abipy package

abipy is a set of tools for the analysis of ABINIT results.

Subpackages

• abipy.abio package
• abipy.core package
• abipy.dfpt package
• abipy.display package
• abipy.dynamics package
• abipy.electrons package
• abipy.embedding namespace
• abipy.eph package
• abipy.flowtk package
• abipy.iotools package
• abipy.lumi package
• abipy.ml package
• abipy.panels package
• abipy.ppcodes package
• abipy.scripts package
• abipy.tools package
• abipy.wannier90 package
• abipy.waves package

Submodules

abipy.abilab module

This module gathers the most important classes and helper functions used for scripting.

abipy.abilab.abiopen_ext2class_table()

Print the association table between file extensions and File classes.

abipy.abilab.extcls_supporting_panel(as_table=True, **tabulate_kwargs)

Return list of (ext, cls) tuples where ext is the file extension supporting the get_panel method i.e files that can be integrated with the panel dashboard and cls the python class associated to it.

abipy.abilab.abipanel(**kwargs)

Activate panel extensions used by AbiPy. Return panel module.

abipy.abilab.abifile_subclass_from_filename(filename)

Returns the appropriate class associated to the given filename.

Return type:

ClassVar

abipy.abilab.dir2abifiles(top, recurse=True)

Analyze the filesystem starting from directory top and return an ordered dictionary mapping the directory name to the list of files supported by abiopen contained within that directory. If not recurse, children directories are not analyzed.

Return type:

dict

abipy.abilab.isabifile(filepath)

Return True if filepath can be opened with abiopen.

Return type:

bool

abipy.abilab.abiopen(filepath)

Factory function that opens any file supported by abipy. File type is detected from the extension

Parameters:

filepath (str) – string with the filename.

abipy.abilab.abirobot(filepaths)

Factory function to create and return a Robot subclass from a list of filenames The Robot subclass is detected from the extension of the first file hence all files are assumed to have the same extension.

Parameters:

filepaths (str | list[str]) – List of strings with the filename.

Return type:

Robot

abipy.abilab.software_stack(as_dataframe=False)

Import all the hard dependencies and some optional packages. Returns ordered dict: package –> string with version info or pandas dataframe if as_dataframe.

abipy.abilab.abicheck(verbose=0)

This function tests if the most important ABINIT executables can be found in $PATH and whether the python modules needed at run-time can be imported. Return string with error messages, empty if success.

Return type:

str

abipy.abilab.check_netcdf4_for_shell_submission(manager=None)

Check whether the netCDF4/h5py packages importable in this Python process are safe to use together with a “shell” TaskManager qadapter.

Background

AbiPy’s ShellAdapter (used when manager.yml declares qtype: shell) launches the Abinit job as a direct child process of the running Python interpreter, via Popen(("/bin/bash", script_file), ...). Other adapters (slurm, pbs, sge, …) instead submit through a small helper command (sbatch, qsub, …) whose process tree is unrelated to the compute job, so they are not affected by this issue.

If netCDF4 (or h5py) was built against a parallel (MPI-enabled) HDF5/NetCDF stack, merely import netCDF4 loads the MPI runtime (libmpi/libopen-rte/…) into this process – even if the calling script never uses MPI directly (this happens transitively: abipy.iotools eagerly imports pymatgen’s netcdf reader, which eagerly imports netCDF4, so it is triggered by plain import abipy.flowtk). When AbiPy later forks mpirun out of this same process via ShellAdapter, the freshly-launched mpirun can fail immediately and silently: zero-byte log/err files, task status QCritical, no diagnostic anywhere, because it inherits/detects stale MPI runtime state from its live ancestor.

type manager:

param manager:

Optional TaskManager (or object exposing .qadapter.QTYPE). When given, the returned message is tailored to whether the active qadapter is actually “shell” (the only one at risk).

returns:

is_safe is False only if netCDF4 and/or h5py report they were built with MPI (parallel I/O) support. message explains the finding and, if unsafe, how to fix it.

rtype:

tuple[bool, str]

abipy.abilab.install_config_files(workdir=None, force_reinstall=False)

Install pre-defined configuration files for the TaskManager and the Scheduler in the workdir directory.

Parameters:

• workdir (str | None) – Directory when configuration files should be produced. Use ~/abinit/abipy/ if None

• force_reinstall (bool | None) – Allow overwrite pre-existent configuration files. By default, the function raises RuntimeError if configuration files are already present.

abipy.abilab.abipy_logo1()

http://www.text-image.com/convert/pic2ascii.cgi

Return type:

str

abipy.abilab.abipy_logo2()

http://www.text-image.com/convert/pic2ascii.cgi

Return type:

str

abipy.abilab.abipy_logo3()

http://www.text-image.com/convert/pic2ascii.cgi

Return type:

str