glio

A Python package for reading and writing Gadget, and Gadget-like, data files. (glio ~ Gadget-like I/O.)

glio is released under the MIT licence (see LICENCE.md). glio is compatible with Python 2.7.x and Python 3.x, and requires numpy.

New file types may be added by defining a schema for the header and block data, and subclassing SnapshotBase (or one of its subclasses). See for example GadgetSnapshot in gadget.py and SPHRAYSnapshot in sphray.py.

To load a Gadget-2 data file,

>>> import glio
>>> s = glio.GadgetSnapshot('filename')
>>> s.load()

and similarly for the SPHRAYSnapshot class. Header data is accessible as

>>> s.header.header_item

where valid header_item attributes are defined by the current class' header schema. They can also be obtained as a list of strings, or looped through,

>>> s.header.fields
['npart', 'mass', ... ]
>>> for (name, value) in s.header.iterfields():
>>>    # Do something with header data.

Block data is accessed similarly, and can be iterated over similarly,

>>> s.ID
[array([1370534, 1370614, 1348895, ..., 1370615, 1370533, 1370531], dtype=uint32), ... ]
>>> s.fields
['pos', 'vel', 'ID', ... ]
>>> for (name, field) in s.iterfields():
>>>     # Do something with block data.

Each block data (field above) is a list, with each index i corresponding to that block's particle data for Gadget particle type i. When reading from file, if a block contains no data for a particle type i, an empty numpy.ndarray is set for that particle type in the list; if a particle type is not valid for the block, None is set.

For convenience, particle types may be aliased to readable attribute names. A dict mapping the available attributes to their corresponding particle type indices is availabe as s.ptype_aliases.

For example, for a GadgetSnapshot,

>>> s.ptype_aliases
{'star': 4, 'bulge': 3, 'gas': 0, 'boundary': 5, 'disk': 2, 'halo': 1}
>>> s.gas.pos is s.pos[0]
True
>>> s.star.vel is s.vel[4]
True

However, the alias attribute (e.g. s.gas or s.star) returns a SnapshotView object, which is read-only. Modification of the snapshot data in general requires directly modifying the field list(s).

A snapshot can be written to file, optionally with a new filename,

>>> import glio
>>> s = glio.GadgetSnapshot('filename')
>>> # Double all gas-particle smoothing lengths.
>>> s.hsml[0] *= 2
>>> s.save('new_filename')