pymatgen.io.cp2k.inputs module
This module defines the building blocks of a CP2K input file. The cp2k input structure is essentially a collection of “sections” which are similar to dictionary objects that activate modules of the cp2k executable, and then “keywords” which adjust variables inside of those modules. For example, FORCE_EVAL section will activate CP2K’s ability to calculate forces, and inside FORCE_EVAL, the Keyword “METHOD can be set to “QS” to set the method of force evaluation to be the quickstep (DFT) module.
A quick overview of the module:
- – Section class defines the basis of Cp2k input and contains methods for manipulating these
objects similarly to Dicts.
- – Keyword class defines the keywords used inside of Section objects that changes variables in
Cp2k programs.
- – SectionList and KeywordList classes are lists of Section and Keyword objects that have
the same dictionary key. This deals with repeated sections and keywords.
- – Cp2kInput class is special instantiation of Section that is used to represent the full cp2k
calculation input.
- – The rest of the classes are children of Section intended to make initialization of common
sections easier.
- class BrokenSymmetry(l_alpha: Sequence = (- 1,), n_alpha: Sequence = (0,), nel_alpha: Sequence = (- 1,), l_beta: Sequence = (- 1,), n_beta: Sequence = (0,), nel_beta: Sequence = (- 1,))[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Define the required atomic orbital occupation assigned in initialization of the density matrix, by adding or subtracting electrons from specific angular momentum channels. It works only with GUESS ATOMIC.
Initialize the broken symmetry section
- Parameters
l_alpha – Angular momentum quantum number of the orbitals whose occupation is changed
n_alpha – Principal quantum number of the orbitals whose occupation is changed. Default is the first not occupied
nel_alpha – Orbital occupation change per angular momentum quantum number. In unrestricted calculations applied to spin alpha
l_beta – Same as L_alpha for beta channel
n_beta – Same as N_alpha for beta channel
nel_beta – Same as NEL_alpha for beta channel
- class Cell(lattice: pymatgen.core.lattice.Lattice, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Defines the simulation cell (lattice)
Initialize the cell section.
- Parameters
lattice – pymatgen lattice object
- class Coord(structure: pymatgen.core.structure.Structure | pymatgen.core.structure.Molecule, aliases: Optional[dict] = None, subsections: Optional[dict] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Specifies the coordinates of the atoms using a pymatgen structure object.
- Parameters
structure – Pymatgen structure object
alias (bool) – whether or not to identify the sites by Element + number so you can do things like assign unique magnetization do different elements.
- class Cp2kInput(name: str = 'CP2K_INPUT', subsections: Optional[dict] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Special instance of ‘Section’ class that is meant to represent the overall cp2k input. Distinguishes itself from Section by overriding get_string() to not print this section’s title and by implementing the file i/o.
Initialize Cp2kInput by calling the super
- write_file(input_filename: str = 'cp2k.inp', output_dir: str = '.', make_dir_if_not_present: bool = True)[source]
Write input to a file.
- Parameters
input_filename (str, optional) – Defaults to “cp2k.inp”.
output_dir (str, optional) – Defaults to “.”.
make_dir_if_not_present (bool, optional) – Defaults to True.
- class Davidson(new_prec_each: int = 20, preconditioner: str = 'FULL_SINGLE_INVERSE', **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Parameters for davidson diagonalization
- Parameters
new_prec_each (int) – How often to recalculate the preconditioner.
preconditioner (str) –
Preconditioner to use. “FULL_ALL”: Most effective state selective preconditioner based on diagonalization,
requires the ENERGY_GAP parameter to be an underestimate of the HOMO-LUMO gap. This preconditioner is recommended for almost all systems, except very large systems where make_preconditioner would dominate the total computational cost.
- ”FULL_KINETIC”: Cholesky inversion of S and T, fast construction, robust, use for
very large systems.
- ”FULL_SINGLE”: Based on H-eS diagonalisation, not as good as FULL_ALL, but
somewhat cheaper to apply.
- ”FULL_SINGLE_INVERSE”: Based on H-eS cholesky inversion, similar to FULL_SINGLE
in preconditioning efficiency but cheaper to construct, might be somewhat less robust. Recommended for large systems.
- ”FULL_S_INVERSE”: Cholesky inversion of S, not as good as FULL_KINETIC,
yet equally expensive.
”NONE”: skip preconditioning
- class Dft(basis_set_filenames='BASIS_MOLOPT', potential_filename='GTH_POTENTIALS', uks: bool = True, wfn_restart_file_name: Optional[str] = None, subsections: Optional[dict] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls the DFT parameters in Cp2k
Initialize the DFT section.
- Parameters
basis_set_filenames (str, optional) – Name of the file that contains the basis set information. Defaults to “BASIS_MOLOPT”.
potential_filename (str, optional) – Name of the file that contains the pseudopotential information. Defaults to “GTH_POTENTIALS”.
uks (bool, optional) – Whether to run unrestricted Kohn Sham (spin polarized). Defaults to True.
wfn_restart_file_name (str, optional) – Defaults to None.
subsections (dict, optional) – Any subsections to initialize with. Defaults to None.
- class DftPlusU(eps_u_ramping=1e-05, init_u_ramping_each_scf=False, l=- 1, u_minus_j=0, u_ramping=0)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls DFT+U for an atom kind
Initialize the DftPlusU section.
- Parameters
eps_u_ramping – (float) SCF convergence threshold at which to start ramping the U value
init_u_ramping_each_scf – (bool) Whether or not to do u_ramping each scf cycle
l – (int) angular moment of the orbital to apply the +U correction
u_minus_j – (float) the effective U parameter, Ueff = U-J
u_ramping – (float) stepwise amount to increase during ramping until u_minus_j is reached
- class Diagonalization(eps_adapt: float = 0, eps_iter: float = 1e-08, eps_jacobi: float = 0, jacobi_threshold: float = 1e-07, subsections: Optional[dict] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls diagonalization settings (if using traditional diagonalization).
Initialize the diagronalization section
- class E_Density_Cube(keywords=None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls printing of the electron density cube file
Initialize the E_DENSITY_CUBE Section
- class ForceEval(subsections: Optional[dict] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls the calculation of energy and forces in Cp2k
Initialize the ForceEval section
- Parameters
subsections (dict, optional) – Defaults to None.
- class Global(project_name: str = 'CP2K', run_type: str = 'ENERGY_FORCE', **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls ‘global’ settings for cp2k execution such as RUN_TYPE and PROJECT_NAME
Initialize the global section
- Parameters
project_name (str, optional) – Defaults to “CP2K”.
run_type (str, optional) –
- class Keyword(name: str, *values, description: Optional[str] = None, units: Optional[str] = None, verbose: bool = True, repeats: bool = False)[source]
Bases:
monty.json.MSONable
Class representing a keyword argument in CP2K. Within CP2K Sections, which activate features of the CP2K code, the keywords are arguments that control the functionality of that feature. For example, the section “FORCE_EVAL” activates the evaluation of forces/energies, but within “FORCE_EVAL” the keyword “METHOD” controls whether or not this will be done with, say, “Quickstep” (DFT) or “EIP” (empirical interatomic potential).
Initializes a keyword. These Keywords and the value passed to them are sometimes as simple as KEYWORD VALUE, but can also be more elaborate such as KEYWORD [UNITS] VALUE1 VALUE2, which is why this class exists: to handle many values and control easy printing to an input file.
- Parameters
name (str) – The name of this keyword. Must match an acceptable keyword from CP2K
args – All non-keyword arguments after ‘name’ are interpreted as the values to set for this keyword. i.e: KEYWORD ARG1 ARG2 would provide two values to the keyword.
description (str) – The description for this keyword. This can make readability of input files easier for some. Default=None.
units (str) – The units for this keyword. If not specified, CP2K default units will be used. Consult manual for default units. Default=None.
repeats (bool) – Whether or not this keyword may be repeated. Default=False.
- class KeywordList(keywords: Sequence[pymatgen.io.cp2k.inputs.Keyword])[source]
Bases:
monty.json.MSONable
Some keywords can be repeated, which makes accessing them via the normal dictionary methods a little unnatural. This class deals with this by defining a collection of same-named keywords that are accessed by one name.
Initializes a keyword list given a sequence of keywords.
- Parameters
keywords – A list of keywords. Must all have the same name (case-insensitive)
- class Kind(specie: str, alias: Optional[str] = None, magnetization: float = 0.0, subsections: Optional[dict] = None, basis_set: str = 'GTH_BASIS', potential: str = 'GTH_POTENTIALS', ghost: bool = False, aux_basis: Optional[str] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Specifies the information for the different atom types being simulated.
Initialize a KIND section
- Parameters
alias (str) – Alias for the atom, can be used for specifying modifications to certain atoms but not all, e.g. Mg_1 and Mg_2 to force difference oxidation states on the two atoms.
magnetization (float) – From the CP2K Manual: The magnetization used in the atomic initial guess. Adds magnetization/2 spin-alpha electrons and removes magnetization/2 spin-beta electrons.
basis_set (str) – Basis set for this atom, accessible from the basis set file specified
potential (str) – Pseudopotential for this atom, accessible from the potential file
kwargs – Additional kwargs to pass to Section()
- class Kpoints(kpts: Union[Sequence, Sequence[Sequence[int]]], weights: Optional[Sequence] = None, eps_geo: float = 1e-06, full_grid: bool = False, parallel_group_size: int = - 1, scheme: str = 'MONKHORST-PACK', symmetry: bool = False, units: str = 'B_VECTOR', verbose: bool = False, wavefunctions: str = 'COMPLEX')[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Description of the k-points to use for the calculation.
- Parameters
kpts (list, tuple) – a 2D array for the kpoints of the form [(1,1,1),]. If len(kpts) == 1. Then it is taken as subdivisions for automatic kpoint scheme. If it has more entries, it is taken as manual entries for kpoints.
weights (list, tuple) – a weight for each kpoint. Default is to weigh each by 1
eps_geo (float) – tolerance for symmetry. Default=1e-6
full_grid (bool) – use full (not reduced) kpoint grid. Default=False.
parallel_group_size (int) – from cp2k manual: Number of processors to be used for a single kpoint. This number must divide the total number of processes. The number of groups must divide the total number of kpoints. Value=-1 (smallest possible number of processes per group, satisfying the constraints). Value=0 (all processes). Value=n (exactly n processes). Default=-1.
scheme (str) – kpoint generation scheme. Default=’Monkhorst-Pack’
symmetry (bool) – Use symmetry to reduce the number of kpoints. Default=False.
units (str) – Units for the kpoint coordinates (reciprocal coordinates or cartesian). Default=’B_VECTOR’ (reciprocal)
verbose (bool) – verbose output for kpoints. Default=False
wavefunctions (str) – Whether to use complex or real valued wavefunctions (if available). Default=’complex’
- classmethod from_kpoints(kpoints, structure=None, reduce=True)[source]
Initialize the section from a Kpoints object (pymatgen.io.vasp.inputs). CP2K does not have an automatic gamma-point constructor, so this is generally used to get the number of divisions from a kpoint static constructor and then build a Monkhorst-Pack grid, which is sufficient for gamma-recommended systems so long as the grid is fine enough.
- Parameters
kpoints – A pymatgen kpoints object.
structure – Pymatgen structure object. Required for automatically performing symmetry analysis and reducing the kpoint grid.
reduce – whether or not to reduce the grid using symmetry. CP2K itself cannot do this automatically without spglib present at execution time.
- class LDOS(index: int = 1, alias: Optional[str] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls printing of the LDOS (List-Density of states). i.e. projects onto specific atoms.
Initialize the LDOS section
- Parameters
index – Index of the atom to project onto
- class MO_Cubes(write_cube: bool = False, nhomo: int = 1, nlumo: int = 1, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls printing of the molecular orbital eigenvalues
Initialize the MO_CUBES section
- class Mgrid(cutoff: int | float = 1200, rel_cutoff: int | float = 80, ngrids: int = 5, progression_factor: int = 3, subsections: Optional[dict] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls the multigrid for numerical integration
Initialize the MGRID section
- Parameters
cutoff – Cutoff energy (in Rydbergs for historical reasons) defining how find of Gaussians will be used
rel_cutoff – The relative cutoff energy, which defines how to map the Gaussians onto the multigrid. If the value is too low then, even if you have a high cutoff with sharp Gaussians, they will be mapped to the course part of the multigrid
ngrids – number of grids to use
progression_factor – divisor that decides how to map Gaussians the multigrid after the highest mapping is decided by rel_cutoff
- class OrbitalTransformation(minimizer: str = 'CG', preconditioner: str = 'FULL_ALL', algorithm: str = 'STRICT', rotation: bool = False, occupation_preconditioner: bool = False, energy_gap: float = - 1, linesearch: str = '2PNT', subsections: Optional[dict] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Turns on the Orbital Transformation scheme for diagonalizing the Hamiltonian. Much faster and with guaranteed convergence compared to normal diagonalization, but requires the system to have a band gap.
NOTE: OT has poor convergence for metallic systems and cannot use SCF mixing or smearing. Therefore, you should not use it for metals or systems with ‘small’ band gaps. In that case, use normal diagonalization, which will be slower, but will converge properly.
Initialize the OT section
- Parameters
minimizer (str) – The minimizer to use with the OT method. Default is conjugate gradient method, which is more robust, but more well-behaved systems should use DIIS, which can be as much as 50% faster.
preconditioner (str) – Preconditioner to use for OT, FULL_ALL tends to be most robust, but is not always most efficient. For difficult systems, FULL_SINGLE_INVERSE can be more robust, and is reasonably efficient with large systems. For huge, but well behaved, systems, where construction of the preconditioner can take a very long time, FULL_KINETIC can be a good choice.
energy_gap (float) – Guess for the band gap. For FULL_ALL, should be smaller than the actual band gap, so simply using 0.01 is a robust value. Choosing a larger value will help if you start with a bad initial guess though. For FULL_SINGLE_INVERSE, energy_gap is treated as a lower bound. Values lower than 0.05 in this case can lead to stability issues.
algorithm (str) – What algorithm to use for OT. ‘Strict’: Taylor or diagonalization based algorithm. IRAC: Orbital Transformation based Iterative Refinement of the Approximative Congruence transformation (OT/IR).
linesearch (str) – From the manual: 1D line search algorithm to be used with the OT minimizer, in increasing order of robustness and cost. MINIMIZER CG combined with LINESEARCH GOLD should always find an electronic minimum. Whereas the 2PNT minimizer is almost always OK, 3PNT might be needed for systems in which successive OT CG steps do not decrease the total energy.
- class PBE(parameterization: str = 'ORIG', scale_c: float | int = 1, scale_x: float | int = 1)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Info about the PBE functional.
- Parameters
parameterization (str) – ORIG: original PBE PBESOL: PBE for solids/surfaces REVPBE: revised PBE
scale_c (float) – scales the correlation part of the functional.
scale_x (float) – scales the exchange part of the functional.
- class PDOS(nlumo: int = - 1, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls printing of projected density of states onto the different atom KINDS (elemental decomposed DOS).
Initialize the PDOS section
- Parameters
nlumo – how many unoccupied orbitals to include (-1==ALL)
- class QS(method: str = 'GPW', eps_default: float = 1e-10, extrapolation: str = 'ASPC', subsections: Optional[dict] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls the quickstep settings (DFT driver)
Initialize the QS Section
- Parameters
method ("GPW" | "GAPW") – What DFT methodology to use. GPW (Gaussian Plane Waves) for DFT with pseudopotentials or GAPW (Gaussian Augmented Plane Waves) for all electron calculations.
eps_default (float) – The default level of convergence accuracy. NOTE: This is a global value for all the numerical value of all EPS_* values in QS module. It is not the same as EPS_SCF, which sets convergence accuracy of the SCF cycle alone.
extrapolation ("PS" | "ASPC") – Method use for extrapolation. If using gamma-point-only calculation, then one should either PS or ASPC (ASPC especially for MD runs). See the manual for other options.
subsections (dict) – Subsections to initialize with.
- class Scf(max_scf: int = 50, eps_scf: float = 1e-06, scf_guess: str = 'RESTART', subsections: Optional[dict] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls the self consistent field loop
Initialize the Scf section
- Parameters
max_scf (int) – Maximum number of SCF loops before terminating. Defaults to 50.
eps_scf (float) – Convergence criteria for SCF loop. Defaults to 1e-6.
scf_guess –
Initial guess for SCF loop. “ATOMIC”: Generate an atomic density using the atomic code “CORE”: Diagonalize the core Hamiltonian for an initial guess. “HISTORY_RESTART”: Extrapolated from previous RESTART files. “MOPAC”: Use same guess as MOPAC for semi-empirical methods or a simple
diagonal density matrix for other methods.
”NONE”: Skip initial guess (only for NON-SCC DFTB). “RANDOM”: Use random wavefunction coefficients. “RESTART”: Use the RESTART file as an initial guess (and ATOMIC if not present). “SPARSE”: Generate a sparse wavefunction using the atomic code (for OT based
methods).
- class Section(name: str, subsections: Optional[dict] = None, repeats: bool = False, description: Optional[str] = None, keywords: Optional[dict] = None, section_parameters: Optional[Sequence[str]] = None, location: Optional[str] = None, verbose: bool = True, alias: Optional[str] = None, **kwargs)[source]
Bases:
monty.json.MSONable
Basic input representation of input to Cp2k. Activates functionality inside of the Cp2k executable.
Basic object representing a CP2K Section. Sections activate different parts of the calculation. For example, FORCE_EVAL section will activate CP2K’s ability to calculate forces.
- Parameters
name (str) – The name of the section (must match name in CP2K)
subsections (dict) – A dictionary of subsections that are nested in this section. Format is {‘NAME’: Section(args, **kwargs). The name you chose for ‘NAME’ to index that subsection does not *have to be the same as the section’s true name, but we recommend matching them. You can specify a blank dictionary if there are no subsections, or if you want to insert the subsections later.
repeats (bool) – Whether or not this section can be repeated. Most sections cannot. Default=False.
description (str) – Description of this section for easier readability
keywords (list) – the keywords to be set for this section. Each element should be a Keyword object. This can be more cumbersome than simply using kwargs for building a class in a script, but is more convenient for the class instantiations of CP2K sections (see below).
section_parameters (list) – the section parameters for this section. Section parameters are specialized keywords that modify the behavior of the section overall. Most sections do not have section parameters, but some do. Unlike normal Keywords, these are specified as strings and not as Keyword objects.
location (str) – the path to the section in the form ‘SECTION/SUBSECTION1/SUBSECTION3’, example for QS module: ‘FORCE_EVAL/DFT/QS’. This location is used to automatically determine if a subsection requires a supersection to be activated.
verbose (str) – Controls how much is printed to Cp2k input files (Also see Keyword). If True, then a description of the section will be printed with it as a comment (if description is set). Default=True.
keyword (kwargs are interpreted as) –
as (value pairs and added to the keywords array) –
objects (Keyword) –
- by_path(path: str)[source]
Access a sub-section using a path. Used by the file parser.
- Parameters
path (str) – Path to section of form ‘SUBSECTION1/SUBSECTION2/SUBSECTION_OF_INTEREST’
- check(path: str)[source]
Check if section exists within the current using a path. Can be useful for cross-checking whether or not required dependencies have been satisfied, which CP2K does not enforce.
- Parameters
path (str) – Path to section of form ‘SUBSECTION1/SUBSECTION2/SUBSECTION_OF_INTEREST’
- get(d, default=None)[source]
Similar to get for dictionaries. This will attempt to retrieve the section or keyword matching d. Will not raise an error if d does not exist.
- Parameters
d – the key to retrieve, if present
default – what to return if d is not found
- get_keyword(d, default=None)[source]
Get function, only for subsections
- Parameters
d – Name of keyword to get
default – return if d is not found in keyword list
- get_section(d, default=None)[source]
Get function, only for subsections
- Parameters
d – Name of section to get
default – return if d is not found in subsections
- update(d: dict, strict=False)[source]
Update the Section according to a dictionary argument. This is most useful for providing user-override settings to default parameters. As you pass a dictionary the class variables like “description”, “location”, or “repeats” are not included. Therefore, it is recommended that this be used to modify existing Section objects to a user’s needs, but not used for the creation of new Section child-classes.
- Parameters
d (dict) –
A dictionary containing the update information. Should use nested dictionaries to specify the full path of the update. If a section or keyword does not exist, it will be created, but only with the values that are provided in “d”, not using default values from a Section object. Example: {
- ’SUBSECTION1’: {
‘SUBSEC2’: {‘NEW_KEYWORD’: ‘NEW_VAL’}, ‘NEW_SUBSEC’: {‘NEW_KWD’: ‘NEW_VAL’} }
}
strict (bool) – If true, only update existing sections and keywords. If false, allow new sections and keywords. Default: False
- class SectionList(sections: Sequence[pymatgen.io.cp2k.inputs.Section])[source]
Bases:
monty.json.MSONable
Section list
Initializes a SectionList object using a sequence of sections.
- Parameters
sections – A list of keywords. Must all have the same name (case-insensitive)
- class Smear(elec_temp: int | float = 300, method: str = 'FERMI_DIRAC', fixed_magnetic_moment: float = - 100.0, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Control electron smearing
Initialize the SMEAR section
- class Subsys(subsections: Optional[dict] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls the definition of the system to be simulated
Initialize the subsys section
- class V_Hartree_Cube(keywords=None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Controls printing of the hartree potential as a cube file.
Initialize the V_HARTREE_CUBE section
- class Xc_Functional(functionals: Optional[Iterable] = None, subsections: Optional[dict] = None, **kwargs)[source]
Bases:
pymatgen.io.cp2k.inputs.Section
Defines the XC functional(s) to use.
Initialize the XC_FUNCTIONAL class