vayesta.solver

Subpackages

• vayesta.solver.eb_fci
  • Submodules
  • vayesta.solver.eb_fci.ebfci
    • REBFCI
      • REBFCI.solver
      • REBFCI.Options
      • REBFCI.norb
      • REBFCI.nelec
      • REBFCI.nbos
      • REBFCI.kernel()
      • REBFCI.get_hamil()
      • REBFCI.make_rdm1()
      • REBFCI.make_rdm2()
      • REBFCI.make_rdm12()
      • REBFCI.make_rdm_eb()
      • REBFCI.make_dd_moms()
    • UEBFCI
      • UEBFCI.solver
      • UEBFCI.make_rdm12()
      • UEBFCI.Options
      • UEBFCI.get_hamil()
      • UEBFCI.kernel()
      • UEBFCI.make_dd_moms()
      • UEBFCI.make_rdm1()
      • UEBFCI.make_rdm2()
      • UEBFCI.make_rdm_eb()
      • UEBFCI.nbos
      • UEBFCI.nelec
      • UEBFCI.norb
  • vayesta.solver.eb_fci.ebfci_slow
    • contract_all()
    • make_shape()
    • contract_1e()
    • contract_2e()
    • contract_ep()
    • contract_pp()
    • apply_bos_annihilation()
    • apply_bos_creation()
    • contract_pp_for_future()
    • slices_for()
    • slices_for_cre()
    • slices_for_des()
    • slices_for_occ_reduction()
    • make_hdiag()
    • kernel()
    • kernel_multiroot()
    • make_rdm1()
    • make_rdm12()
    • make_rdm12s()
    • make_eb_rdm()
    • calc_dd_resp_mom()
    • run()
    • run_hub_test()
    • run_ep_hubbard()
  • vayesta.solver.eb_fci.uebfci_slow
    • contract_all()
    • make_shape()
    • contract_1e()
    • contract_2e()
    • contract_ep()
    • contract_pp()
    • apply_bos_annihilation()
    • apply_bos_creation()
    • contract_pp_for_future()
    • slices_for()
    • slices_for_cre()
    • slices_for_des()
    • slices_for_occ_reduction()
    • make_hdiag()
    • kernel()
    • kernel_multiroot()
    • make_rdm1()
    • make_rdm12()
    • make_rdm12s()
    • make_eb_rdm()
    • calc_dd_resp_mom()
    • run()
    • run_hub_test()
    • run_ep_hubbard()
  • Module contents

Submodules

vayesta.solver.callback

class vayesta.solver.callback.CallbackSolver(hamil, log=None, **kwargs)

Bases: ClusterSolver

class Options(callback: int = None)

Bases: Options

callback: int = None

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

kernel(*args, **kwargs)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

calc_v_ext(v_ext_0, cpt)

get_init_guess()

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

vayesta.solver.ccsd

class vayesta.solver.ccsd.RCCSD_Solver(hamil, log=None, **kwargs)

Bases: ClusterSolver

class Options(max_cycle: int = 100, conv_tol: float = None, conv_tol_normt: float = None, diis_space: int = None, diis_start_cycle: int = None, iterative_damping: float = None, level_shift: float = None, init_guess: str = 'MP2', solve_lambda: bool = True, n_moments: tuple = None, sc_mode: int = None)

Bases: Options

max_cycle: int = 100

conv_tol: float = None

conv_tol_normt: float = None

diis_space: int = None

diis_start_cycle: int = None

iterative_damping: float = None

level_shift: float = None

init_guess: str = 'MP2'

solve_lambda: bool = True

n_moments: tuple = None

sc_mode: int = None

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

kernel(t1=None, t2=None, l1=None, l2=None, coupled_fragments=None, t_diagnostic=True)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

get_solver_class(mf)

generate_init_guess(eris=None)

t_diagnostic(solver)

get_callback()

print_extra_info(mycc)

calc_v_ext(v_ext_0, cpt)

get_init_guess()

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

class vayesta.solver.ccsd.UCCSD_Solver(hamil, log=None, **kwargs)

Bases: UClusterSolver, RCCSD_Solver

class Options(max_cycle: int = 100, conv_tol: float = None, conv_tol_normt: float = None, diis_space: int = None, diis_start_cycle: int = None, iterative_damping: float = None, level_shift: float = None, init_guess: str = 'MP2', solve_lambda: bool = True, n_moments: tuple = None, sc_mode: int = None)

Bases: Options

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

conv_tol: float = None

conv_tol_normt: float = None

static dict_with_defaults(**kwargs)

diis_space: int = None

diis_start_cycle: int = None

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

init_guess: str = 'MP2'

items()

iterative_damping: float = None

keys()

level_shift: float = None

max_cycle: int = 100

n_moments: tuple = None

replace(**kwargs)

sc_mode: int = None

solve_lambda: bool = True

update(**kwargs)

values()

get_solver_class(mf)

t_diagnostic(solver)

T diagnostic not implemented for UCCSD in PySCF.

calc_v_ext(v_ext_0, cpt)

generate_init_guess(eris=None)

get_callback()

get_init_guess()

kernel(t1=None, t2=None, l1=None, l2=None, coupled_fragments=None, t_diagnostic=True)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

print_extra_info(mycc)

property v_ext

class vayesta.solver.ccsd.UCCSD(mf, frozen=None, mo_coeff=None, mo_occ=None)

Bases: UCCSD

ao2mo(mo_coeff=None)

EOMEA(*args, **kwargs)

EOMEE(*args, **kwargs)

EOMEESpinFlip(*args, **kwargs)

EOMEESpinKeep(*args, **kwargs)

EOMIP(*args, **kwargs)

amplitudes_from_rccsd(t1, t2)

amplitudes_to_vector(t1, t2, out=None)

apply(fn, *args, **kwargs)

Apply the fn to rest arguments: return fn(*args, **kwargs). The return value of method set is the object itself. This allows a series of functions/methods to be executed in pipe.

as_scanner()

Generating a scanner/solver for CCSD PES.

The returned solver is a function. This function requires one argument “mol” as input and returns total CCSD energy.

The solver will automatically use the results of last calculation as the initial guess of the new calculation. All parameters assigned in the CCSD and the underlying SCF objects (conv_tol, max_memory etc) are automatically applied in the solver.

Note scanner has side effects. It may change many underlying objects (_scf, with_df, with_x2c, …) during calculation.

Examples:

>>> from pyscf import gto, scf, cc
>>> mol = gto.M(atom='H 0 0 0; F 0 0 1')
>>> cc_scanner = cc.CCSD(scf.RHF(mol)).as_scanner()
>>> e_tot = cc_scanner(gto.M(atom='H 0 0 0; F 0 0 1.1'))
>>> e_tot = cc_scanner(gto.M(atom='H 0 0 0; F 0 0 1.5'))

async_io = True

callback = None

cc2 = False

ccsd(t1=None, t2=None, eris=None, mbpt2=False)

Ground-state unrestricted (U)CCSD.

Kwargs:

mbpt2bool

Use one-shot MBPT2 approximation to CCSD.

ccsd_t(t1=None, t2=None, eris=None)

check_sanity()

Check input of class/object attributes, check whether a class method is overwritten. It does not check the attributes which are prefixed with “_”. The return value of method set is the object itself. This allows a series of functions/methods to be executed in pipe.

conv_tol = 1e-07

conv_tol_normt = 1e-06

copy()

Returns a shallow copy

density_fit(auxbasis=None, with_df=None)

diis = True

diis_file = None

diis_space = 6

diis_start_cycle = 0

diis_start_energy_diff = 1000000000.0

direct = False

dump_chk(t1_t2=None, frozen=None, mo_coeff=None, mo_occ=None)

dump_flags(verbose=None)

property e_tot

eaccsd(nroots=1, left=False, koopmans=False, guess=None, partition=None, eris=None)

property ecc

eeccsd(nroots=1, koopmans=False, guess=None, eris=None)

energy(t1=None, t2=None, eris=None)

UCCSD correlation energy

eomea_method()

eomee_ccsd(nroots=1, koopmans=False, guess=None, eris=None)

eomee_ccsd_singlet(nroots=1, koopmans=False, guess=None, eris=None)

eomee_ccsd_triplet(nroots=1, koopmans=False, guess=None, eris=None)

eomee_method()

eomip_method()

eomsf_ccsd(nroots=1, koopmans=False, guess=None, eris=None)

get_e_hf(mo_coeff=None)

get_frozen_mask()

Get boolean mask for the unrestricted reference orbitals.

In the returned boolean (mask) array of frozen orbital indices, the element is False if it corresponds to the frozen orbital.

get_init_guess(eris=None)

get_nmo()

get_nocc()

incore_complete = False

init_amps(eris=None)

ipccsd(nroots=1, left=False, koopmans=False, guess=None, partition=None, eris=None)

iterative_damping = 1.0

kernel(t1=None, t2=None, eris=None, mbpt2=False)

Kernel function is the main driver of a method. Every method should define the kernel function as the entry of the calculation. Note the return value of kernel function is not strictly defined. It can be anything related to the method (such as the energy, the wave-function, the DFT mesh grids etc.).

make_rdm1(t1=None, t2=None, l1=None, l2=None, ao_repr=False, with_frozen=True, with_mf=True)

Un-relaxed 1-particle density matrix in MO space

Returns:

dm1a, dm1b

make_rdm2(t1=None, t2=None, l1=None, l2=None, ao_repr=False, with_frozen=True, with_dm1=True)

2-particle density matrix in spin-orbital basis.

max_cycle = 50

property nmo

property nocc

nuc_grad_method()

post_kernel(envs)

A hook to be run after the main body of the kernel function. Internal variables are exposed to post_kernel through the “envs” dictionary. Return value of post_kernel function is not required.

pre_kernel(envs)

A hook to be run before the main body of kernel function is executed. Internal variables are exposed to pre_kernel through the “envs” dictionary. Return value of pre_kernel function is not required.

reset(mol=None)

restore_from_diis_(diis_file, inplace=True)

Reuse an existed DIIS object in the CCSD calculation.

The CCSD amplitudes will be restored from the DIIS object to generate t1 and t2 amplitudes. The t1/t2 amplitudes of the CCSD object will be overwritten by the generated t1 and t2 amplitudes. The amplitudes vector and error vector will be reused in the CCSD calculation.

run(*args, **kwargs)

Call the kernel function of current object. args will be passed to kernel function. kwargs will be used to update the attributes of current object. The return value of method run is the object itself. This allows a series of functions/methods to be executed in pipe.

run_diis(t1, t2, istep, normt, de, adiis)

set(*args, **kwargs)

Update the attributes of the current object. The return value of method set is the object itself. This allows a series of functions/methods to be executed in pipe.

set_frozen(method='auto', window=(-1000.0, 1000.0))

solve_lambda(t1=None, t2=None, l1=None, l2=None, eris=None)

spin_square(mo_coeff=None, s=None)

stdout = <_io.TextIOWrapper name='<stdout>' mode='w' encoding='utf-8'>

to_gpu(out=None)

Convert a method to its corresponding GPU variant, and recursively converts all attributes of a method to cupy objects or gpu4pyscf objects.

uccsd_t(t1=None, t2=None, eris=None)

update_amps(t1, t2, eris)

vector_size(nmo=None, nocc=None)

vector_to_amplitudes(vector, nmo=None, nocc=None)

verbose = 0

view(cls)

New view of object with the same attributes.

vayesta.solver.ccsdtq

vayesta.solver.ccsdtq.t2_residual_rhf_t3v(solver, fragment, t3, v)

vayesta.solver.ccsdtq.t_residual_rhf(solver, fragment, t1, t2, t3, t4, f, v, include_t3v=False)

vayesta.solver.ccsdtq.t2_residual_uhf_t3v(solver, fragment, t3, v)

vayesta.solver.ccsdtq.t_residual_uhf(solver, fragment, t1, t2, t3, t4, f, v, include_t3v=False)

vayesta.solver.cisd

vayesta.solver.cisd.CISD_Solver(hamil, *args, **kwargs)

class vayesta.solver.cisd.RCISD_Solver(hamil, log=None, **kwargs)

Bases: ClusterSolver

class Options(conv_tol: float = None)

Bases: Options

conv_tol: float = None

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

kernel(*args, **kwargs)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

get_solver_class()

calc_v_ext(v_ext_0, cpt)

get_init_guess()

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

class vayesta.solver.cisd.UCISD_Solver(hamil, log=None, **kwargs)

Bases: UClusterSolver, RCISD_Solver

class Options(conv_tol: float = None)

Bases: Options

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

conv_tol: float = None

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

get_solver_class()

calc_v_ext(v_ext_0, cpt)

get_init_guess()

kernel(*args, **kwargs)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

class vayesta.solver.cisd.UCISD(mf, frozen=None, mo_coeff=None, mo_occ=None)

Bases: UCISD

ao2mo(mo_coeff=None)

amplitudes_to_cisdvec(c0, c1, c2)

apply(fn, *args, **kwargs)

Apply the fn to rest arguments: return fn(*args, **kwargs). The return value of method set is the object itself. This allows a series of functions/methods to be executed in pipe.

as_scanner()

Generating a scanner/solver for CISD PES.

The returned solver is a function. This function requires one argument “mol” as input and returns total CISD energy.

The solver will automatically use the results of last calculation as the initial guess of the new calculation. All parameters assigned in the CISD and the underlying SCF objects (conv_tol, max_memory etc) are automatically applied in the solver.

Note scanner has side effects. It may change many underlying objects (_scf, with_df, with_x2c, …) during calculation.

Examples:

>>> from pyscf import gto, scf, ci
>>> mol = gto.M(atom='H 0 0 0; F 0 0 1')
>>> ci_scanner = ci.CISD(scf.RHF(mol)).as_scanner()
>>> e_tot = ci_scanner(gto.M(atom='H 0 0 0; F 0 0 1.1'))
>>> e_tot = ci_scanner(gto.M(atom='H 0 0 0; F 0 0 1.5'))

async_io = True

check_sanity()

Check input of class/object attributes, check whether a class method is overwritten. It does not check the attributes which are prefixed with “_”. The return value of method set is the object itself. This allows a series of functions/methods to be executed in pipe.

cisd(ci0=None, eris=None)

cisdvec_to_amplitudes(civec, nmo=None, nocc=None, copy=True)

contract(civec, eris)

Application of CISD hamiltonian onto civec.

Parameters:

• myci – CISD (inheriting) object

• civec – numpy array, same length as a CI vector.

• eris – ccsd._ChemistsERIs (inheriting) object (poss diff for df) Contains the various (pq|rs) integrals needed.

Returns:

numpy array, same length as a CI vector.

conv_tol = 1e-09

copy()

Returns a shallow copy

density_fit()

direct = False

dump_chk(ci=None, frozen=None, mo_coeff=None, mo_occ=None)

dump_flags(verbose=None)

property e_tot

from_fcivec(fcivec, nmo=None, nocc=None)

get_e_hf(mo_coeff=None)

get_frozen_mask()

Get boolean mask for the unrestricted reference orbitals.

In the returned boolean (mask) array of frozen orbital indices, the element is False if it corresponds to the frozen orbital.

get_init_guess(eris=None, nroots=1, diag=None)

MP2 energy and MP2 initial guess(es) for CISD coefficients.

Kwargs:

erisccsd._ChemistsERIs (inheriting) object (poss diff for df)

Contains the various (pq|rs) integrals needed.

nrootsinteger

Number of CISD solutions to be found.

diagnumpy array (1D)

e.g. CISD Hamiltonian diagonal in Slater determinant space with HF energy subtracted.

Returns:

Tuple of float and numpy array or tuple of float and list of numpy arrays (if nroots > 1) MP2 energy and initial guess(es) for CISD coefficients.

get_nmo()

get_nocc()

kernel(ci0=None, eris=None)

Kernel function is the main driver of a method. Every method should define the kernel function as the entry of the calculation. Note the return value of kernel function is not strictly defined. It can be anything related to the method (such as the energy, the wave-function, the DFT mesh grids etc.).

level_shift = 0.001

lindep = 1e-14

make_diagonal(eris)

Return diagonal of CISD hamiltonian in Slater determinant basis.

Note that a constant has been substracted of all elements. The first element is the HF energy (minus the constant), the next elements are the diagonal elements with singly excited determinants (<D_i^a|H|D_i^a> within the constant), then doubly excited determinants (<D_ij^ab|H|D_ij^ab> within the constant).

Parameters:

• myci – CISD (inheriting) object

• eris – ccsd._ChemistsERIs (inheriting) object (poss diff for df) Contains the various (pq|rs) integrals needed.

Returns:

(1, 1 + #single excitations from HF det

• #double excitations from HF det))

Diagonal elements of hamiltonian matrix within a constant, see above.

Return type:

numpy array (size

make_rdm1(civec=None, nmo=None, nocc=None, ao_repr=False)

One-particle spin density matrices dm1a, dm1b in MO basis (the occupied-virtual blocks due to the orbital response contribution are not included).

dm1a[p,q] = <q_alpha^dagger p_alpha> dm1b[p,q] = <q_beta^dagger p_beta>

The convention of 1-pdm is based on McWeeney’s book, Eq (5.4.20).

make_rdm2(civec=None, nmo=None, nocc=None, ao_repr=False)

Two-particle spin density matrices dm2aa, dm2ab, dm2bb in MO basis

dm2aa[p,q,r,s] = <q_alpha^dagger s_alpha^dagger r_alpha p_alpha> dm2ab[p,q,r,s] = <q_alpha^dagger s_beta^dagger r_beta p_alpha> dm2bb[p,q,r,s] = <q_beta^dagger s_beta^dagger r_beta p_beta>

(p,q correspond to one particle and r,s correspond to another particle) Two-particle density matrix should be contracted to integrals with the pattern below to compute energy

E = numpy.einsum(‘pqrs,pqrs’, eri_aa, dm2_aa) E+= numpy.einsum(‘pqrs,pqrs’, eri_ab, dm2_ab) E+= numpy.einsum(‘pqrs,rspq’, eri_ba, dm2_ab) E+= numpy.einsum(‘pqrs,pqrs’, eri_bb, dm2_bb)

where eri_aa[p,q,r,s] = (p_alpha q_alpha | r_alpha s_alpha ) eri_ab[p,q,r,s] = ( p_alpha q_alpha | r_beta s_beta ) eri_ba[p,q,r,s] = ( p_beta q_beta | r_alpha s_alpha ) eri_bb[p,q,r,s] = ( p_beta q_beta | r_beta s_beta )

max_cycle = 50

max_space = 12

property nmo

property nocc

property nstates

nuc_grad_method()

post_kernel(envs)

A hook to be run after the main body of the kernel function. Internal variables are exposed to post_kernel through the “envs” dictionary. Return value of post_kernel function is not required.

pre_kernel(envs)

A hook to be run before the main body of kernel function is executed. Internal variables are exposed to pre_kernel through the “envs” dictionary. Return value of pre_kernel function is not required.

reset(mol=None)

run(*args, **kwargs)

Call the kernel function of current object. args will be passed to kernel function. kwargs will be used to update the attributes of current object. The return value of method run is the object itself. This allows a series of functions/methods to be executed in pipe.

set(*args, **kwargs)

Update the attributes of the current object. The return value of method set is the object itself. This allows a series of functions/methods to be executed in pipe.

stdout = <_io.TextIOWrapper name='<stdout>' mode='w' encoding='utf-8'>

to_fcivec(cisdvec, nmo=None, nocc=None)

to_gpu(out=None)

Convert a method to its corresponding GPU variant, and recursively converts all attributes of a method to cupy objects or gpu4pyscf objects.

trans_rdm1(cibra, ciket, nmo=None, nocc=None)

One-particle spin density matrices dm1a, dm1b in MO basis (the occupied-virtual blocks due to the orbital response contribution are not included).

dm1a[p,q] = <q_alpha^dagger p_alpha> dm1b[p,q] = <q_beta^dagger p_beta>

The convention of 1-pdm is based on McWeeney’s book, Eq (5.4.20).

vector_size()

The size of the vector which was returned from amplitudes_to_cisdvec()

verbose = 0

view(cls)

New view of object with the same attributes.

vayesta.solver.coupled_ccsd

class vayesta.solver.coupled_ccsd.coupledRCCSD_Solver(hamil, log=None, **kwargs)

Bases: RCCSD_Solver

class Options(max_cycle: int = 100, conv_tol: float = None, conv_tol_normt: float = None, diis_space: int = None, diis_start_cycle: int = None, iterative_damping: float = None, level_shift: float = None, init_guess: str = 'MP2', solve_lambda: bool = True, n_moments: tuple = None, sc_mode: int = None, fragments: List | NoneType = None)

Bases: Options

fragments: List | None = None

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

conv_tol: float = None

conv_tol_normt: float = None

static dict_with_defaults(**kwargs)

diis_space: int = None

diis_start_cycle: int = None

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

init_guess: str = 'MP2'

items()

iterative_damping: float = None

keys()

level_shift: float = None

max_cycle: int = 100

n_moments: tuple = None

replace(**kwargs)

sc_mode: int = None

solve_lambda: bool = True

update(**kwargs)

values()

set_coupled_fragments(fragments)

get_callback()

calc_v_ext(v_ext_0, cpt)

generate_init_guess(eris=None)

get_init_guess()

get_solver_class(mf)

kernel(t1=None, t2=None, l1=None, l2=None, coupled_fragments=None, t_diagnostic=True)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

print_extra_info(mycc)

t_diagnostic(solver)

property v_ext

vayesta.solver.coupling

vayesta.solver.coupling.transform_amplitude(t, u_occ, u_vir, u_occ2=None, u_vir2=None, inverse=False)

u: (old basis|new basis)

vayesta.solver.coupling.get_amplitude_norm(t1, t2)

vayesta.solver.coupling.project_t2(t2, proj, projectors)

vayesta.solver.coupling.project_t2_rspin(t2, proj, projectors)

vayesta.solver.coupling.project_t2_uspin(t2, proj, projectors)

vayesta.solver.coupling.couple_ccsd_iterations(solver, fragments)

Requires MPI.

vayesta.solver.coupling.tailor_with_fragments(solver, fragments, project=False, tailor_t1=True, tailor_t2=True, ovlp_tol=1e-06)

Tailor current CCSD calculation with amplitudes of other fragments.

This assumes orthogonal fragment spaces.

Parameters:

• project (int, optional) –

  Level of external correction of T2 amplitudes: 1: Both occupied indices are projected to each other fragment X. 2: Both occupied indices are projected to each other fragment X

  and combinations of other fragments X,Y.

  3: Only the first occupied indices is projected to each other fragment X.

• coupled_fragments (list, optional) – List of fragments, which are used for the external correction. Each fragment x must have the following attributes defined: c_active_occ : Active occupied MO orbitals of fragment x c_active_vir : Active virtual MO orbitals of fragment x results.t1 : T1 amplitudes of fragment x results.t2 : T2 amplitudes of fragment x

Returns:

tailor_func – Tailoring function for CCSD.

Return type:

function(cc, t1, t2) -> t1, t2

vayesta.solver.coupling.externally_correct(solver, external_corrections, hamil=None)

Build callback function for CCSD, to add external correction from other fragments.

TODO: combine with tailor_with_fragments?

Parameters:

• solver (CCSD_Solver) – Vayesta CCSD solver.

• external_corrections (list of tuple of (int, str, int, bool)) – List of external corrections. Each tuple contains the fragment ID, type of correction, and number of projectors for the given external correction. Final element is boolean giving the low_level_coul optional argument.

• eris (_ChemistsERIs) – ERIs for parent CCSD fragment. Used for MO energies in residual contraction, and for the case of low_level_coul, where the parent Coulomb integral is contracted. If not passed in, MO energy if needed will be constructed from the diagonal of get_fock() of embedding base class, and the eris will be also be obtained from the embedding base class. Optional.

Returns:

callback – Callback function for PySCF’s CCSD solver.

Return type:

callable

vayesta.solver.dump

class vayesta.solver.dump.DumpSolver(hamil, log=None, **kwargs)

Bases: ClusterSolver

class Options(dumpfile: str = None)

Bases: Options

dumpfile: str = None

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

kernel(*args, **kwargs)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

calc_v_ext(v_ext_0, cpt)

get_init_guess()

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

vayesta.solver.ebcc

class vayesta.solver.ebcc.REBCC_Solver(hamil, log=None, **kwargs)

Bases: ClusterSolver

class Options(ansatz: str = 'CCSD', solve_lambda: bool = False, max_cycle: int = 200, conv_tol: float = None, conv_tol_normt: float = None, store_as_ccsd: bool = False, c_cas_occ: <built-in function array> = None, c_cas_vir: <built-in function array> = None)

Bases: Options

ansatz: str = 'CCSD'

solve_lambda: bool = False

max_cycle: int = 200

conv_tol: float = None

conv_tol_normt: float = None

store_as_ccsd: bool = False

c_cas_occ: array = None

c_cas_vir: array = None

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

kernel()

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

get_space(mo_coeff, mo_occ, frozen=None)

get_nonnull_solver_opts()

construct_wavefunction(mycc, mo, mbos=None)

calc_v_ext(v_ext_0, cpt)

get_init_guess()

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

class vayesta.solver.ebcc.UEBCC_Solver(hamil, log=None, **kwargs)

Bases: UClusterSolver, REBCC_Solver

class Options(ansatz: str = 'CCSD', solve_lambda: bool = False, max_cycle: int = 200, conv_tol: float = None, conv_tol_normt: float = None, store_as_ccsd: bool = False, c_cas_occ: <built-in function array> = (None, None), c_cas_vir: <built-in function array> = (None, None))

Bases: Options, Options

c_cas_occ: array = (None, None)

c_cas_vir: array = (None, None)

ansatz: str = 'CCSD'

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

conv_tol: float = None

conv_tol_normt: float = None

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

max_cycle: int = 200

replace(**kwargs)

solve_lambda: bool = False

store_as_ccsd: bool = False

update(**kwargs)

values()

get_space(mo_coeff, mo_occ, frozen=None)

construct_wavefunction(mycc, mo, mbos=None)

calc_v_ext(v_ext_0, cpt)

get_init_guess()

get_nonnull_solver_opts()

kernel()

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

class vayesta.solver.ebcc.EB_REBCC_Solver(hamil, log=None, **kwargs)

Bases: REBCC_Solver

class Options(ansatz: str = 'CCSD-S-1-1', solve_lambda: bool = False, max_cycle: int = 200, conv_tol: float = None, conv_tol_normt: float = None, store_as_ccsd: bool = False, c_cas_occ: <built-in function array> = None, c_cas_vir: <built-in function array> = None)

Bases: Options

ansatz: str = 'CCSD-S-1-1'

store_as_ccsd: bool = False

asdict(deepcopy=False)

c_cas_occ: array = None

c_cas_vir: array = None

classmethod change_dict_defaults(field, **kwargs)

conv_tol: float = None

conv_tol_normt: float = None

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

max_cycle: int = 200

replace(**kwargs)

solve_lambda: bool = False

update(**kwargs)

values()

get_nonnull_solver_opts()

get_couplings()

construct_wavefunction(mycc, mo, mbos=None)

calc_v_ext(v_ext_0, cpt)

get_init_guess()

get_space(mo_coeff, mo_occ, frozen=None)

kernel()

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

class vayesta.solver.ebcc.EB_UEBCC_Solver(hamil, log=None, **kwargs)

Bases: EB_REBCC_Solver, UEBCC_Solver

class Options(ansatz: str = 'CCSD', solve_lambda: bool = False, max_cycle: int = 200, conv_tol: float = None, conv_tol_normt: float = None, store_as_ccsd: bool = False, c_cas_occ: <built-in function array> = (None, None), c_cas_vir: <built-in function array> = (None, None))

Bases: Options, Options

ansatz: str = 'CCSD-S-1-1'

asdict(deepcopy=False)

c_cas_occ: array = (None, None)

c_cas_vir: array = (None, None)

classmethod change_dict_defaults(field, **kwargs)

conv_tol: float = None

conv_tol_normt: float = None

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

max_cycle: int = 200

replace(**kwargs)

solve_lambda: bool = False

store_as_ccsd: bool = False

update(**kwargs)

values()

get_couplings()

construct_wavefunction(mycc, mo, mbos=None)

calc_v_ext(v_ext_0, cpt)

get_init_guess()

get_nonnull_solver_opts()

get_space(mo_coeff, mo_occ, frozen=None)

kernel()

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

vayesta.solver.ebcc.gen_space(c_occ, c_vir, co_active=None, cv_active=None, frozen_orbs=None)

Given the occupied and virtual orbital coefficients in the local cluster basis, which orbitals are frozen, and any active space orbitals in this space generate appropriate coefficients and ebcc.Space inputs for a calculation. Inputs:

c_occ: occupied orbital coefficients in local cluster basis. c_vir: virtual orbital coefficients in local cluster basis. co_active: occupied active space orbitals in local cluster basis. cv_active: virtual active space orbitals in local cluster basis. frozen_orbs: indices of orbitals to freeze in local cluster basis.

Outputs:

c: coefficients for the active space orbitals in the local cluster basis. space: ebcc.Space object defining the resulting active space behaviours.

vayesta.solver.ebfci

class vayesta.solver.ebfci.EB_EBFCI_Solver(hamil, log=None, **kwargs)

Bases: ClusterSolver

class Options(max_cycle: int = 100, conv_tol: float = None, max_boson_occ: int = 2)

Bases: Options

max_cycle: int = 100

conv_tol: float = None

max_boson_occ: int = 2

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

get_solver(*args, **kwargs)

kernel()

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

calc_v_ext(v_ext_0, cpt)

get_init_guess()

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

class vayesta.solver.ebfci.EB_UEBFCI_Solver(hamil, log=None, **kwargs)

Bases: UClusterSolver, EB_EBFCI_Solver

get_solver(*args, **kwargs)

class Options(max_cycle: int = 100, conv_tol: float = None, max_boson_occ: int = 2)

Bases: Options

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

conv_tol: float = None

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

max_boson_occ: int = 2

max_cycle: int = 100

replace(**kwargs)

update(**kwargs)

values()

calc_v_ext(v_ext_0, cpt)

get_init_guess()

kernel()

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

vayesta.solver.ext_ccsd

class vayesta.solver.ext_ccsd.extRCCSD_Solver(hamil, log=None, **kwargs)

Bases: RCCSD_Solver

class Options(max_cycle: int = 100, conv_tol: float = None, conv_tol_normt: float = None, diis_space: int = None, diis_start_cycle: int = None, iterative_damping: float = None, level_shift: float = None, init_guess: str = 'MP2', solve_lambda: bool = True, n_moments: tuple = None, sc_mode: int = None, external_corrections: Union[List[Any], NoneType] = <factory>)

Bases: Options

external_corrections: List[Any] | None

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

conv_tol: float = None

conv_tol_normt: float = None

static dict_with_defaults(**kwargs)

diis_space: int = None

diis_start_cycle: int = None

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

init_guess: str = 'MP2'

items()

iterative_damping: float = None

keys()

level_shift: float = None

max_cycle: int = 100

n_moments: tuple = None

replace(**kwargs)

sc_mode: int = None

solve_lambda: bool = True

update(**kwargs)

values()

get_callback()

calc_v_ext(v_ext_0, cpt)

generate_init_guess(eris=None)

get_init_guess()

get_solver_class(mf)

kernel(t1=None, t2=None, l1=None, l2=None, coupled_fragments=None, t_diagnostic=True)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

print_extra_info(mycc)

t_diagnostic(solver)

property v_ext

class vayesta.solver.ext_ccsd.extUCCSD_Solver(hamil, log=None, **kwargs)

Bases: UCCSD_Solver, extRCCSD_Solver

class Options(max_cycle: int = 100, conv_tol: float = None, conv_tol_normt: float = None, diis_space: int = None, diis_start_cycle: int = None, iterative_damping: float = None, level_shift: float = None, init_guess: str = 'MP2', solve_lambda: bool = True, n_moments: tuple = None, sc_mode: int = None, external_corrections: Union[List[Any], NoneType] = <factory>)

Bases: Options, Options

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

conv_tol: float = None

conv_tol_normt: float = None

static dict_with_defaults(**kwargs)

diis_space: int = None

diis_start_cycle: int = None

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

init_guess: str = 'MP2'

items()

iterative_damping: float = None

keys()

level_shift: float = None

max_cycle: int = 100

n_moments: tuple = None

replace(**kwargs)

sc_mode: int = None

solve_lambda: bool = True

update(**kwargs)

values()

external_corrections: List[Any] | None

calc_v_ext(v_ext_0, cpt)

generate_init_guess(eris=None)

get_callback()

get_init_guess()

get_solver_class(mf)

kernel(t1=None, t2=None, l1=None, l2=None, coupled_fragments=None, t_diagnostic=True)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

print_extra_info(mycc)

t_diagnostic(solver)

T diagnostic not implemented for UCCSD in PySCF.

property v_ext

vayesta.solver.fci

class vayesta.solver.fci.FCI_Solver(*args, **kwargs)

Bases: ClusterSolver

class Options(threads: int = 1, max_cycle: int = 300, lindep: float = None, conv_tol: float = 1e-12, solver_spin: bool = True, fix_spin: float = 0.0, fix_spin_penalty: float = 1.0, davidson_only: bool = True, init_guess: str = 'default', init_guess_noise: float = 1e-05, n_moments: tuple = None)

Bases: Options

threads: int = 1

max_cycle: int = 300

lindep: float = None

conv_tol: float = 1e-12

solver_spin: bool = True

fix_spin: float = 0.0

fix_spin_penalty: float = 1.0

davidson_only: bool = True

init_guess: str = 'default'

init_guess_noise: float = 1e-05

n_moments: tuple = None

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

cisd_solver

alias of RCISD_Solver

get_solver_class()

kernel(ci=None)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

calc_v_ext(v_ext_0, cpt)

get_init_guess()

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

class vayesta.solver.fci.UFCI_Solver(*args, **kwargs)

Bases: UClusterSolver, FCI_Solver

class Options(threads: int = 1, max_cycle: int = 300, lindep: float = None, conv_tol: float = 1e-12, solver_spin: bool = True, fix_spin: float = None, fix_spin_penalty: float = 1.0, davidson_only: bool = True, init_guess: str = 'default', init_guess_noise: float = 1e-05, n_moments: tuple = None)

Bases: Options

fix_spin: float = None

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

conv_tol: float = 1e-12

davidson_only: bool = True

static dict_with_defaults(**kwargs)

fix_spin_penalty: float = 1.0

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

init_guess: str = 'default'

init_guess_noise: float = 1e-05

items()

keys()

lindep: float = None

max_cycle: int = 300

n_moments: tuple = None

replace(**kwargs)

solver_spin: bool = True

threads: int = 1

update(**kwargs)

values()

cisd_solver

alias of UCISD_Solver

get_solver_class()

kernel(ci=None)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

calc_v_ext(v_ext_0, cpt)

get_init_guess()

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

vayesta.solver.hamiltonian

vayesta.solver.hamiltonian.is_ham(ham)

vayesta.solver.hamiltonian.is_uhf_ham(ham)

vayesta.solver.hamiltonian.is_eb_ham(ham)

vayesta.solver.hamiltonian.ClusterHamiltonian(fragment, mf, log=None, **kwargs)

class vayesta.solver.hamiltonian.DummyERIs(getter, valid_blocks, **kwargs)

Bases: object

class vayesta.solver.hamiltonian.RClusterHamiltonian(fragment, mf, log=None, cluster=None, **kwargs)

Bases: object

class Options(screening: str | NoneType = None, cache_eris: bool = True, match_fock: bool = True)

Bases: OptionsBase

screening: str | None = None

cache_eris: bool = True

match_fock: bool = True

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

property cluster

property mo

property mbos

property nelec

property ncas

property has_screening

target_space_projector(c=None)

Projector to the target fragment space within our cluster.

get_mo(mo_coeff=None)

get_integrals(bare_eris=None, with_vext=True)

get_fock(with_vext=True, use_seris=None, with_exxdiv=False)

get_heff(eris=None, fock=None, with_vext=True, with_exxdiv=False)

get_eris_screened(block=None)

get_eris_bare(block=None)

get_cderi_bare(only_ov=False, compress=False, svd_threshold=1e-12)

to_pyscf_mf(allow_dummy_orbs=False, force_bare_eris=False, overwrite_fock=False, allow_df=False)

Generate pyscf.scf object representing this active space Hamiltonian. This should be able to be passed into a standard post-hf pyscf solver without modification.

Parameters:

• allow_dummy_orbs (bool, optional) – Whether the introduction of dummy orbitals into the mean-field, which are then frozen, is permitted. Default is False

• force_bare_eris (bool, optional) – Forces resultant mean-field object to use unscreened eris. Default is False

• overwrite_fock (bool, optional) – Whether mf.get_fock should be set to self.get_fock. Mainly for use in UHF. Default is False

• allow_df (bool, optional) – Whether the resultant mean-field object should include a .with_df object containing the projection of the CDERIs into the cluster space. Default is False

Returns:

• clusmf (pyscf.scf.SCF) – Representation of cluster as pyscf mean-field.

• orbs_to_freeze (list of lists) – Which orbitals to freeze, split by spin channel if UHF.

get_clus_mf_info(ao_basis=False, with_vext=True, with_exxdiv=False)

calc_loc_erpa(m0, amb, only_cumulant=False)

add_screening(seris_intermed=None)

Adds appropriate screening according to the value of self.opts.screening.

-None: gives bare interactions, but this function shouldn’t be called in that case. -‘mrpa’: moment-conserving interactions. -‘crpa’: gives cRPA interactions. Including ‘ov’ after ‘crpa’ will only apply cRPA screening in the o-v channel.

Including ‘pcoupling’ similarly will account for the polarisability in non-canonical cluster spaces.

seris_intermed is only required for mRPA interactions.

assert_equal_spin_channels(message='')

with_new_cluster(cluster)

get_dummy_eri_object(force_bare=False, with_vext=True, with_exxdiv=False)

class vayesta.solver.hamiltonian.UClusterHamiltonian(fragment, mf, log=None, cluster=None, **kwargs)

Bases: RClusterHamiltonian

property ncas

property nelec

get_mo(mo_coeff=None)

get_fock(with_vext=True, use_seris=True, with_exxdiv=False)

get_heff(eris=None, fock=None, with_vext=True, with_exxdiv=False)

get_eris_bare(block=None)

get_cderi_bare(only_ov=False, compress=False, svd_threshold=1e-12)

to_pyscf_mf(allow_dummy_orbs=True, force_bare_eris=False, overwrite_fock=True, allow_df=False)

Generate pyscf.scf object representing this active space Hamiltonian. This should be able to be passed into a standard post-hf pyscf solver without modification.

Parameters:

• allow_dummy_orbs (bool, optional) – Whether the introduction of dummy orbitals into the mean-field, which are then frozen, is permitted. Default is False

• force_bare_eris (bool, optional) – Forces resultant mean-field object to use unscreened eris. Default is False

• overwrite_fock (bool, optional) – Whether mf.get_fock should be set to self.get_fock. Mainly for use in UHF. Default is False

• allow_df (bool, optional) – Whether the resultant mean-field object should include a .with_df object containing the projection of the CDERIs into the cluster space. Default is False

Returns:

• clusmf (pyscf.scf.SCF) – Representation of cluster as pyscf mean-field.

• orbs_to_freeze (list of lists) – Which orbitals to freeze, split by spin channel if UHF.

get_clus_mf_info(ao_basis=False, with_vext=True, with_exxdiv=False)

get_dummy_eri_object(force_bare=False, with_vext=True, with_exxdiv=False)

add_screening(seris_intermed=None)

Add screened interactions into the Hamiltonian.

class Options(screening: str | NoneType = None, cache_eris: bool = True, match_fock: bool = True)

Bases: OptionsBase

asdict(deepcopy=False)

cache_eris: bool = True

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

match_fock: bool = True

replace(**kwargs)

screening: str | None = None

update(**kwargs)

values()

assert_equal_spin_channels(message='')

calc_loc_erpa(m0, amb, only_cumulant=False)

property cluster

get_eris_screened(block=None)

get_integrals(bare_eris=None, with_vext=True)

property has_screening

property mbos

property mo

target_space_projector(c=None)

Projector to the target fragment space within our cluster.

with_new_cluster(cluster)

class vayesta.solver.hamiltonian.EB_RClusterHamiltonian(*args, **kwargs)

Bases: RClusterHamiltonian

class Options(screening: str | NoneType = None, cache_eris: bool = True, match_fock: bool = True, polaritonic_shift: bool = True)

Bases: Options

polaritonic_shift: bool = True

asdict(deepcopy=False)

cache_eris: bool = True

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

match_fock: bool = True

replace(**kwargs)

screening: str | None = None

update(**kwargs)

values()

property polaritonic_shift

property couplings

property unshifted_couplings

property bos_freqs

property mbos

initialise_bosons(bos_freqs, bos_couplings, rotation=None, boson_nonconserving=None)

generate_bosonic_interactions()

set_polaritonic_shift(freqs, couplings)

get_heff(eris=None, fock=None, with_vext=True)

get_polaritonic_fock_shift(couplings)

get_polaritonic_shifted_couplings()

get_eb_dm_polaritonic_shift(dm1)

add_screening(seris_intermed=None)

Adds appropriate screening according to the value of self.opts.screening.

-None: gives bare interactions, but this function shouldn’t be called in that case. -‘mrpa’: moment-conserving interactions. -‘crpa’: gives cRPA interactions. Including ‘ov’ after ‘crpa’ will only apply cRPA screening in the o-v channel.

Including ‘pcoupling’ similarly will account for the polarisability in non-canonical cluster spaces.

seris_intermed is only required for mRPA interactions.

assert_equal_spin_channels(message='')

calc_loc_erpa(m0, amb, only_cumulant=False)

property cluster

get_cderi_bare(only_ov=False, compress=False, svd_threshold=1e-12)

get_clus_mf_info(ao_basis=False, with_vext=True, with_exxdiv=False)

get_dummy_eri_object(force_bare=False, with_vext=True, with_exxdiv=False)

get_eris_bare(block=None)

get_eris_screened(block=None)

get_fock(with_vext=True, use_seris=None, with_exxdiv=False)

get_integrals(bare_eris=None, with_vext=True)

get_mo(mo_coeff=None)

property has_screening

property mo

property ncas

property nelec

target_space_projector(c=None)

Projector to the target fragment space within our cluster.

to_pyscf_mf(allow_dummy_orbs=False, force_bare_eris=False, overwrite_fock=False, allow_df=False)

Generate pyscf.scf object representing this active space Hamiltonian. This should be able to be passed into a standard post-hf pyscf solver without modification.

Parameters:

• allow_dummy_orbs (bool, optional) – Whether the introduction of dummy orbitals into the mean-field, which are then frozen, is permitted. Default is False

• force_bare_eris (bool, optional) – Forces resultant mean-field object to use unscreened eris. Default is False

• overwrite_fock (bool, optional) – Whether mf.get_fock should be set to self.get_fock. Mainly for use in UHF. Default is False

• allow_df (bool, optional) – Whether the resultant mean-field object should include a .with_df object containing the projection of the CDERIs into the cluster space. Default is False

Returns:

• clusmf (pyscf.scf.SCF) – Representation of cluster as pyscf mean-field.

• orbs_to_freeze (list of lists) – Which orbitals to freeze, split by spin channel if UHF.

with_new_cluster(cluster)

class vayesta.solver.hamiltonian.EB_UClusterHamiltonian(*args, **kwargs)

Bases: UClusterHamiltonian, EB_RClusterHamiltonian

class Options(screening: str | NoneType = None, cache_eris: bool = True, match_fock: bool = True, polaritonic_shift: bool = True)

Bases: Options

polaritonic_shift: bool = True

asdict(deepcopy=False)

cache_eris: bool = True

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

match_fock: bool = True

replace(**kwargs)

screening: str | None = None

update(**kwargs)

values()

property couplings

get_heff(eris=None, fock=None, with_vext=True)

add_screening(seris_intermed=None)

Add screened interactions into the Hamiltonian.

assert_equal_spin_channels(message='')

property bos_freqs

property cluster

generate_bosonic_interactions()

get_cderi_bare(only_ov=False, compress=False, svd_threshold=1e-12)

get_clus_mf_info(ao_basis=False, with_vext=True, with_exxdiv=False)

get_dummy_eri_object(force_bare=False, with_vext=True, with_exxdiv=False)

get_eris_bare(block=None)

get_eris_screened(block=None)

get_fock(with_vext=True, use_seris=True, with_exxdiv=False)

get_integrals(bare_eris=None, with_vext=True)

get_mo(mo_coeff=None)

get_polaritonic_fock_shift(couplings)

get_polaritonic_shifted_couplings()

property has_screening

initialise_bosons(bos_freqs, bos_couplings, rotation=None, boson_nonconserving=None)

property mbos

property mo

property ncas

property nelec

property polaritonic_shift

set_polaritonic_shift(freqs, couplings)

target_space_projector(c=None)

Projector to the target fragment space within our cluster.

to_pyscf_mf(allow_dummy_orbs=True, force_bare_eris=False, overwrite_fock=True, allow_df=False)

Generate pyscf.scf object representing this active space Hamiltonian. This should be able to be passed into a standard post-hf pyscf solver without modification.

Parameters:

• allow_dummy_orbs (bool, optional) – Whether the introduction of dummy orbitals into the mean-field, which are then frozen, is permitted. Default is False

• force_bare_eris (bool, optional) – Forces resultant mean-field object to use unscreened eris. Default is False

• overwrite_fock (bool, optional) – Whether mf.get_fock should be set to self.get_fock. Mainly for use in UHF. Default is False

• allow_df (bool, optional) – Whether the resultant mean-field object should include a .with_df object containing the projection of the CDERIs into the cluster space. Default is False

Returns:

• clusmf (pyscf.scf.SCF) – Representation of cluster as pyscf mean-field.

• orbs_to_freeze (list of lists) – Which orbitals to freeze, split by spin channel if UHF.

property unshifted_couplings

with_new_cluster(cluster)

get_eb_dm_polaritonic_shift(dm1)

calc_loc_erpa(m0, amb)

vayesta.solver.mp2

class vayesta.solver.mp2.RMP2_Solver(hamil, log=None, **kwargs)

Bases: ClusterSolver

class Options(compress_cderi: bool = False)

Bases: Options

compress_cderi: bool = False

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

kernel(*args, **kwargs)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

get_ov_eris(eris, nocc)

make_t2(mo_energy, eris=None, cderi=None, cderi_neg=None, blksize=None)

Make T2 amplitudes

calc_v_ext(v_ext_0, cpt)

get_init_guess()

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

class vayesta.solver.mp2.UMP2_Solver(hamil, log=None, **kwargs)

Bases: UClusterSolver, RMP2_Solver

get_ov_eris(eris, nocc)

make_t2(mo_energy, eris=None, cderi=None, cderi_neg=None, blksize=None, workmem=1000000000)

Make T2 amplitudes

class Options(compress_cderi: bool = False)

Bases: Options

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

compress_cderi: bool = False

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

calc_v_ext(v_ext_0, cpt)

get_init_guess()

kernel(*args, **kwargs)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

vayesta.solver.solver

class vayesta.solver.solver.ClusterSolver(hamil, log=None, **kwargs)

Bases: object

Base class for cluster solver

class Options

Bases: OptionsBase

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

property v_ext

kernel(*args, **kwargs)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

calc_v_ext(v_ext_0, cpt)

get_init_guess()

class vayesta.solver.solver.UClusterSolver(hamil, log=None, **kwargs)

Bases: ClusterSolver

class Options

Bases: OptionsBase

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

static dict_with_defaults(**kwargs)

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

items()

keys()

replace(**kwargs)

update(**kwargs)

values()

calc_v_ext(v_ext_0, cpt)

get_init_guess()

kernel(*args, **kwargs)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

property v_ext

vayesta.solver.tccsd

class vayesta.solver.tccsd.TRCCSD_Solver(hamil, log=None, **kwargs)

Bases: RCCSD_Solver

class Options(max_cycle: int = 100, conv_tol: float = None, conv_tol_normt: float = None, diis_space: int = None, diis_start_cycle: int = None, iterative_damping: float = None, level_shift: float = None, init_guess: str = 'MP2', solve_lambda: bool = True, n_moments: tuple = None, sc_mode: int = None, fci_opts: dict = <factory>, c_cas_occ: <built-in function array> = None, c_cas_vir: <built-in function array> = None)

Bases: Options

fci_opts: dict

c_cas_occ: array = None

c_cas_vir: array = None

asdict(deepcopy=False)

classmethod change_dict_defaults(field, **kwargs)

conv_tol: float = None

conv_tol_normt: float = None

static dict_with_defaults(**kwargs)

diis_space: int = None

diis_start_cycle: int = None

get(attr, default=None)

Dictionary-like access to attributes. Allows the definition of a default value, of the attribute is not present.

classmethod get_default(field)

classmethod get_default_factory(field)

init_guess: str = 'MP2'

items()

iterative_damping: float = None

keys()

level_shift: float = None

max_cycle: int = 100

n_moments: tuple = None

replace(**kwargs)

sc_mode: int = None

solve_lambda: bool = True

update(**kwargs)

values()

get_callback()

print_extra_info(mycc)

calc_v_ext(v_ext_0, cpt)

generate_init_guess(eris=None)

get_init_guess()

get_solver_class(mf)

kernel(t1=None, t2=None, l1=None, l2=None, coupled_fragments=None, t_diagnostic=True)

Set up everything for a calculation on the CAS and pass this to the solver-specific kernel that runs on this information.

optimize_cpt(nelectron, c_frag, cpt_guess=0, atol=1e-06, rtol=1e-06, cpt_radius=0.5)

Enables chemical potential optimization to match a number of electrons in the fragment space.

Parameters:

• nelectron (float) – Target number of electrons.

• c_frag (array) – Fragment orbitals.

• cpt_guess (float, optional) – Initial guess for fragment chemical potential. Default: 0.

• atol (float, optional) – Absolute electron number tolerance. Default: 1e-6.

• rtol (float, optional) – Relative electron number tolerance. Default: 1e-6

• cpt_radius (float, optional) – Search radius for chemical potential. Default: 0.5.

Returns:

Solver results.

Return type:

results

t_diagnostic(solver)

property v_ext

Module contents

vayesta.solver.get_solver_class(ham, solver)

vayesta.solver.check_solver_config(solver, is_uhf, is_eb, log)