Skip to main content

Python package to load PLUTO simulation data

Project description

plutoplot - Plutocode data loading and plotting library

GitHub release PyPI Documentation pre-commit testing License Code style: black Imports: isort

plutoplot is a Python package for loading, analyzing and plotting data from the astrophysical gas dynamics simulation code PLUTO. By handling simulations as a whole and loading data only when needed (lazy loading), it is fast and easy to use.

It features:

  • Loading on data actually needed (by lazy loading in memory mapping)
  • easy slicing into large datasets (e.g. plotting 2D slice of 3D dataset)
  • plotting non-cartesian coordinates projected into cartesian space
  • interactive plotting in Jupyter Notebooks
  • CLI tool to quickly acquire information about simulations
  • Access to parameters set in definitions.h and pluto.ini
  • Access to simulation metadata and grid properties
  • Indexing with ijk (like pyPLUTO) and kji (like PLUTO internally) schemes

There is currently no support for:

  • Particles
  • Adaptive grids (AMR)
  • 3D plotting (only plotting 2D slices of 3D data is supported.)

To get started with plutoplot check out the documentation and examples.

Installation

plutoplot requires Python >=3.6. Make sure for installation to install in the right environment, and/or adapt the pip commands accordingly (e.g. pip3).

The base version of plutoplot can be installed from Pypi with:

pip install --upgrade plutoplot

To keep the installation size small the dependencies for interactive plotting and loading of HDF5 data are declared as optional dependencies. To install the optional dependencies use either of the following, depeding on your needs:

pip install --upgrade plutoplot[interactive]
pip install --upgrade plutoplot[hdf5]
pip install --upgrade plutoplot[interactive,hdf5]

To install the current development version from this branch:

pip install --upgrade "plutoplot @ git+https://github.com/simske/plutoplot"

Quick Start

CLI: check simulation info

To quickly check the status of a PLUTO simulation, you can use the pluto-info command line tool:

pluto-info /path/to/simulation

Sample:

>> pluto-info test-problems/HD/Jet/01/
PLUTO simulation at 'test-problems/HD/Jet/01'
Polar grid with dimensions (160, 1, 640)
Domain: x1: 0.00e+00 .. 1.00e+01 (Lx1 = 1.00e+01)
        x2: 0.00e+00 .. 1.00e+00 (Lx2 = 1.00e+00)
        x2: 0.00e+00 .. 4.00e+01 (Lx3 = 4.00e+01)
Available variables: rho vx1 vx2 vx3 prs
Data files:
    Format dbl: 16 files, last time 15.0, data timestep 1.00e+00
    Format flt: 61 files, last time 15.0, data timestep 2.50e-01

(note: depending on your Python environment, the entrypoint pluto-info might not be installed correctly. Check the documentation for more details).

Quick start in Python scripts / Jupyter notebooks

import plutoplot as pp

Basics

Simulations can be loaded by just providing the path to the simulation directory (directory with definitions.h and pluto.ini). The data directory is the found automatically, as well as the grid geometry. The formats are searched in the order (dbl, flt, vtk, dbl.h5, flt.h5) and the first found loaded, this can be overriden.

sim = pp.Simulation("/path/to/simulation", format="flt")

The simulation object provides access to all outputs (of this format), the grid and simulation metadata. A few examples of accesible attributes:

sim.dims                  # domain resolution
sim.x1, sim.x2, sim.x3    # cell centered coordinates, redirect to sim.grid.x{1,2,3}
sim.x1i, sim.x2i, sim.x3i # cell interface coordinates
sim.Lx1, sim.Lx2, sim.Lx3

sim.t                     # simulation output times
sim.vars                  # simulation output variables

The data at the specific output steps can be accessed with the sim[n] syntax. This results in a PlutoData object, which can be used to access the data:

sim[-1].rho

The data object as well as the data array are generated when needed, and the data arrays are memory mapped, which means only data actually accessed is read from disk. The names of the data variables can be found with sim.vars.

Be aware that data arrays are always 3-dimensional, even if the domain has a resolution of 1 in some directions. This is for clarity and consistency, so that an axis always has a direct correspondence to its coordinates in the grid.

Simulations are iterable:

for step in sim:
    print("Total mean squared velocity at t={step.t}: ", (step.vx1**2 + step.vx2**2 + step.vx3**2).mean())

And also provide a reduce() function to run over the whole simulation:

prs_mean = sim.reduce(lambda step: step.prs.mean())

This results in an array containing the output of the supplied function for each step, in the example the mean pressure.

Working with simulation slices

To only look at a part of a simulation we can slice into it. This can be particularly useful for analyzing and plotting 2D slices of 3D data, as full 3D plotting is not supported yet. To slice into a simulation the usual slicing syntax on Simulation.slicer:

sliced_sim = sim.slicer[:,0,32:64]

This sliced simulation object can now be used as the normal simulation object. The grid and data arrays will automatically return the data for this slice. (Note: for clarity a 3D slice is always required. The data arrays will still always be 3D, see above).

Plotting

plutoplot provides automatic plotting for 1D & 2D simulations. To plot slices of 3D simulations, first create a simulation slice (section above), then just follow normally.

The PlutoData object provides a plot() function, which will automatically project the data onto a cartesian grid, and annotating the axes and colorbar correctly:

sim[-1].plot('rho')

Density plot

The projection can also be turned off, an arbitrary options which will be passed to matplotlib can be given:

sim[-1].plot('rho', projection=False)

Density plot

If a plot of a custom quantitity is needed, a array instead of a variable name can be given:

step = sim[-1]
step.plot(step.prs / step.rho, label="T")

In Jupyter Notebooks interactive plotting can be used to step through the simulation steps. Simulation.iplot() has the same interface as PlutoData.plot(), but it will show a slider to step through simulation steps:

sim.iplot("rho")

Interactive plot

Testing

The test suite for plutoplot can be found in test/ and is run by Github actions using tox and pytest.

The test suite is currently minimal and has low code coverage. It will be expanded in the future to ensure correctness.

Contributing

If you have any questions, suggestions or find a bug, feel free to open an issue or Pull Request, I am happy to accept contributions to make plutoplot better.

plutoplot uses pre-commit to ensure a consistend code style, please use it when opening a Pull Request.

License

GPLv3

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

plutoplot-0.3.2.tar.gz (23.1 kB view hashes)

Uploaded Source

Built Distribution

plutoplot-0.3.2-py3-none-any.whl (23.1 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page