datastart_idxs[0]].split(' ')))
    # sys.exit(0)
    header = purge(lines[datastart_idxs[1]].split(' ')) # should be the same header for all  but isn't!! so use second one
    # we only want the TI data not initial equilibration
    data = []
    for idx in range(len(datastart_idxs)): # index of indices
            data_chunk = lines[datastart_idxs[idx]+1:dataend_idxs[idx]]
            for d in data_chunk:
                    # print(purge(d.split(' ')))
                    # sys.exit(0)
                    # if len(purge(d.split(' ')))>5:
                    #     print(purge(d.split(' ')))
                    if d.split(' ')[0] != 'WARNING:':
                        data.append(purge(d.split(' ')))
    print(header)
    print(np.array(data).shape)
    df = pd.DataFrame(data, columns=header)
    outfile = infile + '.csv'
    df.to_csv(outfile)

import matplotlib.pyplot as plt

df = pd.read_csv(infile + '.csv')
print(df)
plt.plot(df.PotEng)
# plt.yscale('log')
plt.savefig(f'{infile}.pdf')