dgamore.config#

Global configuration singleton. This module holds the process-wide mutable state of a DGAmore run as module-level instances of the *Config classes (box, lattice, sys, dmft, eliashberg, lambda_correction, self_consistency, self_energy_interpolation, output, memory, ana_cont, and the logger). ConfigParser populates these from a YAML file on rank 0, after which they are broadcast to all MPI ranks; most modules read their parameters directly off this module rather than receiving them as arguments, so mutating a field changes behavior everywhere. Each *Config class documents its fields; the example YAML config is dgamore/dga_config.yaml.

class dgamore.config.AnaContConfig[source]#

Bases: object

Stores the analytic-continuation (maximum-entropy) configuration.

Variables:

  • do_spectrum_dga (bool) – Whether to continue the DGA Green’s function to real frequencies.

  • do_spectrum_dmft (bool) – Whether to continue the DMFT Green’s function to real frequencies.

  • w_count (int) – Number of real-frequency points.

  • plot_spectrum (bool) – Whether to plot the resulting spectral function.

  • k_path (list) – The k-path (list of (kx, ky, kz, label) tuples) for the spectral function.

  • energy_window (tuple) – The real-frequency window (w_min, w_max).

class dgamore.config.BoxConfig[source]#

Bases: object

Stores the Matsubara frequency box sizes. The main quantities live in the core region; explicit asymptotics correct it with shell-region quantities. The full region is the sum of core and shell and exists for convenience.

Variables:

  • niw_core (int) – Number of positive bosonic core frequencies \(\omega\).

  • niv_core (int) – Number of positive fermionic core frequencies \(\nu\).

  • niv_shell (int) – Number of positive fermionic shell frequencies (asymptotic correction region).

  • niv_full (int) – Number of positive fermionic full-region frequencies (core + shell).

  • niv_dmft (int) – Number of positive fermionic frequencies available in the DMFT 1-particle input.

class dgamore.config.DmftConfig[source]#

Bases: object

Stores the DMFT input-file parameters: the input path, the 1- and 2-particle data filenames, symmetrization options and the inequivalent-atom structure.

Variables:

  • type (str) – DMFT solver/format of the input (e.g. "w2dyn").

  • input_path (str) – Directory containing the DMFT input files.

  • fname_1p (str) – Filename of the 1-particle data.

  • fname_2p (str) – Filename of the 2-particle data.

  • do_sym_v_vp (bool) – Whether to symmetrize the 2-particle data with respect to \((\nu, \nu')\).

  • symmetrize_orbitals (list) – 1-based orbital indices to symmetrize over (empty for none).

  • n_ineq (int) – Number of inequivalent atoms.

  • ineq_ordering (list) – Ordering of the inequivalent atoms.

  • n_bands_per_ineq (list) – Number of bands per inequivalent atom.

class dgamore.config.EliashbergConfig[source]#

Bases: object

Stores the Eliashberg-step configuration: whether to run it, power-iteration settings, and saving options for the pairing/full vertex in pp notation.

Variables:

  • perform_eliashberg (bool) – Whether to solve the Eliashberg equation.

  • save_pairing_vertex (bool) – Whether to save the singlet/triplet pairing vertices.

  • save_fq (bool) – Whether to save the full ladder vertex (in ph notation) in the irreducible BZ.

  • construct_fq_cheap (bool) – Whether to build the full vertex on the smaller pp frequency box (cheaper).

  • n_eig (int) – Number of leading eigenvalues/gap functions to compute per channel.

  • epsilon (float) – Convergence tolerance for the Lanczos eigensolver.

  • symmetry (str) – Initial gap-function symmetry ("d-wave", "p-wave-x", "p-wave-y", or "random").

  • include_local_part (bool) – Whether to add the local reducible pp diagrams to the pairing vertex.

  • subfolder_name (str) – Output subfolder name for Eliashberg results.

class dgamore.config.InteractionConfig[source]#

Bases: object

Stores the interaction parameters. Currently only udd, vdd, jdd are used (local and Kanamori-type interactions); the remaining parameters are reserved for future use.

Variables:

  • udd (float) – Intra-orbital Hubbard interaction \(U_{dd}\) on the d orbitals.

  • udp (float) – Inter-orbital d-p Hubbard interaction.

  • upp (float) – Intra-orbital Hubbard interaction \(U_{pp}\) on the p orbitals.

  • uppod (float) – Off-diagonal p-p Hubbard interaction.

  • jdd (float) – Hund’s exchange \(J_{dd}\) on the d orbitals.

  • jdp (float) – Inter-orbital d-p exchange.

  • jpp (float) – Hund’s exchange \(J_{pp}\) on the p orbitals.

  • jppod (float) – Off-diagonal p-p exchange.

  • vdd (float) – Inter-orbital interaction \(V_{dd}\) on the d orbitals.

  • vpp (float) – Inter-orbital interaction \(V_{pp}\) on the p orbitals.

class dgamore.config.LambdaCorrectionConfig[source]#

Bases: object

Stores the lambda-correction configuration.

Variables:

  • perform_lambda_correction (bool) – Whether to apply the Moriya lambda correction.

  • type (str) – Correction type: "sp" (magnetic channel only) or "spch" (density and magnetic channels).

class dgamore.config.LatticeConfig[source]#

Bases: object

Stores the lattice parameters: the symmetries, lattice type, input Hamiltonian and input interaction. The k and q grids are built from the number of k/q points and the lattice symmetries. See dga_config.yaml or the author’s master’s thesis for details.

Variables:

  • symmetries (list) – The lattice symmetries (a list of KnownSymmetries or the auto sentinel).

  • type (str) – How the kinetic Hamiltonian is provided (e.g. "from_wannier90").

  • er_input (str | list) – Path(s) to the hopping input.

  • interaction_type (str) – How the interaction is provided (e.g. "one_band_from_dmft").

  • interaction_input (str | list) – Path(s) to the interaction input.

  • nk (tuple) – Number of k-points per spatial direction.

  • nq (tuple) – Number of q-points per spatial direction (defaults to nk).

  • interaction (InteractionConfig) – The InteractionConfig.

  • hamiltonian (Hamiltonian) – The Hamiltonian instance.

  • k_grid (KGrid) – The k-space KGrid.

  • q_grid (KGrid) – The q-space KGrid.

class dgamore.config.MemoryConfig[source]#

Bases: object

Stores the speed-vs-memory trade-off switches. Each flag, when True, selects a slower but more memory-lean code path for the corresponding quantity.

Variables:

  • save_memory_for_chi0q (bool) – Use the per-q einsum bubble instead of the FFT bubble.

  • save_memory_for_chiq_aux (bool) – Use the per-q auxiliary-susceptibility path and per-rank BZ mapping.

  • save_memory_for_sde (bool) – Use the q-loop self-energy contraction instead of the FFT one.

  • save_memory_for_fq (bool) – Use the per-q full-vertex construction in the Eliashberg step.

  • save_memory_for_lanczos (bool) – Use the frequency-distributed Lanczos solver.

class dgamore.config.OutputConfig[source]#

Bases: object

Stores the output paths.

Variables:

  • output_path (str) – Directory where the main results are written.

  • do_plotting (bool) – Whether to produce plots (rank 0 only).

  • plotting_path (str) – Directory where plots are written.

  • plotting_subfolder_name (str) – Subfolder name (under plotting_path) for the plots.

  • eliashberg_path (str) – Directory where Eliashberg results are written.

class dgamore.config.SelfConsistencyConfig[source]#

Bases: object

Stores the self-consistency-loop parameters: the maximum iteration count, the convergence criterion, the mixing parameter/scheme and continuation options. If previous_sc_path is set, the loop resumes from a previous run. The mixing scheme can be "linear", "pulay" or "anderson" (the latter two use an iteration history).

Variables:

  • max_iter (int) – Maximum number of self-consistency iterations.

  • epsilon (float) – Relative-residual convergence threshold on the self-energy.

  • mixing (float) – The mixing parameter \(\alpha\).

  • mixing_strategy (str) – The mixing scheme ("linear", "pulay" or "anderson").

  • mixing_history_length (int) – Number of past iterations used by the accelerated mixing schemes.

  • previous_sc_path (str) – Path to a previous self-consistency run to resume from (empty to start fresh).

  • use_interpolated_sigma (bool) – Whether to resume from the interpolated rather than the raw self-energy.

  • use_lambda_correction (bool) – Whether the self-consistency loop applies the lambda correction.

  • restrict_chi_phys (bool) – Whether to restrict the physical susceptibility to positive values.

  • anderson_prev_res (float | None) – Cached previous Anderson residual (internal use).

class dgamore.config.SelfEnergyInterpolationConfig[source]#

Bases: object

Stores the self-energy interpolation parameters (re-gridding to a different temperature/frequency box).

Variables:

  • do_interpolation (bool) – Whether to additionally save an interpolated self-energy each iteration.

  • beta_target (float) – Target inverse temperature \(\beta\) of the interpolation.

  • niv_target (int) – Target number of positive fermionic frequencies of the interpolation.

class dgamore.config.SystemConfig[source]#

Bases: object

Stores the physical system parameters and derived quantities updated during the run.

Variables:

  • beta (float) – Inverse temperature \(\beta\).

  • mu (float) – Chemical potential \(\mu\) (updated during self-consistency).

  • mu_dmft (float) – Chemical potential of the DMFT input.

  • n (float) – Total filling \(n\).

  • n_bands (int) – Number of bands.

  • occ (ndarray) – Local (k-averaged) occupation matrix.

  • occ_k (ndarray) – Full (k-resolved) occupation matrix.

  • occ_dmft (ndarray) – Local occupation matrix from the DMFT input.

  • occ_dmft_per_ineq (list) – DMFT occupation matrices per inequivalent atom.

dgamore.config.logger: DgaLogger#