import xml.etree.ElementTree as ET import numpy as np from pymatgen.io.vasp.outputs import Vasprun import csv Nline = 20 # number of kpoints in a line class KPOINTS_OPT_Parser: def __init__(self, file_path, Nline=20): self.Nline = Nline self.file_path = file_path self.vasprun, self.xml_root = self.parse_vasprun() self.fermi_level = self.get_fermi_level() self.kpoints = self.get_kpoints() self.weights_list = self.get_weights() self.reciprocal_lattice = self.vasprun.final_structure.lattice.reciprocal_lattice.matrix self.bands_data = self.get_eigenvalues() self.kpath = self.get_kpath() def parse_vasprun(self): vasprun = Vasprun(self.file_path) tree = ET.parse(self.file_path) xml_root = tree.getroot() return vasprun, xml_root def get_fermi_level(self): kpoints_opt_dos = self.xml_root.find(".//dos[@comment='kpoints_opt']") fermi_level = float(kpoints_opt_dos.find(".//i[@name='efermi']").text.strip()) return fermi_level def get_kpoints(self): kpoints_opt = self.xml_root.find(".//eigenvalues_kpoints_opt") kpointlist = kpoints_opt.find(".//varray[@name='kpointlist']") kpoints = [np.array([float(x) for x in v.text.split()]) for v in kpointlist.findall('v')] return kpoints def get_weights(self): kpoints_opt = self.xml_root.find(".//eigenvalues_kpoints_opt") weights = kpoints_opt.find(".//varray[@name='weights']") weights_list = [float(v.text) for v in weights.findall('v')] return weights_list def get_eigenvalues(self): fermi_level = self.fermi_level kpoints_opt = self.xml_root.find(".//eigenvalues_kpoints_opt") eigenvalues = kpoints_opt.find(".//eigenvalues") bands_data = [] for spin_set in eigenvalues.findall('.//set[@comment]'): spin = spin_set.get('comment') for kpoint_set in spin_set.findall('.//set[@comment]'): kpoint = kpoint_set.get('comment') for r in kpoint_set.findall('r'): bands_data.append({ 'spin': spin, 'kpoint': kpoint, 'eigenvalue': float(r.text.strip()) - fermi_level }) return bands_data def kpoints_opt_dos(self): kpoints_opt_dos = self.xml_root.find(".//dos[@comment='kpoints_opt']") dos_data = kpoints_opt_dos.find(".//total/array") return dos_data def dos_spin(self, dos_data): fermi_level = self.fermi_level dos_spin = {} for spin_set in dos_data.findall(".//set[@comment]"): spin = spin_set.get('comment').split()[1] # "spin 1" -> "1" dos_spin[spin] = [] for r in spin_set.findall('r'): energy, total, integrated = [float(x) for x in r.text.split()] dos_spin[spin].append({ 'energy': energy - fermi_level, # フェルミ準位を0とする 'total': total, 'integrated': integrated }) return dos_spin def get_kpath(self): reciprocal_lattice = self.reciprocal_lattice kpoints = self.kpoints Nline = self.Nline kpath = [0] for i in range(1, len(kpoints)): if i % Nline == 0: kpath.append(kpath[-1]) else: dk = np.linalg.norm(np.dot(kpoints[i] - kpoints[i - 1], reciprocal_lattice)) kpath.append(kpath[-1] + dk) return kpath def eigenvalues_per_kpoint(self): bands_data = self.get_eigenvalues() eigenvalues_per_kpoint = {} for band in bands_data: kpoint_index = int(band['kpoint'].split()[1]) - 1 spin = band['spin'] if (spin, kpoint_index) not in eigenvalues_per_kpoint: eigenvalues_per_kpoint[(spin, kpoint_index)] = [] eigenvalues_per_kpoint[(spin, kpoint_index)].append(band['eigenvalue']) return eigenvalues_per_kpoint def write_kptopt_eig_csv(file_path, Nline=20): parser = KPOINTS_OPT_Parser(file_path, Nline) kpoints = parser.kpoints weights_list = parser.weights_list kpath = parser.kpath # 各k点に対する固有値をまとめる eigenvalues_dict = parser.eigenvalues_per_kpoint() # CSVファイルに書き込む with open('kpoints_opt_eigenvalues.csv', 'w', newline='') as csvfile: writer = csv.writer(csvfile) # 最大バンド数を取得 max_bands = max(len(eigenvalues) for eigenvalues in eigenvalues_dict.values()) # ヘッダーの書き込み header = ['spin', 'kpoint_index', 'kpath', 'weight', 'kx', 'ky', 'kz'] + [f'band_{i+1}' for i in range(max_bands)] writer.writerow(header) # スピンとk点インデックスでソート sorted_keys = sorted(eigenvalues_dict.keys()) # 各k点ごとに行を書き込む for key in sorted_keys: spin, kpoint_index = key kpoint = kpoints[kpoint_index] weight = weights_list[kpoint_index] row = [spin, kpoint_index, kpath[kpoint_index], weight, *kpoint] + eigenvalues_dict[key] writer.writerow(row) def write_kptopt_dos_csv(file_path, Nline=20): # vasprunParserオブジェクトの作成 parser = KPOINTS_OPT_Parser(file_path, Nline) # vasprun.xmlファイルの読み込み dos_data = parser.kpoints_opt_dos() dos_spin = parser.dos_spin(dos_data) # CSVファイルに書き込み with open('kpoints_opt_dos.csv', 'w', newline='') as csvfile: writer = csv.writer(csvfile) # ヘッダーの書き込み writer.writerow(['spin', 'energy', 'total', 'integrated']) # データの書き込み for spin, data in dos_spin.items(): for record in data: writer.writerow([spin, record['energy'], record['total'], record['integrated']]) def write_high_symmetry_kpoints_opt(file_path, Nline=20): # make vasprunParser object parser = KPOINTS_OPT_Parser(file_path, Nline) # read vasprun.xml bands_data = parser.bands_data kpath = parser.get_kpath() # Calculate the minimum and maximum eigenvalues eigenvalues = [band['eigenvalue'] for band in bands_data] min_eigenvalue = min(eigenvalues) * 1.1 max_eigenvalue = max(eigenvalues) * 1.1 # Create a list for the first column kpath_column = [kpath[i] for i in range(Nline-1, len(kpath), Nline) for _ in range(2)] # (min_eigenvalue, max_eigenvalue, max_eigenvalue, min_eigenvalue)のパターンでリストに追加 pattern = [min_eigenvalue, max_eigenvalue, max_eigenvalue, min_eigenvalue] eigenvalue_column = [pattern[i % 4] for i in range(len(kpath_column))] # return kpath_column, eigenvalue_column # Write the data to the CSV file with open('kpath_origin.csv', 'w', newline='') as csvfile: writer = csv.writer(csvfile) # Write the header writer.writerow(['kpath', 'eigenvalue']) # Write the data rows for kpath_val, eigenvalue_val in zip(kpath_column, eigenvalue_column): writer.writerow([kpath_val, eigenvalue_val]) if __name__ == '__main__': write_kptopt_eig_csv("vasprun.xml", Nline) write_kptopt_dos_csv("vasprun.xml", Nline) write_high_symmetry_kpoints_opt("vasprun.xml", Nline)