GitLocker: The Coding Marketplace

Description:

phantomconfig 0.3.4

Phantom config

phantom-config: parse, convert, modify, and generate Phantom config files

phantom-config can read Phantom .in and .setup files. (They have the same format.) You can, for example:

modify config values or comment strings,
add new variables or delete old ones,
write the config to a JSON or TOML file,
generate a config file from a Python dictionary.

Installation
Install phantom-config with pip
python -m pip install phantomconfig

Requirements
Python 3.7+ only. Optionally tomlkit for read/write to TOML format.
Usage
Basic usage
Import phantom-config.
>>> import phantomconfig

To read in a Phantom config file
>>> input_file = phantomconfig.read_config('prefix.in')

Print a summary
>>> input_file.summary()

The variables, with their values, comment string, and the block they are a member of, are stored in a dictionary accessed by the .config method.
>>> dtmax = input_file.config['dtmax']

The keys of this dictionary correspond to the variable name, and values are a ConfigVariable named tuple with the variable name, value, comment, and block.
>>> dtmax.name
>>> dtmax.value
>>> dtmax.comment
>>> dtmax.block

You can just get the value if you want.
input_file.get_value('dtmax')

If you like, you can write the Phantom config as a JSON file, and you can read the JSON file.
>>> input_file.write_json('prefix-in.json')
>>> json_file = phantomconfig.read_json('prefix-in.json')

Check that the configs are equal
>>> input_file.config == json_file.config

You can also read and write TOML files.
>>> input_file.write_toml('prefix-in.toml')
>>> toml_file = phantomconfig.read_toml('prefix-in.toml')

You can add a new variable, remove a variable, and change the value of a variable.
# Add new variable
>>> input_file.add_variable(
... 'new_var',
... 12345678,
... comment='Sets thing',
... block='options controlling things',
... )

# Remove a variable
>>> input_file.remove_variable('dtmax')

# Change the value of a variable
>>> input_file.change_value('dumpfile', 'new_dumpfile_name')

Then you can write the Phantom config file with the modified values.
>>> input_file.write_phantom('new.in')

Examples
Generate a config from a dictionary
You can create a Phantom .setup file from a Python dictionary. First create the dictionary
>>> setup = dict()
>>> setup['gas properties'] = {
... 'cs': (cs, 'sound speed'),
... 'npart': (npart, 'number of particles in x direction'),
... 'rhozero': (rhozero, 'initial density'),
... 'ilattice': (ilattice, 'lattice type'),
... }

Then you can read the dictionary with phantomconfig, and write to a Phantom .setup file
>>> setup_config = phantomconfig.read_dict(setup)
>>> setup_config.header = [
... 'input file for some particular setup routine',
... 'short description of what it does',
... ]
>>> setup_config.write_phantom('filename.setup')

This writes a file like
# input file for some particular setup routine
# short description of what it does

# gas properties
cs = 1.000 ! sound speed
npart = 9999 ! number of particles in x direction
rhozero = 0.100 ! initial density
ilattice = 2 ! lattice type

Writing multiple configs
Say you want to write multiple configs, each with a different parameter value. For example, you have a template .in file and you want to vary the alpha parameter. The following

reads the template file
loops over a list of alpha values, writing a new .in file for each value in the list

alphas = [0.1, 0.2, 0.3]
infile = phantomconfig.read_config('template.in')

for alpha in alphas:
infile.change_value('alpha', alpha)
infile.write_phantom(f'alpha={alpha}.in')

See also
phantom-build
phantom-build is a Python package designed to make it easy to generate reproducible Phantom builds for writing reproducible papers. You can generate .in and .setup files with phantom-config and then, with phantom-build, you can compile Phantom and set up multiple runs, and schedule them via, for example, the Slurm job scheduler.
phantom-setup
phantom-setup is a (work in progress) Python package designed to set up Phantom initial conditions in pure Python, i.e. with no Fortran dependencies. It uses NumPy and Numba to achieve Fortran like performance for computationally expensive operations.