scida

Scalable analysis for large astrophysical datasets

Process cosmological simulations and observational data with dask-powered parallel computing and automatic physical units.

Load any supported dataset with a single function call. Automatic type detection selects the right handler.

import scida
ds = scida.load("snapshot_099.hdf5")

Scale from your laptop to HPC clusters. All data is loaded as lazy dask arrays, computed only when needed.

masses = ds.data["PartType0"]["Masses"]
total = masses.sum().compute()  # runs in parallel

Automatic unit support via pint. Easily compare results across different simulation codes and observational surveys with consistent physical units.

ds = scida.load("snapshot_099.hdf5", units=True)
coords = ds.data["PartType0"]["Coordinates"]
coords_kpc = coords.to("kpc")

Native support for HDF5, zarr, and FITS. Works with AREPO, SWIFT, GIZMO, Gadget, and more out of the box.