josefkaufmann / ana_cont

Package for analytic continuation of many-body Green's functions

ana_cont: Python package for analytic continuation

Update: Documentation and resources can be found on https://josefkaufmann.github.io/ana_cont/.

Any questions, suggestions and bug reports will be received gratefully. (Mail to: josefkaufma at gmail dot com)

If you used this package to generate results for a publication, please cite this preprint by Kaufmann and Held (this bibtex file),

where some implementation details are described in the Supplementary Material.

Short description

Analytic continuation of Matsubara Green's functions is a difficult task, for which there exists a large variety of methods and codes. This package does, for now, not contain any new method, but its concept differs from most other codes: I do not provide compiled program that is able to do many different tasks, but instead a small Python library containing the necessary classes and functions. Based on this, the user is encouraged to write their own scripts, fitting their needs.

Currently, the Pade approximation and the Maximum Entropy method are implemented.

Package structure

• ana_cont contains the main code files.
• gui contains the code for the graphical user interface.
• doc contains some learning resources.
• scripts contains some simple examples how the library may be used; also the GUI executables are located there.

Requirements

The code was checked to run with the following versions: * python3-dev * Python 3.7 * numpy 1.18.1 * scipy 1.4.1 * Cython 0.29.15 (only for Pade) * matplotlib 3.1.3 * h5py 2.10.0 * PyQt5 (only for GUI)

It is likely that the code runs also with older package versions, but I cannot guarantee for that.

Installation and usage

Installation via pip

The ana_cont library can be installed via pip:

• sudo apt install python3-dev
• python3 -m pip install -U pyqt5 matplotlib wheel cython h5py scipy
• pip install -i https://test.pypi.org/simple/ ana-cont

This installs the ana_cont python library. The GUI scripts are also installed and can be called from the command line as maxent.py, pade.py, maxent_bosonic.py.

Installation from source via git

I recommend the following steps:

git clone https://github.com/josefkaufmann/ana_cont

If you want to use Pade, you have to compile: python setup.py build_ext --inplace

Now you can add the code directory to the python path in your script and import the package:

sys.path.insert(0, '/path/to/ana_cont')

import ana_cont.continuation as cont

The graphical user interface scripts can be executed as /path/to/ana_cont/scripts/maxent.py

Running tests after installation from source

python3 -m unittest tests/ana_cont_tests.py -v