---

DGAmore is a Python toolbox that computes the multi-orbital, self-consistent ladder Dynamical Vertex Approximation and solves the Eliashberg equation for (strongly) correlated electron systems described by the multi-band Hubbard model. Starting from the one- and two-particle output of a dynamical mean-field theory (DMFT) calculation, it assembles the local vertex functions, solves the momentum-dependent ladder equations for the non-local self-energy and, optionally, extracts the leading superconducting eigenvalues and gap functions. It relies on vectorized numpy operations, parallelizes the heavy momentum-dependent work with mpi4py, and reads its DMFT input from HDF5 via h5py. It combines and extends the one-shot multi-orbital AbinitioDGA and the single-band DGApy, on which it is partially based.

For the implemented equations, see the author's Master's thesis (Chapters 3 and 4).

Features

• Multi-orbital ladder DGA in the density and magnetic channels, with explicit high-frequency asymptotics of the vertex functions and support for non-local interactions and several inequivalent atoms.
• A self-consistency loop for the non-local self-energy with linear, Pulay and Anderson mixing, a cooldown procedure that warm-starts low-temperature runs from converged higher-temperature ones, and three stabilization techniques for the regime close to the Bethe-Salpeter instability; the one-shot Moriya lambda correction is available as well.
• The multi-orbital Eliashberg equation, yielding the leading singlet and triplet eigenvalues and gap functions, optionally resolved by frequency parity and sorted into symmetry-adapted multiplets.
• Optional analytic continuation of the DGA and DMFT Green's functions to real frequencies with the maximum entropy method.
• MPI parallelization over the irreducible Brillouin zone with automatic symmetry discovery, shared-memory windows for replicated objects, and an upfront check that a run fits the memory of the nodes it received.

Documentation

Full documentation is hosted at dgamore.readthedocs.io.

Topic	Description
Installation	Environment setup, MPI dependencies, and installing the package.
Usage	Running the routine single-core, with MPI, and on a SLURM cluster.
Configuration	The YAML configuration file and its parameters.
Cooldown and stabilization	Reaching low temperatures by chaining runs, and the stabilization techniques of the self-consistency cycle.
Output files	Every file a run writes, its array layout, and how to read it back.
Contributing	Reporting issues and submitting pull requests.
API reference	Module-by-module reference.
About	Background, citation, license, and contact.

Quick start

DGAmore needs Python 3.12 or newer and a working MPI installation; it is tested on Linux and macOS with Python 3.12 to 3.14 (Windows is not supported). Install mpich and mpi4py, then the package:

conda install -c conda-forge mpich mpi4py
git clone https://github.com/Julpe/DGAmore.git
cd DGAmore
pip install .

To check the installation, run pytest tests from the repository directory. A calculation needs three inputs: the one- and two-particle output of a w2dynamics DMFT run (the two-particle file goes through the installed symmetrize script first), a Hamiltonian in real space (Wannier90 format) or in momentum space, and a YAML configuration file. Edit the configuration, then start the routine with -p pointing at the directory that holds it and -c naming it (defaults: the current working directory and dga_config.yaml):

mpiexec -np 8 DGAmore -p /configs/ -c my_config.yaml   # or: DGAmore for a single-core test run

See the installation and usage pages for the full instructions and an example SLURM submit script, the configuration page for every parameter, and the cooldown page before attempting low-temperature calculations.

Contributing

Contributions are welcome. Please open an issue for bugs and feature requests, or submit a pull request. See CONTRIBUTING.md for details.

Citation and license

DGAmore is released under the MIT license. If you use it, please consider citing it together with the author's Master's thesis. For questions, get in touch by e-mail.