neutrinoceros / yt_idefix

GNU General Public License v3.0
5 stars 5 forks source link

yt_idefix

PyPI yt-project

CI pre-commit.ci status Ruff

A maturing yt frontend for Idefix and Pluto, packaged as an extension for yt.

Installation

python -m pip install yt_idefix

Supported formats

Code format supported since additional dependencies
Idefix .dmp v0.1.0
Idefix .vtk v0.3.0
Pluto .vtk v0.9.0
Pluto XDMF v1.1.0 h5py

Usage

Integration with yt is seamless. Installing this plugin is all that's required to make yt compatible with data formats supported by yt_idefix !

Additional arguments to yt.load

The metadata are parsed from data file, definitions header file and inifile when loading dataset.

Definitions header file (definitions.h for Pluto, or definitions.hpp for Idefix) and inifile (pluto.ini and idefix.ini respectively) are discovered automatically if they match default names, are located along with data files, and unique. Otherwise, they can be specified explicitly as paths (either relative to data files or absolute paths) with parameters definitions_header and inifile respectively.

ds = yt.load(
    "data.0010.vtk",
    definitions_header="../definitions.h",
    inifile="example.ini",
)

Geometry is parsed automatically whenever possible, but as a last resort, it can also be specified as a keyword argument (possible values are "cartesian", "spherical", "cylindrical" and "polar").

ds = yt.load("data.0010.vtk", geometry="spherical")

The data are loaded as physical quantities with units. The default unit system is cgs in yt. Data is always interpreted as dimensionful. For Pluto simulation, metadata is retrieved from definitions.h and pluto.ini to guess the proper on-disk units automatically.

Units may also be provided at runtime using the units_override argument

ds = yt.load(
    "data.0010.vtk",
    units_override={
        "length_unit": (100.0, "au"),
        "mass_unit": yt.units.mass_sun,
    },
)

Note that other units will also be changed for consistency (Pluto).

Displayed units can also be controled using the unit_system argument. Accepted values are "cgs" (default), "mks" and "code".

ds = yt.load("data.0010.vtk", unit_system="mks")

With Pluto data, units not specified with units_override will be derived consistently with given units, within the following rules:

  1. Temperature unit cannot be overridden (always set to Kelvin)
  2. No more than three units can be overridden at once (overconstrained systems are never validated for simplicity)
  3. When given less than three overrides, base units in Pluto (ordered: velocity_unit, density_unit, length_unit) are assumed
  4. The following combinations are not allowed
{"magnetic_unit", "velocity_unit", "density_unit"},
{"velocity_unit", "time_unit", "length_unit"},
{"density_unit", "length_unit", "mass_unit"}

yt is able to provide some derived fields from existed fields, e.g., "cell_volume". Fields related to element species can be created according to primordial abundances of H and He, through default_species_fields ("neutral" and "ionized") parameters.

ds = yt.load("data.0010.vtk", default_species_fields="ionized")

Conventions on field names

Field names of on-disk fields for density, pressure, velocity and magnetic field components are always normalized to upper case, even if Pluto may use lowercase in some versions.

>>> ds.field_list
[('pluto-vtk', 'PRS'),
 ('pluto-vtk', 'RHO'),
 ('pluto-vtk', 'VX1'),
 ('pluto-vtk', 'VX2'),
 ('pluto-vtk', 'VX3')]

This normalization is only applied to non-user-defined outputs and Pluto's ion fraction outputs.