TB2J 0.7.7.1__py3-none-any.whl → 0.8.0__py3-none-any.whl

TB2J/abacus/stru_api.py

@@ -0,0 +1,2020 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Wed Jun 13 10:31:30 2018
+@author: shenzx
+
+Modified on Wed Aug 01 11:44:51 2022
+@author: Ji Yu-yang
+"""
+
+import re
+import warnings
+import numpy as np
+import os
+import shutil
+from pathlib import Path
+
+from ase import Atoms
+from ase.units import Bohr, Hartree, GPa, mol, _me, Rydberg
+from ase.utils import lazymethod, lazyproperty, reader, writer
+from ase.calculators.singlepoint import SinglePointDFTCalculator, arrays_to_kpoints
+
+_re_float = r"[-+]?\d+\.*\d*(?:[Ee][-+]\d+)?"
+AU_to_MASS = mol * _me * 1e3
+UNIT_V = np.sqrt(Hartree / AU_to_MASS)
+
+# --------WRITE---------
+
+# WRITE ABACUS INPUT -START-
+
+
+@writer
+def write_input(fd, parameters=None):
+    """Write the INPUT file for ABACUS
+
+    Parameters
+    ----------
+    fd: str
+        The file object to write to
+    parameters: dict
+        The dictionary of all paramters for the calculation.
+    """
+    from copy import deepcopy
+
+    params = deepcopy(parameters)
+    params["dft_functional"] = (
+        params.pop("xc") if params.get("xc") else params.get("dft_functional", "pbe")
+    )
+    for key in [
+        "pp",
+        "basis",
+        "pseudo_dir",
+        "basis_dir",
+        "orbital_dir",
+        "offsite_basis_dir",
+        "kpts",
+        "knumber",
+        "kmode",
+        "knumbers",
+        "scaled",
+    ]:
+        params.pop(key, None)
+
+    lines = []
+    lines.append("INPUT_PARAMETERS")
+    lines.append("# Created by Atomic Simulation Enviroment")
+    for key, val in params.items():
+        if val is not None:
+            lines.append(str(key) + " " * (40 - len(key)) + str(val))
+    lines.append("")
+    fd.write("\n".join(lines))
+
+
+# WRITE ABACUS INPUT -END-
+
+# WRITE ABACUS KPT -START-
+
+
+@writer
+def write_kpt(fd=None, parameters=None, atoms=None):
+    """Write the KPT file for ABACUS
+
+    Parameters
+    ----------
+    fd: str
+        The file object to write to
+    parameters: dict
+        The dictionary of all paramters for the calculation.
+        If `gamma_only` or `kspacing` in `parameters`, it will not output any files by ase
+    atoms: Atoms
+        It should be set, when `parameters['kpts']` is `dict`. Parameters of `cell.bandpath` and
+        `ase.calculators.calculator.kpts2sizeandoffsets` are supported to be keys
+        of the dictionary `parameters['kpts']`, and k-points will be generated by ASE.
+    """
+
+    gamma_only = parameters.get("gamma_only", 0)
+    kspacing = parameters.get("kspacing", 0.0)
+    kpts = parameters.get("kpts", None)
+    koffset = parameters.get("koffset", 0)
+    if gamma_only is not None and gamma_only == 1:
+        return
+    elif kspacing is not None and kspacing > 0.0:
+        return
+    elif kpts is not None:
+        if isinstance(kpts, dict) and "path" not in kpts:
+            from ase.calculators.calculator import kpts2sizeandoffsets
+
+            kgrid, shift = kpts2sizeandoffsets(atoms=atoms, **kpts)
+            koffset = []
+            for i, x in enumerate(shift):
+                assert x == 0 or abs(x * kgrid[i] - 0.5) < 1e-14
+                koffset.append(0 if x == 0 else 1)
+        else:
+            kgrid = kpts
+    else:
+        kgrid = "gamma"
+
+    if isinstance(koffset, int):
+        koffset = [koffset] * 3
+
+    if isinstance(kgrid, dict) or hasattr(kgrid, "kpts"):
+        from ase.calculators.calculator import kpts2ndarray
+
+        kmode = "Direct"
+        kgrid = kpts2ndarray(kgrid, atoms=atoms)
+    elif isinstance(kgrid, str) and (kgrid == "gamma"):
+        kmode = "Gamma"
+        knumber = 0
+        kgrid = [0, 0, 0]
+    elif "gamma" in kpts:
+        kmode = "Gamma" if kpts["gamma"] else "MP"
+    else:
+        kmode = parameters.get("kmode", "Gamma")
+
+    lines = []
+    lines.append("K_POINTS")
+    if kmode in ["Gamma", "MP"]:
+        knumber = 0
+        lines.append(f"{knumber}")
+        lines.append(f"{kmode}")
+        lines.append(" ".join(map(str, kgrid)) + " " + " ".join(map(str, koffset)))
+    elif kmode in ["Direct", "Cartesian"]:
+        knumber = parameters.get("knumber", len(kgrid))
+        lines.append(f"{knumber}")
+        lines.append(f"{kmode}")
+        assert isinstance(knumber, int) and knumber > 0
+        for n in range(knumber):
+            lines.append(
+                f"{kgrid[n][0]:0<12f} {kgrid[n][1]:0<12f} {kgrid[n][2]:0<12f} {1/knumber}"
+            )
+    elif kmode in ["Line"]:
+        knumber = parameters.get("knumber", len(kgrid))
+        lines.append(f"{knumber}")
+        lines.append(f"{kmode}")
+        knumbers = parameters.get("knumbers", [10] * (knumber - 1) + [1])
+        for n in range(knumber):
+            lines.append(
+                f"{kgrid[n][0]:0<12f} {kgrid[n][1]:0<12f} {kgrid[n][2]:0<12f} {knumbers[n]}"
+            )
+    else:
+        raise ValueError(
+            "The value of kmode is not right, set to "
+            "Gamma, MP, Direct, Cartesian, or Line."
+        )
+    lines.append("")
+    fd.write("\n".join(lines))
+
+
+# WRITE ABACUS KPT -END-
+
+
+def _copy_files(file_list, src, dst, env, name):
+    if not src:
+        # environment variable for PP paths
+        if env in os.environ:
+            src = os.environ[env]
+        else:
+            src = "./"
+            # raise NotFoundErr(
+            #     f"Can not set directory of {name} according to environment variable {env}")
+
+    for val in file_list:
+        src_file = os.path.join(src, val.strip())
+        dst_file = os.path.join(dst, val.strip())
+        if os.path.exists(dst_file):
+            continue
+        elif os.path.exists(src_file):
+            shutil.copyfile(src_file, dst_file)
+        else:
+            raise FileNotFoundError(f"Can't find {name} for ABACUS calculation")
+
+
+# WRITE ABACUS PP -START-
+def copy_pp(pp_list, pseudo_dir=None, directory="./"):
+    """Copy pseudo-potential files from `pseudo_dir` to `directory`
+
+    Parameters
+    ----------
+    pp_list: list
+        List of pseudo-potential files, e.g. ['Si.UPF', 'C.UPF']
+    pseudo_dir: str
+        The src directory includes pseudo-potential files. If None,
+        it will get directory from environment variable `ABACUS_PP_PATH`
+    directory: str
+        The dst directory
+    """
+    _copy_files(
+        pp_list, pseudo_dir, directory, "ABACUS_PP_PATH", "pseudo-potential files"
+    )
+
+
+# WRITE ABACUS PP -END-
+
+
+# WRITE ABACUS basis -START-
+def copy_basis(basis_list, basis_dir=None, directory="./"):
+    """Copy LCAO basis files from `basis_dir` to `directory`
+
+    Parameters
+    ----------
+    basis_list: list
+        List of LCAO basis files, e.g. ['Si.orb', 'C.orb']
+    basis_dir: str
+        The src directory includes LCAO basis files. If None,
+        it will get directory from environment variable `ABACUS_ORBITAL_PATH`
+    directory: str
+        The dst directory
+    """
+    _copy_files(basis_list, basis_dir, directory, "ABACUS_ORBITAL_PATH", "basis files")
+
+
+# WRITE ABACUS basis -END-
+
+# WRITE ABACUS basis -START-
+
+
+def copy_offsite_basis(offsite_basis_list, offsite_basis_dir=None, directory="./"):
+    """Copy off-site ABFs basis files from `basis_dir` to `directory`
+
+    Parameters
+    ----------
+    offsite_basis_list: list
+        List of off-site ABFs basis files, e.g. ['abfs_Si.dat', 'abfs_C.dat']
+    offsite_basis_dir: str
+        The src directory includes off-site ABFs basis files. If None,
+        it will get directory from environment variable `ABACUS_ABFS_PATH`
+    directory: str
+        The dst directory
+    """
+    _copy_files(
+        offsite_basis_list,
+        offsite_basis_dir,
+        directory,
+        "ABACUS_ABFS_PATH",
+        "off-site ABFs basis files",
+    )
+
+
+# WRITE ABACUS basis -END-
+
+
+# WRITE ABACUS STRU -START-
+
+
+def judge_exist_stru(stru=None):
+    if stru is None:
+        return False
+    else:
+        return True
+
+
+def read_ase_stru(stru=None, coordinates_type="Cartesian"):
+    from ase.constraints import FixAtoms, FixCartesian
+
+    fix_cart = np.ones((len(stru), 3), dtype=int).tolist()
+    for constr in stru.constraints:
+        for i in constr.index:
+            if isinstance(constr, FixAtoms):
+                fix_cart[i] = [0, 0, 0]
+            elif isinstance(constr, FixCartesian):
+                fix_cart[i] = constr.mask
+            else:
+                UserWarning("Only `FixAtoms` and `FixCartesian` are supported now.")
+
+    if judge_exist_stru(stru):
+        atoms_list = []
+        atoms_sort = []
+        atoms_position = []
+        atoms_masses = []
+        atoms_magnetism = []
+        atoms_fix = []
+        atoms_all = stru.get_chemical_symbols()
+
+        # sort atoms according to atoms
+        atoms_dict = {}
+        for idx, atoms_all_name in enumerate(atoms_all):
+            if atoms_all_name not in atoms_dict:
+                atoms_dict[atoms_all_name] = []
+            atoms_dict[atoms_all_name].append(idx)
+        for symbol in atoms_dict:
+            atoms_sort.extend(atoms_dict[symbol])
+        atoms_list = list(
+            atoms_dict.keys()
+        )  # Python >= 3.7 for keeping the order of keys
+
+        for atoms_list_name in atoms_list:
+            atoms_position.append([])
+            atoms_masses.append([])
+            atoms_magnetism.append(0)
+            atoms_fix.append([])
+
+        # get position, masses, magnetism from ase atoms
+        # TODO: property 'magmoms' is not implemented in ABACUS
+        if coordinates_type == "Cartesian":
+            for i in range(len(atoms_list)):
+                for j in range(len(atoms_all)):
+                    if atoms_all[j] == atoms_list[i]:
+                        atoms_position[i].append(list(stru.get_positions()[j]))
+                        atoms_masses[i] = stru.get_masses()[j]
+                        # atoms_magnetism[i] += np.array(stru[j].magmom)
+                        atoms_fix[i].append(fix_cart[j])
+                # atoms_magnetism[i] = np.linalg.norm(atoms_magnetism[i])
+
+        elif coordinates_type == "Direct":
+            for i in range(len(atoms_list)):
+                for j in range(len(atoms_all)):
+                    if atoms_all[j] == atoms_list[i]:
+                        atoms_position[i].append(list(stru.get_scaled_positions()[j]))
+                        atoms_masses[i] = stru.get_masses()[j]
+                        # atoms_magnetism[i] += np.array(stru[j].magmom)
+                        atoms_fix[i].append(fix_cart[j])
+                # atoms_magnetism[i] = np.linalg.norm(atoms_magnetism[i])
+
+        else:
+            raise ValueError(
+                "'coordinates_type' is ERROR," "please set to 'Cartesian' or 'Direct'"
+            )
+
+        return (
+            atoms_list,
+            atoms_sort,
+            atoms_masses,
+            atoms_position,
+            atoms_magnetism,
+            atoms_fix,
+        )
+
+
+def write_input_stru_sort(atoms_sort=None):
+    if atoms_sort is None:
+        return "Please set right atoms sort"
+    else:
+        with open("ase_sort.dat", "w") as fd:
+            for idx in atoms_sort:
+                fd.write("%s\n" % idx)
+
+
+def write_input_stru_core(
+    fd,
+    stru=None,
+    pp=None,
+    basis=None,
+    offsite_basis=None,
+    coordinates_type="Cartesian",
+    atoms_list=None,
+    atoms_position=None,
+    atoms_masses=None,
+    atoms_magnetism=None,
+    fix=None,
+    init_vel=False,
+):
+    if not judge_exist_stru(stru):
+        return "No input structure!"
+
+    elif atoms_list is None:
+        return "Please set right atoms list"
+    elif atoms_position is None:
+        return "Please set right atoms position"
+    elif atoms_masses is None:
+        return "Please set right atoms masses"
+    elif atoms_magnetism is None:
+        return "Please set right atoms magnetism"
+    else:
+        fd.write("ATOMIC_SPECIES\n")
+        for i, elem in enumerate(atoms_list):
+            if pp:
+                pseudofile = pp.get(elem, "")
+            else:
+                pseudofile = ""
+            temp1 = " " * (4 - len(atoms_list[i]))
+            temp2 = " " * (14 - len(str(atoms_masses[i])))
+            atomic_species = (
+                atoms_list[i] + temp1 + str(atoms_masses[i]) + temp2 + pseudofile
+            )
+
+            fd.write(atomic_species)
+            fd.write("\n")
+
+        if basis:
+            fd.write("\n")
+            fd.write("NUMERICAL_ORBITAL\n")
+            for i, elem in enumerate(atoms_list):
+                orbitalfile = basis[elem]
+                fd.write(orbitalfile)
+                fd.write("\n")
+
+        if offsite_basis:
+            fd.write("\n")
+            fd.write("ABFS_ORBITAL\n")
+            for i, elem in enumerate(atoms_list):
+                orbitalfile = offsite_basis[elem]
+            fd.write(orbitalfile)
+            fd.write("\n")
+        # modified output by QuantumMisaka to synchroize with ATOMKIT
+        fd.write("\n")
+        fd.write("LATTICE_CONSTANT\n")
+        fd.write(f"{1/Bohr:.6f}\n")
+        fd.write("\n")
+
+        fd.write("LATTICE_VECTORS\n")
+        for i in range(3):
+            for j in range(3):
+                temp3 = str("{:0<12f}".format(stru.get_cell()[i][j])) + " " * 3
+                fd.write(temp3)
+                fd.write("   ")
+            fd.write("\n")
+        fd.write("\n")
+
+        fd.write("ATOMIC_POSITIONS\n")
+        fd.write(coordinates_type)
+        fd.write("\n")
+        fd.write("\n")
+        k = 0
+        for i in range(len(atoms_list)):
+            fd.write(atoms_list[i])
+            fd.write("\n")
+            fd.write(str("{:0<12f}".format(float(atoms_magnetism[i]))))
+            fd.write("\n")
+            fd.write(str(len(atoms_position[i])))
+            fd.write("\n")
+
+            for j in range(len(atoms_position[i])):
+                temp4 = str("{:0<12f}".format(atoms_position[i][j][0])) + " "
+                temp5 = str("{:0<12f}".format(atoms_position[i][j][1])) + " "
+                temp6 = str("{:0<12f}".format(atoms_position[i][j][2])) + " "
+                sym_pos = (
+                    temp4
+                    + temp5
+                    + temp6
+                    + f"{fix[i][j][0]:.0f} {fix[i][j][1]:.0f} {fix[i][j][2]:.0f} "
+                )
+                if init_vel:  # velocity in unit A/fs ?
+                    sym_pos += f"v {stru.get_velocities()[j][0]} {stru.get_velocities()[j][1]} {stru.get_velocities()[j][2]} "
+
+                if isinstance(stru[k].magmom, float):
+                    sym_pos += f"mag {stru[k].magmom} "
+                elif isinstance(stru[k].magmom, list) or isinstance(
+                    stru[k].magmom, np.ndarray
+                ):
+                    if len(stru[k].magmom) == 3:
+                        sym_pos += f"mag {stru[k].magmom[0]} {stru[k].magmom[1]} {stru[k].magmom[2]} "
+                    elif len(stru[k].magmom) == 1:
+                        sym_pos += f"mag {stru[k].magmom[0]} "
+                k += 1
+                fd.write(sym_pos)
+                fd.write("\n")
+            fd.write("\n")
+
+
+@writer
+def write_abacus(
+    fd, atoms=None, pp=None, basis=None, offsite_basis=None, scaled=True, init_vel=False
+):
+    """Write the STRU file for ABACUS
+
+    Parameters
+    ----------
+    fd: str
+        The file object to write to
+    atoms: atoms.Atoms
+        The Atoms object for the requested calculation
+    pp: dict
+        The pseudo-potential file of each elements, e.g. for SiC, {'Si':'Si.UPF', 'C':'C.UPF'}
+    basis: dict
+        The basis file of each elements for LCAO calculations, e.g. for SiC, {'Si':'Si.orb', 'C':'C.orb'}
+    offsite_basis: dict
+        The offsite basis file of each elements for HSE calculations with off-site ABFs, e.g. for SiC, {'Si':'Si.orb', 'C':'C.orb'}
+    scaled: bool
+        If output STRU file with scaled positions
+    init_vel: bool
+        if initialize velocities in STRU file
+    """
+
+    if scaled:
+        coordinates_type = "Direct"
+    else:
+        coordinates_type = "Cartesian"
+
+    if not judge_exist_stru(atoms):
+        return "No input structure!"
+
+    else:
+        (
+            atoms_list,
+            atoms_sort,
+            atoms_masses,
+            atoms_position,
+            atoms_magnetism,
+            atoms_fix,
+        ) = read_ase_stru(atoms, coordinates_type)
+
+        write_input_stru_core(
+            fd,
+            atoms,
+            pp,
+            basis,
+            offsite_basis,
+            coordinates_type,
+            atoms_list,
+            atoms_position,
+            atoms_masses,
+            atoms_magnetism,
+            atoms_fix,
+            init_vel,
+        )
+        write_input_stru_sort(atoms_sort)
+
+
+# WRITE ABACUS STRU -END-
+
+# --------READ---------
+
+# Read KPT file  -START-
+
+
+@reader
+def read_kpt(fd, cell=None):
+    """Read ABACUS KPT file and return results dict.
+
+    If `cell` is not None, a BandPath object will be returned.
+    """
+    contents = fd.read()
+    contents = re.compile(r"#.*|//.*").sub("", contents)
+    lines = [i.strip() for i in contents.split("\n")]
+    kmode = None
+    knumber = None
+    kpts = None
+    koffset = None
+    knumbers = None
+    weights = None
+    kmode = lines[2]
+    knumber = int(lines[1].split()[0])
+    if kmode in ["Gamma", "MP"]:
+        kpts = np.array(lines[3].split()[:3], dtype=int)
+        koffset = np.array(lines[3].split()[3:], dtype=float)
+        return {"mode": kmode, "number": knumber, "kpts": kpts, "offset": koffset}
+    elif kmode in ["Cartesian", "Direct", "Line"]:
+        klines = np.array(
+            [line.split() for line in lines[3 : 3 + knumber]], dtype=float
+        )
+        kpts = klines[:, :3]
+        if kmode in ["Cartesian", "Direct"]:
+            weights = klines[:, 3]
+            return {"mode": kmode, "number": knumber, "kpts": kpts, "weights": weights}
+        else:
+            knumbers = klines[:, 3].astype(int)
+            if cell is not None:
+                from ase.dft.kpoints import bandpath
+
+                return bandpath(kpts, cell, npoints=knumbers.sum())
+            else:
+                return {
+                    "mode": kmode,
+                    "number": knumber,
+                    "kpts": kpts,
+                    "knumbers": knumbers,
+                }
+    else:
+        raise ValueError(
+            "The value of kmode is not right, please set to Gamma, MP, Direct, Cartesian, or Line."
+        )
+
+
+# Read KPT file  -END-
+
+# Read INPUT file -START-
+
+
+@reader
+def read_input(fd):
+    """Read ABACUS INPUT file and return parameters dict."""
+    result = {}
+
+    first_line = fd.readline().strip()
+    if first_line != "INPUT_PARAMETERS":
+        raise ValueError("Missing INPUT_PARAMETERS keyword in INPUT file.")
+    for line in fd:
+        if line.startswith("#"):
+            continue
+        items = line.strip().split()
+        if not items:
+            continue
+        key = items[0]
+        value = " ".join(items[1:])
+        result[key] = value
+
+    return result
+
+
+# Read INPUT file -End-
+
+# READ ABACUS STRU -START-
+
+# Read UPF file -START-
+
+
+@reader
+def read_pp_upf(fd):
+    """Read PP UPF file and return parameters dict."""
+    result = {}
+
+    for line in fd:
+        if "<UPF version=" in line:
+            result["version"] = (
+                line.split("=")[-1].strip('"').strip().strip(r"\>").strip('"')
+            )
+        if "element" in line:
+            result["element"] = (
+                line.split("=")[-1].strip('"').strip().strip('"').strip()
+            )
+        if "pseudo_type" in line:
+            result["pseudo_type"] = line.split("=")[-1].strip('"').strip().strip('"')
+        if "relativistic" in line:
+            result["relativistic"] = line.split("=")[-1].strip('"').strip().strip('"')
+        if "is_ultrasoft" in line:
+            result["is_ultrasoft"] = (
+                False
+                if line.split("=")[-1].strip('"').strip().strip('"') == "F"
+                else True
+            )
+        if "is_paw" in line:
+            result["is_paw"] = (
+                False
+                if line.split("=")[-1].strip('"').strip().strip('"') == "F"
+                else True
+            )
+        if "is_coulomb" in line:
+            result["is_coulomb"] = (
+                False
+                if line.split("=")[-1].strip('"').strip().strip('"') == "F"
+                else True
+            )
+        if "core_correction" in line:
+            result["core_correction"] = (
+                False
+                if line.split("=")[-1].strip('"').strip().strip('"') == "F"
+                else True
+            )
+        if "functional" in line:
+            result["functional"] = line.split("=")[-1].strip('"').strip().strip('"')
+        if "z_valence" in line:
+            result["z_valence"] = float(
+                line.split("=")[-1].strip('" ').strip().strip('"')
+            )
+        if "l_max" in line:
+            result["l_max"] = int(line.split("=")[-1].strip('"').strip().strip('"'))
+
+    return result
+
+
+# Read UPF file -END-
+
+
+@reader
+def read_abacus(fd, latname=None, verbose=False):
+    """Read structure information from abacus structure file.
+
+    If `latname` is not None, 'LATTICE_VECTORS' should be removed in structure files of ABACUS.
+    Allowed values: 'sc', 'fcc', 'bcc', 'hexagonal', 'trigonal', 'st', 'bct', 'so', 'baco', 'fco', 'bco', 'sm', 'bacm', 'triclinic'
+
+    If `verbose` is True, pseudo-potential, basis and other information along with the Atoms object will be output as a dict.
+    """
+
+    from ase.constraints import FixCartesian
+
+    contents = fd.read()
+    title_str = r"(?:LATTICE_CONSTANT|NUMERICAL_DESCRIPTOR|NUMERICAL_ORBITAL|ABFS_ORBITAL|LATTICE_VECTORS|LATTICE_PARAMETERS|ATOMIC_POSITIONS)"
+
+    # remove comments and empty lines
+    contents = re.compile(r"#.*|//.*").sub("", contents)
+    contents = re.compile(r"\n{2,}").sub("\n", contents)
+
+    # specie, mass, pps
+    specie_pattern = re.compile(rf"ATOMIC_SPECIES\s*\n([\s\S]+?)\s*\n{title_str}")
+    specie_lines = np.array(
+        [line.split() for line in specie_pattern.search(contents).group(1).split("\n")]
+    )
+    symbols = specie_lines[:, 0]
+    ntype = len(symbols)
+    mass = specie_lines[:, 1].astype(float)
+    try:
+        atom_potential = dict(zip(symbols, specie_lines[:, 2].tolist()))
+    except IndexError:
+        atom_potential = None
+
+    # basis
+    aim_title = "NUMERICAL_ORBITAL"
+    aim_title_sub = title_str.replace("|" + aim_title, "")
+    orb_pattern = re.compile(rf"{aim_title}\s*\n([\s\S]+?)\s*\n{aim_title_sub}")
+    orb_lines = orb_pattern.search(contents)
+    if orb_lines:
+        atom_basis = dict(zip(symbols, orb_lines.group(1).split("\n")))
+    else:
+        atom_basis = None
+
+    # ABFs basis
+    aim_title = "ABFS_ORBITAL"
+    aim_title_sub = title_str.replace("|" + aim_title, "")
+    abf_pattern = re.compile(rf"{aim_title}\s*\n([\s\S]+?)\s*\n{aim_title_sub}")
+    abf_lines = abf_pattern.search(contents)
+    if abf_lines:
+        atom_offsite_basis = dict(zip(symbols, abf_lines.group(1).split("\n")))
+    else:
+        atom_offsite_basis = None
+
+    # deepks for ABACUS
+    aim_title = "NUMERICAL_DESCRIPTOR"
+    aim_title_sub = title_str.replace("|" + aim_title, "")
+    deep_pattern = re.compile(rf"{aim_title}\s*\n([\s\S]+?)\s*\n{aim_title_sub}")
+    deep_lines = deep_pattern.search(contents)
+    if deep_lines:
+        atom_descriptor = deep_lines.group(1)
+    else:
+        atom_descriptor = None
+
+    # lattice constant
+    aim_title = "LATTICE_CONSTANT"
+    aim_title_sub = title_str.replace("|" + aim_title, "")
+    a0_pattern = re.compile(rf"{aim_title}\s*\n([\s\S]+?)\s*\n{aim_title_sub}")
+    a0_lines = a0_pattern.search(contents)
+    atom_lattice_scale = float(a0_lines.group(1))
+
+    # lattice vector
+    if latname:
+        aim_title = "LATTICE_PARAMETERS"
+        aim_title_sub = title_str.replace("|" + aim_title, "")
+        lparam_pattern = re.compile(rf"{aim_title}\s*\n([\s\S]+?)\s*\n{aim_title_sub}")
+        lparam_lines = lparam_pattern.search(contents)
+        atom_lattice = get_lattice_from_latname(lparam_lines, latname)
+    else:
+        aim_title = "LATTICE_VECTORS"
+        aim_title_sub = title_str.replace("|" + aim_title, "")
+        vec_pattern = re.compile(rf"{aim_title}\s*\n([\s\S]+?)\s*\n{aim_title_sub}")
+        vec_lines = vec_pattern.search(contents)
+        if vec_lines:
+            atom_lattice = np.array(
+                [
+                    line.split()
+                    for line in vec_pattern.search(contents).group(1).split("\n")
+                ]
+            ).astype(float)
+        else:
+            raise Exception(
+                f"Parameter `latname` or `LATTICE_VECTORS` in {fd.name} must be set."
+            )
+    atom_lattice = atom_lattice * atom_lattice_scale * Bohr
+
+    aim_title = "ATOMIC_POSITIONS"
+    type_pattern = re.compile(rf"{aim_title}\s*\n(\w+)\s*\n")
+    # type of coordinates
+    atom_pos_type = type_pattern.search(contents).group(1)
+    assert atom_pos_type in [
+        "Direct",
+        "Cartesian",
+    ], "Only two type of atomic coordinates are supported: 'Direct' or 'Cartesian'."
+
+    block_pattern = re.compile(rf"{atom_pos_type}\s*\n([\s\S]+)")
+    block = block_pattern.search(contents).group()
+    if block[-1] != "\n":
+        block += "\n"
+    atom_magnetism = []
+    atom_symbol = []
+    # atom_mass = []
+    atom_block = []
+    for i, symbol in enumerate(symbols):
+        pattern = re.compile(rf"{symbol}\s*\n({_re_float})\s*\n(\d+)")
+        sub_block = pattern.search(block)
+        number = int(sub_block.group(2))
+
+        # symbols, magnetism
+        sym = [symbol] * number
+        masses = [mass] * number
+        atom_mags = [float(sub_block.group(1))] * number
+        for j in range(number):
+            atom_symbol.append(sym[j])
+            # atom_mass.append(masses[j])
+            atom_magnetism.append(atom_mags[j])
+
+        if i == ntype - 1:
+            lines_pattern = re.compile(
+                rf"{symbol}\s*\n{_re_float}\s*\n\d+\s*\n([\s\S]+)\s*\n"
+            )
+        else:
+            lines_pattern = re.compile(
+                rf"{symbol}\s*\n{_re_float}\s*\n\d+\s*\n([\s\S]+?)\s*\n\w+\s*\n{_re_float}"
+            )
+        lines = lines_pattern.search(block)
+        for j in [line.split() for line in lines.group(1).split("\n")]:
+            atom_block.append(j)
+    atom_block = np.array(atom_block)
+    atom_magnetism = np.array(atom_magnetism)
+
+    # position
+    atom_positions = atom_block[:, 0:3].astype(float)
+    natoms = len(atom_positions)
+
+    # fix_cart
+    if (atom_block[:, 3] == ["m"] * natoms).all():
+        atom_xyz = ~atom_block[:, 4:7].astype(bool)
+    else:
+        atom_xyz = ~atom_block[:, 3:6].astype(bool)
+    fix_cart = [FixCartesian(ci, xyz) for ci, xyz in enumerate(atom_xyz)]
+
+    def _get_index(labels, num):
+        index = None
+        res = []
+        for l in labels:
+            if l in atom_block:
+                index = np.where(atom_block == l)[-1][0]
+        if index is not None:
+            res = atom_block[:, index + 1 : index + 1 + num].astype(float)
+
+        return res, index
+
+    # velocity
+    v_labels = ["v", "vel", "velocity"]
+    atom_vel, v_index = _get_index(v_labels, 3)
+
+    # magnetism
+    m_labels = ["mag", "magmom"]
+    if "angle1" in atom_block or "angle2" in atom_block:
+        warnings.warn(
+            "Non-colinear angle-settings are not yet supported for this interface."
+        )
+    mags, m_index = _get_index(m_labels, 1)
+    try:  # non-colinear
+        if m_index:
+            atom_magnetism = atom_block[:, m_index + 1 : m_index + 4].astype(float)
+    except IndexError:  # colinear
+        if m_index:
+            atom_magnetism = mags
+
+    # to ase
+    if atom_pos_type == "Direct":
+        atoms = Atoms(
+            symbols=atom_symbol,
+            cell=atom_lattice,
+            scaled_positions=atom_positions,
+            pbc=True,
+        )
+    elif atom_pos_type == "Cartesian":
+        atoms = Atoms(
+            symbols=atom_symbol,
+            cell=atom_lattice,
+            positions=atom_positions * atom_lattice_scale * Bohr,
+            pbc=True,
+        )
+
+    # atom_mass = np.array(atom_mass).flatten()
+    # if atom_mass.any():
+    #     atoms.set_masses(atom_mass)
+    if v_index:
+        atoms.set_velocities(atom_vel * UNIT_V)
+
+    atoms.set_initial_magnetic_moments(atom_magnetism)
+    atoms.set_constraint(fix_cart)
+
+    if verbose:
+        atoms.info["pp"] = atom_potential
+        atoms.info["basis"] = atom_basis
+        atoms.info["offsite_basis"] = atom_offsite_basis
+        atoms.info["descriptor"] = atom_descriptor
+
+    return atoms
+
+
+def get_lattice_from_latname(lines, latname=None):
+    from math import sqrt
+
+    if lines:
+        lines = lines.group(1).split(" ")
+
+    if latname == "sc":
+        return np.eye(3)
+    elif latname == "fcc":
+        return np.array([[-0.5, 0, 0.5], [0, 0.5, 0.5], [-0.5, 0.5, 0]])
+    elif latname == "bcc":
+        return np.array([[0.5, 0.5, 0.5], [-0.5, 0.5, 0.5], [-0.5, -0.5, 0.5]])
+    elif latname == "hexagonal":
+        x = float(lines[0])
+        return np.array([[1.0, 0, 0], [-0.5, sqrt(3) / 2, 0], [0, 0, x]])
+    elif latname == "trigonal":
+        x = float(lines[0])
+        tx = sqrt((1 - x) / 2)
+        ty = sqrt((1 - x) / 6)
+        tz = sqrt((1 + 2 * x) / 3)
+        return np.array([[tx, -ty, tz], [0, 2 * ty, tz], [-tx, -ty, tz]])
+    elif latname == "st":
+        x = float(lines[0])
+        return np.array([[1.0, 0, 0], [0, 1, 0], [0, 0, x]])
+    elif latname == "bct":
+        x = float(lines[0])
+        return np.array([[0.5, -0.5, x], [0.5, 0.5, x], [0.5, 0.5, x]])
+    elif latname == "baco":
+        x, y = list(map(float, lines))
+        return np.array([[0.5, x / 2, 0], [-0.5, x / 2, 0], [0, 0, y]])
+    elif latname == "fco":
+        x, y = list(map(float, lines))
+        return np.array([[0.5, 0, y / 2], [0.5, x / 2, 0], [0.5, x / 2, 0]])
+    elif latname == "bco":
+        x, y = list(map(float, lines))
+        return np.array(
+            [[0.5, x / 2, y / 2], [-0.5, x / 2, y / 2], [-0.5, -x / 2, y / 2]]
+        )
+    elif latname == "bco":
+        x, y, z = list(map(float, lines))
+        return np.array([[1, 0, 0], [x * z, x * sqrt(1 - z**2), 0], [0, 0, y]])
+    elif latname == "bacm":
+        x, y, z = list(map(float, lines))
+        return np.array(
+            [[0.5, 0, -y / 2], [x * z, x * sqrt(1 - z**2), 0], [0.5, 0, y / 2]]
+        )
+    elif latname == "triclinic":
+        x, y, m, n, l = list(map(float, lines))
+        fac = sqrt(1 + 2 * m * n * l - m**2 - n**2 - l**2) / sqrt(1 - m**2)
+        return np.array(
+            [
+                [1, 0, 0],
+                [x * m, x * sqrt(1 - m**2), 0],
+                [y * n, y * (l - n * m / sqrt(1 - m**2)), y * fac],
+            ]
+        )
+
+
+# READ ABACUS STRU -END-
+
+
+# READ ABACUS OUT -START-
+class AbacusOutChunk:
+    """Base class for AbacusOutChunks"""
+
+    def __init__(self, contents):
+        """Constructor
+
+        Parameters
+        ----------
+        contents: str
+            The contents of the output file
+        """
+        self.contents = contents
+
+    def parse_scalar(self, pattern):
+        """Parse a scalar property from the chunk according to specific pattern
+
+        Parameters
+        ----------
+        pattern: str
+            The pattern used to parse
+
+        Returns
+        -------
+        float
+            The scalar value of the property
+        """
+        pattern_compile = re.compile(pattern)
+        res = pattern_compile.search(self.contents)
+        if res:
+            return float(res.group(1))
+        else:
+            return None
+
+    @lazyproperty
+    def coordinate_system(self):
+        """Parse coordinate system (Cartesian or Direct) from the output file"""
+        # for '|', it will match all the patterns which results in '' or None
+        class_pattern = re.compile(r"(DIRECT) COORDINATES|(CARTESIAN) COORDINATES")
+        coord_class = list(class_pattern.search(self.contents).groups())
+        _remove_empty(coord_class)
+
+        return coord_class[0]
+
+    @lazymethod
+    def _parse_site(self):
+        """Parse sites for all the structures in the output file"""
+        pos_pattern = re.compile(
+            rf"(CARTESIAN COORDINATES \( UNIT = {_re_float} Bohr \)\.+\n\s*atom\s*x\s*y\s*z\s*mag(\s*vx\s*vy\s*vz\s*|\s*)\n[\s\S]+?)\n\n|(DIRECT COORDINATES\n\s*atom\s*x\s*y\s*z\s*mag(\s*vx\s*vy\s*vz\s*|\s*)\n[\s\S]+?)\n\n"
+        )
+
+        return pos_pattern.findall(self.contents)
+
+
+class AbacusOutHeaderChunk(AbacusOutChunk):
+    """General information that the header of the running_*.log file contains"""
+
+    def __init__(self, contents):
+        """Constructor
+
+        Parameters
+        ----------
+        contents: str
+            The contents of the output file
+        """
+        super().__init__(contents)
+
+    @lazyproperty
+    def out_dir(self):
+        out_pattern = re.compile(r"global_out_dir\s*=\s*([\s\S]+?)/")
+        return out_pattern.search(self.contents).group(1)
+
+    @lazyproperty
+    def lattice_constant(self):
+        """The lattice constant from the header of the running_*.log"""
+        a0_pattern_str = rf"lattice constant \(Angstrom\)\s*=\s*({_re_float})"
+        return self.parse_scalar(a0_pattern_str)
+
+    @lazyproperty
+    def initial_cell(self):
+        """The initial cell from the header of the running_*.log file"""
+        cell_pattern = re.compile(
+            rf"Lattice vectors: \(Cartesian coordinate: in unit of a_0\)\n\s*({_re_float})\s*({_re_float})\s*({_re_float})\n\s*({_re_float})\s*({_re_float})\s*({_re_float})\n\s*({_re_float})\s*({_re_float})\s*({_re_float})\n"
+        )
+        lattice = np.reshape(cell_pattern.findall(self.contents)[0], (3, 3)).astype(
+            float
+        )
+
+        return lattice * self.lattice_constant
+
+    @lazyproperty
+    def initial_site(self):
+        def str_to_sites(val_in):
+            val = np.array(val_in)
+            labels = val[:, 0]
+            pos = val[:, 1:4].astype(float)
+            if val.shape[1] == 5:
+                mag = val[:, 4].astype(float)
+                vel = np.zeros((3,), dtype=float)
+            elif val.shape[1] == 8:
+                mag = val[:, 4].astype(float)
+                vel = val[:, 5:8].astype(float)
+            return labels, pos, mag, vel
+
+        def parse_block(pos_block):
+            data = list(pos_block)
+            _remove_empty(data)
+            site = list(map(list, site_pattern.findall(data[0])))
+            list(map(_remove_empty, site))
+            labels, pos, mag, vel = str_to_sites(site)
+            if self.coordinate_system == "CARTESIAN":
+                unit = float(unit_pattern.search(self.contents).group(1)) * Bohr
+                positions = pos * unit
+            elif self.coordinate_system == "DIRECT":
+                positions = pos
+            return labels, positions, mag, vel
+
+        site_pattern = re.compile(
+            rf"tau[cd]_([a-zA-Z]+)\d+\s+({_re_float})\s+({_re_float})\s+({_re_float})\s+({_re_float})\s+({_re_float})\s+({_re_float})\s+({_re_float})|tau[cd]_([a-zA-Z]+)\d+\s+({_re_float})\s+({_re_float})\s+({_re_float})\s+({_re_float})"
+        )
+        unit_pattern = re.compile(rf"UNIT = ({_re_float}) Bohr")
+
+        return parse_block(self._parse_site()[0])
+
+    @lazyproperty
+    def initial_atoms(self):
+        """Create an atoms object for the initial structure from the
+        header of the running_*.log file"""
+        labels, positions, mag, vel = self.initial_site
+        if self.coordinate_system == "CARTESIAN":
+            atoms = Atoms(
+                symbols=labels,
+                positions=positions,
+                cell=self.initial_cell,
+                pbc=True,
+                velocities=vel * UNIT_V,
+            )
+        elif self.coordinate_system == "DIRECT":
+            atoms = Atoms(
+                symbols=labels,
+                scaled_positions=positions,
+                cell=self.initial_cell,
+                pbc=True,
+                velocities=vel * UNIT_V,
+            )
+        atoms.set_initial_magnetic_moments(mag)
+
+        return atoms
+
+    @lazyproperty
+    def is_relaxation(self):
+        """Determine if the calculation is an atomic position optimization or not"""
+        return "RELAXATION" in self.contents
+
+    @lazyproperty
+    def is_nscf(self):
+        """Determine if the calculation is a NSCF calculation"""
+        return "NONSELF-CONSISTENT" in self.contents
+
+    @lazyproperty
+    def is_cell_relaxation(self):
+        """Determine if the calculation is an variable cell optimization or not"""
+        return "RELAX CELL" in self.contents
+
+    @lazyproperty
+    def is_md(self):
+        """Determine if calculation is a molecular dynamics calculation"""
+        return "STEP OF MOLECULAR DYNAMICS" in self.contents
+
+    @lazymethod
+    def _parse_k_points(self):
+        """Get the list of k-points used in the calculation"""
+
+        def str_to_kpoints(val_in):
+            lines = (
+                re.search(
+                    rf"KPOINTS\s*DIRECT_X\s*DIRECT_Y\s*DIRECT_Z\s*WEIGHT([\s\S]+?)DONE",
+                    val_in,
+                )
+                .group(1)
+                .strip()
+                .split("\n")
+            )
+            data = []
+            for line in lines:
+                data.append(line.strip().split()[1:5])
+            data = np.array(data, dtype=float)
+            kpoints = data[:, :3]
+            weights = data[:, 3]
+            return kpoints, weights
+
+        k_pattern = re.compile(
+            r"minimum distributed K point number\s*=\s*\d+([\s\S]+?DONE : INIT K-POINTS Time)"
+        )
+        sub_contents = k_pattern.search(self.contents).group(1)
+        k_points, k_point_weights = str_to_kpoints(sub_contents)
+
+        return k_points[: int(self.n_k_points)], k_point_weights[: int(self.n_k_points)]
+
+    @lazyproperty
+    def n_atoms(self):
+        """The number of atoms for the material"""
+        pattern_str = r"TOTAL ATOM NUMBER = (\d+)"
+
+        return int(self.parse_scalar(pattern_str))
+
+    @lazyproperty
+    def n_bands(self):
+        """The number of Kohn-Sham states for the chunk"""
+        pattern_str = r"NBANDS = (\d+)"
+
+        return int(self.parse_scalar(pattern_str))
+
+    @lazyproperty
+    def n_electrons(self):
+        """The number of valence electrons for the chunk"""
+        pattern_str = r"AUTOSET number of electrons:  = (\d+)"
+        res = self.parse_scalar(pattern_str)
+        if res:
+            return int(res)
+        else:
+            return None
+
+    @lazyproperty
+    def n_occupied_bands(self):
+        """The number of occupied Kohn-Sham states for the chunk"""
+        pattern_str = r"occupied bands = (\d+)"
+
+        return int(self.parse_scalar(pattern_str))
+
+    @lazyproperty
+    def n_spins(self):
+        """The number of spin channels for the chunk"""
+        pattern_str = r"nspin = (\d+)"
+
+        return 1 if int(self.parse_scalar(pattern_str)) in [1, 4] else 2
+
+    @lazyproperty
+    def n_k_points(self):
+        """The number of spin channels for the chunk"""
+        nks = (
+            self.parse_scalar(r"nkstot_ibz = (\d+)")
+            if self.parse_scalar(r"nkstot_ibz = (\d+)")
+            else self.parse_scalar(r"nkstot = (\d+)")
+        )
+
+        return int(nks)
+
+    @lazyproperty
+    def k_points(self):
+        """All k-points listed in the calculation"""
+        return self._parse_k_points()[0]
+
+    @lazyproperty
+    def k_point_weights(self):
+        """The k-point weights for the calculation"""
+        return self._parse_k_points()[1]
+
+    @lazyproperty
+    def header_summary(self):
+        """Dictionary summarizing the information inside the header"""
+        return {
+            "lattice_constant": self.lattice_constant,
+            "initial_atoms": self.initial_atoms,
+            "initial_cell": self.initial_cell,
+            "is_nscf": self.is_nscf,
+            "is_relaxation": self.is_relaxation,
+            "is_cell_relaxation": self.is_cell_relaxation,
+            "is_md": self.is_md,
+            "n_atoms": self.n_atoms,
+            "n_bands": self.n_bands,
+            "n_occupied_bands": self.n_occupied_bands,
+            "n_spins": self.n_spins,
+            "n_k_points": self.n_k_points,
+            "k_points": self.k_points,
+            "k_point_weights": self.k_point_weights,
+            "out_dir": self.out_dir,
+        }
+
+
+class AbacusOutCalcChunk(AbacusOutChunk):
+    """A part of the running_*.log file correponding to a single calculated structure"""
+
+    def __init__(self, contents, header, index=-1):
+        """Constructor
+
+        Parameters
+        ----------
+        lines: str
+            The contents of the output file
+        header: dict
+            A summary of the relevant information from the running_*.log header
+        index: slice or int
+            index of image. `index = 0` is the first calculated image rather initial image
+        """
+        super().__init__(contents)
+        self._header = header.header_summary
+        self.index = index
+
+    @lazymethod
+    def _parse_cells(self):
+        """Parse all the cells from the output file"""
+        if self._header["is_relaxation"]:
+            return [self.initial_cell for i in range(self.ion_steps)]
+        elif self._header["is_cell_relaxation"]:
+            cell_pattern = re.compile(
+                rf"Lattice vectors: \(Cartesian coordinate: in unit of a_0\)\n\s*({_re_float})\s*({_re_float})\s*({_re_float})\n\s*({_re_float})\s*({_re_float})\s*({_re_float})\n\s*({_re_float})\s*({_re_float})\s*({_re_float})\n"
+            )
+            _lattice = np.reshape(
+                cell_pattern.findall(self.contents), (-1, 3, 3)
+            ).astype(float)
+            if self.ion_steps and _lattice.shape[0] != self.ion_steps:
+                lattice = np.zeros((self.ion_steps, 3, 3), dtype=float)
+                _indices = np.where(self._parse_relaxation_convergency())[0]
+                for i in range(len(_indices)):
+                    if i == 0:
+                        lattice[: _indices[i] + 1] = self.initial_cell
+                    else:
+                        lattice[_indices[i - 1] + 1 : _indices[i] + 1] = _lattice[i - 1]
+            return lattice * self._header["lattice_constant"]
+        else:
+            return self.initial_cell
+
+    @lazyproperty
+    def ion_steps(self):
+        "The number of ion steps"
+        return len(self._parse_ionic_block())
+
+    @lazymethod
+    def _parse_forces(self):
+        """Parse all the forces from the output file"""
+        force_pattern = re.compile(
+            r"TOTAL\-FORCE\s*\(eV/Angstrom\)\n\n.*\s*atom\s*x\s*y\s*z\n([\s\S]+?)\n\n"
+        )
+        forces = force_pattern.findall(self.contents)
+        if not forces:
+            force_pattern = re.compile(
+                r"TOTAL\-FORCE\s*\(eV/Angstrom\)\s*[\-]{2,}\n([\s\S]+?)\n[\-]{2,}"
+            )
+
+        return force_pattern.findall(self.contents)
+
+    @lazymethod
+    def _parse_stress(self):
+        """Parse the stress from the output file"""
+        stress_pattern = re.compile(
+            rf"(?:TOTAL\-|MD\s*)STRESS\s*\(KBAR\)\n\n.*\n\n\s*({_re_float})\s*({_re_float})\s*({_re_float})\n\s*({_re_float})\s*({_re_float})\s*({_re_float})\n\s*({_re_float})\s*({_re_float})\s*({_re_float})\n"
+        )
+        stresses = stress_pattern.findall(self.contents)
+        if not stresses:
+            stress_pattern = re.compile(
+                r"(?:TOTAL\-|MD\s*)STRESS\s*\(KBAR\)\s*[\-]{2,}\n"
+                + rf"\s*({_re_float})\s*({_re_float})\s*({_re_float})\n\s*({_re_float})\s*({_re_float})\s*({_re_float})\n\s*({_re_float})\s*({_re_float})\s*({_re_float})\n"
+            )
+
+        return stress_pattern.findall(self.contents)
+
+    @lazymethod
+    def _parse_eigenvalues(self):
+        """Parse the eigenvalues and occupations of the system."""
+        scf_eig_pattern = re.compile(
+            r"(STATE ENERGY\(eV\) AND OCCUPATIONS\s*NSPIN\s*==\s*\d+[\s\S]+?(?:\n\n\s*EFERMI|\n\n\n))"
+        )
+        scf_eig_all = scf_eig_pattern.findall(self.contents)
+
+        nscf_eig_pattern = re.compile(
+            r"(band eigenvalue in this processor \(eV\)\s*:\n[\s\S]+?\n\n\n)"
+        )
+        nscf_eig_all = nscf_eig_pattern.findall(self.contents)
+
+        return {"scf": scf_eig_all, "nscf": nscf_eig_all}
+
+    @lazymethod
+    def _parse_energy(self):
+        """Parse the energy from the output file."""
+        _out_dir = self._header["out_dir"].strip("/")
+
+        energy_pattern = re.compile(
+            rf"{_out_dir}\/\s*final etot is\s*({_re_float})\s*eV"
+        )
+        res = energy_pattern.findall(self.contents)
+        if res:
+            return res
+        else:
+            energy_pattern = re.compile(rf"\s*final etot is\s*({_re_float})\s*eV")
+            res = energy_pattern.findall(self.contents)
+            return res
+
+        # energy_pattern = re.compile(
+        #     rf'{_out_dir}\/\s*final etot is\s*({_re_float})\s*eV') if 'RELAXATION' in self.contents or 'RELAX CELL' in self.contents else re.compile(rf'\s*final etot is\s*({_re_float})\s*eV')
+
+        # return energy_pattern.findall(self.contents)
+
+    @lazymethod
+    def _parse_efermi(self):
+        """Parse the Fermi energy from the output file."""
+        fermi_pattern = re.compile(rf"EFERMI\s*=\s*({_re_float})\s*eV")
+
+        return fermi_pattern.findall(self.contents)
+
+    @lazymethod
+    def _parse_ionic_block(self):
+        """Parse the ionic block from the output file"""
+        step_pattern = re.compile(
+            rf"(?:[NON]*SELF-|STEP OF|RELAX CELL)([\s\S]+?)charge density convergence is achieved"
+        )
+
+        return step_pattern.findall(self.contents)
+
+    @lazymethod
+    def _parse_relaxation_convergency(self):
+        """Parse the convergency of atomic position optimization from the output file"""
+        if "Ion relaxation" in self.contents:
+            pattern = re.compile(
+                r"Ion relaxation is converged!|Ion relaxation is not converged yet"
+            )
+            return (
+                np.array(pattern.findall(self.contents))
+                == "Ion relaxation is converged!"
+            )
+        else:
+            pattern = re.compile(
+                r"Relaxation is converged!|Relaxation is not converged yet!"
+            )
+            return (
+                np.array(pattern.findall(self.contents)) == "Relaxation is converged!"
+            )
+
+    @lazymethod
+    def _parse_cell_relaxation_convergency(self):
+        """Parse the convergency of variable cell optimization from the output file"""
+        pattern = re.compile(
+            r"Lattice relaxation is converged!|Lattice relaxation is not converged yet"
+        )
+        lat_arr = (
+            np.array(pattern.findall(self.contents))
+            == "Lattice relaxation is converged!"
+        )
+        res = np.zeros((self.ion_steps), dtype=bool)
+        if lat_arr[-1] == True:
+            res[-1] = 1
+
+        return res.astype(bool)
+
+    @lazymethod
+    def _parse_md(self):
+        """Parse the molecular dynamics information from the output file"""
+        md_pattern = re.compile(
+            rf"Energy\s*\(Ry\)\s*Potential\s*\(Ry\)\s*Kinetic\s*\(Ry\)\s*Temperature\s*\(K\)\s*(?:Pressure\s*\(kbar\)\s*\n|\n)\s*({_re_float})\s*({_re_float})"
+        )
+        result = md_pattern.findall(self.contents)
+        self.md_e_unit = Rydberg
+        if result:
+            return result
+        else:
+            md_pattern = re.compile(
+                rf"Energy\s*Potential\s*Kinetic\s*Temperature\s*(?:Pressure \(KBAR\)\s*\n|\n)\s*({_re_float})\s*({_re_float})"
+            )
+            self.md_e_unit = Hartree
+            return md_pattern.findall(self.contents)
+
+    @lazymethod
+    def get_site(self):
+        """Get site from the output file according to index"""
+
+        def str_to_sites(val_in):
+            val = np.array(val_in)
+            labels = val[:, 0]
+            pos = val[:, 1:4].astype(float)
+            if val.shape[1] == 5:
+                mag = val[:, 4].astype(float)
+                vel = np.zeros((3,), dtype=float)
+            elif val.shape[1] == 8:
+                mag = val[:, 4].astype(float)
+                vel = val[:, 5:8].astype(float)
+            return labels, pos, mag, vel
+
+        def parse_block(pos_block):
+            data = list(pos_block)
+            _remove_empty(data)
+            site = list(map(list, site_pattern.findall(data[0])))
+            list(map(_remove_empty, site))
+            labels, pos, mag, vel = str_to_sites(site)
+            if self.coordinate_system == "CARTESIAN":
+                unit = float(unit_pattern.search(self.contents).group(1)) * Bohr
+                positions = pos * unit
+            elif self.coordinate_system == "DIRECT":
+                positions = pos
+            return labels, positions, mag, vel
+
+        site_pattern = re.compile(
+            rf"tau[cd]_([a-zA-Z]+)\d+\s+({_re_float})\s+({_re_float})\s+({_re_float})\s+({_re_float})\s+({_re_float})\s+({_re_float})\s+({_re_float})|tau[cd]_([a-zA-Z]+)\d+\s+({_re_float})\s+({_re_float})\s+({_re_float})\s+({_re_float})"
+        )
+        unit_pattern = re.compile(rf"UNIT = ({_re_float}) Bohr")
+
+        sites = self._parse_site()
+        if self.get_relaxation_convergency():
+            sites.append(sites[-1])
+        return parse_block(sites[self.index])
+
+    @lazymethod
+    def get_forces(self):
+        """Get forces from the output file according to index"""
+
+        def str_to_force(val_in):
+            data = []
+            val = [v.strip().split() for v in val_in.split("\n")]
+            for v in val:
+                data.append(np.array(v[1:], dtype=float))
+            return np.array(data)
+
+        try:
+            forces = self._parse_forces()[self.index]
+            return str_to_force(forces)
+        except IndexError:
+            return
+
+    @lazymethod
+    def get_forces_sort(self):
+        """Get forces from the output file according to index"""
+
+        def str_to_force(val_in):
+            data = []
+            val = [v.strip().split() for v in val_in.split("\n")]
+            for v in val:
+                data.append(np.array(v[1:], dtype=float))
+            return np.array(data)
+
+        try:
+            forces = self._parse_forces()[self.index]
+            if Path("ase_sort.dat").exists():
+                atoms_sort = np.loadtxt("ase_sort.dat", dtype=int)
+                return str_to_force(forces)[np.argsort(atoms_sort)]
+            else:
+                return str_to_force(forces)
+        except IndexError:
+            return
+
+    @lazymethod
+    def get_stress(self):
+        """Get the stress from the output file according to index"""
+        from ase.stress import full_3x3_to_voigt_6_stress
+
+        try:
+            stress = (
+                -0.1
+                * GPa
+                * np.array(self._parse_stress()[self.index])
+                .reshape((3, 3))
+                .astype(float)
+            )
+            return full_3x3_to_voigt_6_stress(stress)
+        except IndexError:
+            return
+
+    @lazymethod
+    def get_eigenvalues(self):
+        """Get the eigenvalues and occupations of the system according to index."""
+
+        # SCF
+        def str_to_energy_occupation(val_in):
+            def extract_data(val):
+                def func(i):
+                    res = np.array(
+                        list(
+                            map(
+                                lambda x: x.strip().split(),
+                                re.search(
+                                    rf"{i+1}/{nks} kpoint \(Cartesian\)\s*=.*\n([\s\S]+?)\n\n",
+                                    val,
+                                )
+                                .group(1)
+                                .split("\n"),
+                            )
+                        ),
+                        dtype=float,
+                    )
+                    return res[:, 1].astype(float), res[:, 2].astype(float)
+
+                return np.asarray(list(map(func, [i for i in range(nks)])))
+
+            nspin = int(
+                re.search(
+                    r"STATE ENERGY\(eV\) AND OCCUPATIONS\s*NSPIN\s*==\s*(\d+)", val_in
+                ).group(1)
+            )
+            nks = int(re.search(r"\d+/(\d+) kpoint \(Cartesian\)", val_in).group(1))
+            eigenvalues = np.full(
+                (
+                    self._header["n_spins"],
+                    self._header["n_k_points"],
+                    self._header["n_bands"],
+                ),
+                np.nan,
+            )
+            occupations = np.full(
+                (
+                    self._header["n_spins"],
+                    self._header["n_k_points"],
+                    self._header["n_bands"],
+                ),
+                np.nan,
+            )
+            if nspin in [1, 4]:
+                energies, occs = (
+                    extract_data(val_in)[:, 0, :],
+                    extract_data(val_in)[:, 1, :],
+                )
+                eigenvalues[0] = energies
+                occupations[0] = occs
+            elif nspin == 2:
+                val_up = re.search(r"SPIN UP :([\s\S]+?)\n\nSPIN", val_in).group()
+                energies, occs = (
+                    extract_data(val_up)[:, 0, :],
+                    extract_data(val_up)[:, 1, :],
+                )
+                eigenvalues[0] = energies
+                occupations[0] = occs
+
+                val_dw = re.search(
+                    r"SPIN DOWN :([\s\S]+?)(?:\n\n\s*EFERMI|\n\n\n)", val_in
+                ).group()
+                energies, occs = (
+                    extract_data(val_dw)[:, 0, :],
+                    extract_data(val_dw)[:, 1, :],
+                )
+                eigenvalues[1] = energies
+                occupations[1] = occs
+            return eigenvalues, occupations
+
+        # NSCF
+        def str_to_bandstructure(val_in):
+            def extract_data(val):
+                def func(i):
+                    res = np.array(
+                        list(
+                            map(
+                                lambda x: x.strip().split(),
+                                re.search(
+                                    rf"k\-points{i+1}\(\d+\):.*\n([\s\S]+?)\n\n", val
+                                )
+                                .group(1)
+                                .split("\n"),
+                            )
+                        )
+                    )
+                    return res[:, 2].astype(float), res[:, 3].astype(float)
+
+                return np.asarray(list(map(func, [i for i in range(nks)])))
+
+            nks = int(re.search(r"k\-points\d+\((\d+)\)", val_in).group(1))
+            eigenvalues = np.full(
+                (
+                    self._header["n_spins"],
+                    self._header["n_k_points"],
+                    self._header["n_bands"],
+                ),
+                np.nan,
+            )
+            occupations = np.full(
+                (
+                    self._header["n_spins"],
+                    self._header["n_k_points"],
+                    self._header["n_bands"],
+                ),
+                np.nan,
+            )
+            if re.search("spin up", val_in) and re.search("spin down", val_in):
+                val = re.search(r"spin up :\n([\s\S]+?)\n\n\n", val_in).group()
+                energies, occs = (
+                    extract_data(val)[:, 0, :],
+                    extract_data(val_in)[:, 1, :],
+                )
+                eigenvalues[0] = energies[: int(nks / 2)]
+                eigenvalues[1] = energies[int(nks / 2) :]
+                occupations[0] = occs[: int(nks / 2)]
+                occupations[1] = occs[int(nks / 2) :]
+            else:
+                energies, occs = (
+                    extract_data(val_in)[:, 0, :],
+                    extract_data(val_in)[:, 1, :],
+                )
+                eigenvalues[0] = energies
+                occupations[0] = occs
+            return eigenvalues, occupations
+
+        try:
+            return str_to_energy_occupation(
+                self._parse_eigenvalues()["scf"][self.index]
+            )
+        except IndexError:
+            try:
+                return str_to_bandstructure(
+                    self._parse_eigenvalues()["nscf"][self.index]
+                )
+            except IndexError:
+                return np.full(
+                    (
+                        self._header["n_spins"],
+                        self._header["n_k_points"],
+                        self._header["n_bands"],
+                    ),
+                    np.nan,
+                ), np.full(
+                    (
+                        self._header["n_spins"],
+                        self._header["n_k_points"],
+                        self._header["n_bands"],
+                    ),
+                    np.nan,
+                )
+
+    @lazymethod
+    def get_energy(self):
+        """Get the energy from the output file according to index."""
+        try:
+            return float(self._parse_energy()[self.index])
+        except IndexError:
+            return None
+
+    @lazymethod
+    def get_efermi(self):
+        """Get the Fermi energy from the output file according to index."""
+        try:
+            return float(self._parse_efermi()[self.index])
+        except IndexError:
+            return
+
+    @lazymethod
+    def get_relaxation_convergency(self):
+        """Get the convergency of atomic position optimization from the output file"""
+        return self._parse_relaxation_convergency()[self.index]
+
+    @lazymethod
+    def get_cell_relaxation_convergency(self):
+        """Get the convergency of variable cell optimization from the output file"""
+        return self._parse_cell_relaxation_convergency()[self.index]
+
+    @lazymethod
+    def get_md_energy(self):
+        """Get the total energy of each md step"""
+
+        try:
+            return float(self._parse_md()[self.index][0]) * self.md_e_unit
+        except IndexError:
+            return
+
+    @lazymethod
+    def get_md_potential(self):
+        """Get the potential energy of each md step"""
+
+        try:
+            return float(self._parse_md()[self.index][1]) * self.md_e_unit
+        except IndexError:
+            return
+
+    @lazymethod
+    def get_md_steps(self):
+        """Get steps of molecular dynamics"""
+        step_pattern = re.compile(r"STEP OF MOLECULAR DYNAMICS\s*:\s*(\d+)")
+
+        return list(map(int, step_pattern.findall(self.contents)))
+
+    @lazymethod
+    def get_dipole(self):
+        """Get electrical dipole"""
+
+        if self._header["is_md"]:
+            data = np.zeros((len(self.get_md_steps()), 3), dtype=float)
+            out_dir = Path(self._header["out_dir"])
+            data_files = [
+                out_dir / f"SPIN{i+1}_DIPOLE" for i in range(self._header["n_spins"])
+            ]
+            for file in data_files:
+                if file.exists():
+                    data = data + np.loadtxt(file, float)[:, 1:4]
+
+            return data[self.index]
+        else:
+            return
+
+    @lazyproperty
+    def forces(self):
+        """The forces for the chunk"""
+        return self.get_forces()
+
+    @lazyproperty
+    def forces_sort(self):
+        """The forces for the chunk"""
+        return self.get_forces_sort()
+
+    @lazyproperty
+    def stress(self):
+        """The stress for the chunk"""
+        return self.get_stress()
+
+    @lazyproperty
+    def dipole(self):
+        """The dipole for the chunk"""
+        return self.get_dipole()
+
+    @lazyproperty
+    def energy(self):
+        """The energy for the chunk"""
+        if self._header["is_md"]:
+            return self.get_md_potential()
+        else:
+            return self.get_energy()
+
+    @lazyproperty
+    def free_energy(self):
+        """The free energy for the chunk"""
+        if self._header["is_md"]:
+            return self.get_md_energy()
+        else:
+            return self.get_energy()
+
+    @lazyproperty
+    def eigenvalues(self):
+        """The eigenvalues for the chunk"""
+        return self.get_eigenvalues()[0]
+
+    @lazyproperty
+    def occupations(self):
+        """The occupations for the chunk"""
+        return self.get_eigenvalues()[1]
+
+    @lazyproperty
+    def kpts(self):
+        """The SinglePointKPoint objects for the chunk"""
+        return arrays_to_kpoints(
+            self.eigenvalues, self.occupations, self._header["k_point_weights"]
+        )
+
+    @lazyproperty
+    def E_f(self):
+        """The Fermi energy for the chunk"""
+        return self.get_efermi()
+
+    @lazyproperty
+    def _ionic_block(self):
+        """The ionic block for the chunk"""
+
+        return self._parse_ionic_block()[self.index]
+
+    @lazyproperty
+    def magmom(self):
+        """The Fermi energy for the chunk"""
+        magmom_pattern = re.compile(
+            rf"total magnetism \(Bohr mag/cell\)\s*=\s*({_re_float})"
+        )
+
+        try:
+            return float(magmom_pattern.findall(self._ionic_block)[-1])
+        except IndexError:
+            return
+
+    @lazyproperty
+    def n_iter(self):
+        """The number of SCF iterations needed to converge the SCF cycle for the chunk"""
+        step_pattern = re.compile(rf"ELEC\s*=\s*[+]?(\d+)")
+
+        try:
+            return int(step_pattern.findall(self._ionic_block)[-1])
+        except IndexError:
+            return
+
+    @lazyproperty
+    def converged(self):
+        """True if the chunk is a fully converged final structure"""
+        if self._header["is_cell_relaxation"]:
+            return self.get_cell_relaxation_convergency()
+        elif self._header["is_relaxation"]:
+            return self.get_relaxation_convergency()
+        elif self._header["is_nscf"]:
+            return "Total  Time" in self.contents
+        else:
+            return "charge density convergence is achieved" in self.contents
+
+    @lazyproperty
+    def initial_atoms(self):
+        """The initial structure defined in the running_*.log file"""
+        return self._header["initial_atoms"]
+
+    @lazyproperty
+    def initial_cell(self):
+        """The initial lattice vectors defined in the running_*.log file"""
+        return self._header["initial_cell"]
+
+    @lazyproperty
+    def n_atoms(self):
+        """The number of atoms for the material"""
+        return self._header["n_atoms"]
+
+    @lazyproperty
+    def n_bands(self):
+        """The number of Kohn-Sham states for the chunk"""
+        return self._header["n_bands"]
+
+    @lazyproperty
+    def n_occupied_bands(self):
+        """The number of occupied Kohn-Sham states for the chunk"""
+        return self._header["n_occupied_bands"]
+
+    @lazyproperty
+    def n_spins(self):
+        """The number of spin channels for the chunk"""
+        return self._header["n_spins"]
+
+    @lazyproperty
+    def n_k_points(self):
+        """The number of k_points for the chunk"""
+        return self._header["n_k_points"]
+
+    @lazyproperty
+    def k_points(self):
+        """k_points for the chunk"""
+        return self._header["k_points"]
+
+    @lazyproperty
+    def k_point_weights(self):
+        """k_point_weights for the chunk"""
+        return self._header["k_point_weights"]
+
+    @property
+    def results(self):
+        """Convert an AbacusOutChunk to a Results Dictionary"""
+        results = {
+            "energy": self.energy,
+            "free_energy": self.free_energy,
+            "forces": self.forces_sort,
+            "stress": self.stress,
+            "magmom": self.magmom,
+            "fermi_level": self.E_f,
+            "n_iter": self.n_iter,
+            "eigenvalues": self.eigenvalues,
+            "occupations": self.occupations,
+            "ibz_kpoints": self.k_points,
+            "kpoint_weights": self.k_point_weights,
+            "dipole": self.dipole,
+        }
+
+        return {key: value for key, value in results.items() if value is not None}
+
+    @lazyproperty
+    def atoms(self):
+        """Convert AbacusOutChunk to Atoms object and add all non-standard outputs to atoms.info"""
+        """Create an atoms object for the subsequent structures
+        calculated in the output file"""
+        atoms = None
+        if self._header["is_md"]:
+            _stru_dir = Path(self._header["out_dir"]) / "STRU"
+            md_stru_dir = (
+                _stru_dir if _stru_dir.exists() else Path(self._header["out_dir"])
+            )
+            atoms = read_abacus(
+                open(md_stru_dir / f"STRU_MD_{self.get_md_steps()[self.index]}", "r")
+            )
+
+        # self._header['is_relaxation'] or self._header['is_cell_relaxation']:
+        elif self.ion_steps > 1:
+            labels, positions, mag, vel = self.get_site()
+            if self.coordinate_system == "CARTESIAN":
+                atoms = Atoms(
+                    symbols=labels,
+                    positions=positions,
+                    cell=self._parse_cells()[self.index],
+                    pbc=True,
+                    velocities=vel * UNIT_V,
+                )
+            elif self.coordinate_system == "DIRECT":
+                atoms = Atoms(
+                    symbols=labels,
+                    scaled_positions=positions,
+                    cell=self._parse_cells()[self.index],
+                    pbc=True,
+                    velocities=vel * UNIT_V,
+                )
+
+        else:
+            atoms = self.initial_atoms.copy()
+
+        calc = SinglePointDFTCalculator(
+            atoms,
+            energy=self.energy,
+            free_energy=self.free_energy,
+            efermi=self.E_f,
+            forces=self.forces,
+            stress=self.stress,
+            magmom=self.magmom,
+            dipole=self.dipole,
+            ibzkpts=self.k_points,
+            kpts=self.kpts,
+        )
+
+        calc.name = "Abacus"
+        atoms.calc = calc
+
+        return atoms
+
+
+def _slice2indices(s, n=None):
+    """Convert a slice object into indices"""
+    if isinstance(s, slice):
+        return range(*s.indices(n))
+    elif isinstance(s, int):
+        return [s]
+    elif isinstance(s, list):
+        return s
+    else:
+        raise ValueError(
+            "Indices must be scalar integer, list of integers, or slice object"
+        )
+
+
+@reader
+def _get_abacus_chunks(fd, index=-1, non_convergence_ok=False):
+    """Import ABACUS output files with all data available, i.e.
+    relaxations, MD information, force information ..."""
+    contents = fd.read()
+    header_pattern = re.compile(
+        r"READING GENERAL INFORMATION([\s\S]+?([NON]*SELF-CONSISTENT|STEP OF MOLECULAR DYNAMICS|STEP OF [ION]*\s*RELAXATION|RELAX CELL))"
+    )
+    header_chunk = AbacusOutHeaderChunk(header_pattern.search(contents).group(1))
+
+    time_pattern = re.compile(
+        r"Total\s*Time\s*:\s*[0-9]+\s*h\s*[0-9]+\s*mins\s*[0-9]+\s*secs"
+    )
+    time = time_pattern.findall(contents)[-1]
+
+    calc_pattern = re.compile(
+        rf"(([NON]*SELF-CONSISTENT|STEP OF MOLECULAR DYNAMICS|STEP OF [ION]*\s*RELAXATION|RELAX CELL)[\s\S]+?(?={time}))"
+    )
+    calc_contents = calc_pattern.search(contents).group(1)
+    final_chunk = AbacusOutCalcChunk(calc_contents, header_chunk, -1)
+
+    if not non_convergence_ok and not final_chunk.converged:
+        raise ValueError("The calculation did not complete successfully")
+
+    _steps = final_chunk.ion_steps if final_chunk.ion_steps else 1
+    indices = _slice2indices(index, _steps)
+
+    return [AbacusOutCalcChunk(calc_contents, header_chunk, i) for i in indices]
+
+
+@reader
+def read_abacus_out(fd, index=-1, non_convergence_ok=False):
+    """Import ABACUS output files with all data available, i.e.
+    relaxations, MD information, force information ..."""
+    chunks = _get_abacus_chunks(fd, index, non_convergence_ok)
+
+    return [chunk.atoms for chunk in chunks]
+
+
+@reader
+def read_abacus_results(fd, index=-1, non_convergence_ok=False):
+    """Import ABACUS output files and summarize all relevant information
+    into a dictionary"""
+    chunks = _get_abacus_chunks(fd, index, non_convergence_ok)
+
+    return [chunk.results for chunk in chunks]
+
+
+def _remove_empty(a: list):
+    """Remove '' and [] in `a`"""
+    while "" in a:
+        a.remove("")
+    while [] in a:
+        a.remove([])
+    while None in a:
+        a.remove(None)
+
+
+# READ ABACUS OUT -END-