def test_array_conversion(path, model):
mmtf_file = mmtf.MMTFFile.read(path)
try:
a1 = mmtf.get_structure(mmtf_file, model=model, include_bonds=True)
except biotite.InvalidFileError:
if model is None:
# The file cannot be parsed into an AtomArrayStack,
# as the models contain different numbers of atoms
# -> skip this test case
return
else:
raise
mmtf_file = mmtf.MMTFFile()
mmtf.set_structure(mmtf_file, a1)
temp = TemporaryFile("w+b")
mmtf_file.write(temp)
temp.seek(0)
mmtf_file = mmtf.MMTFFile.read(temp)
temp.close()
a2 = mmtf.get_structure(mmtf_file, model=model, include_bonds=True)
for category in a1.get_annotation_categories():
assert a1.get_annotation(category).tolist() == \
a2.get_annotation(category).tolist()
assert a1.coord.flatten().tolist() == \
approx(a2.coord.flatten().tolist(), abs=1e-3)
assert a1.bonds == a2.bonds
if a1.box is not None:
assert np.allclose(a1.box, a2.box)
def test_extra_fields():
path = join(data_dir, "1l2y.mmtf")
mmtf_file = mmtf.MMTFFile()
mmtf_file.read(path)
stack1 = mmtf.get_structure(
mmtf_file,
extra_fields=[
"atom_id", "b_factor", "occupancy", "charge"
]
)
mmtf_file == mmtf.MMTFFile()
mmtf.set_structure(mmtf_file, stack1)
stack2 = mmtf.get_structure(
mmtf_file,
extra_fields=[
"atom_id", "b_factor", "occupancy", "charge"
]
)
assert stack1.atom_id.tolist() == stack2.atom_id.tolist()
assert stack1.b_factor.tolist() == approx(stack2.b_factor.tolist())
assert stack1.occupancy.tolist() == approx(stack2.occupancy.tolist())
assert stack1.charge.tolist() == stack2.charge.tolist()
def test_array_conversion(path, single_model):
model = 1 if single_model else None
mmtf_file = mmtf.MMTFFile()
mmtf_file.read(path)
a1 = mmtf.get_structure(mmtf_file, model=model, include_bonds=True)
mmtf_file = mmtf.MMTFFile()
mmtf.set_structure(mmtf_file, a1)
a2 = mmtf.get_structure(mmtf_file, model=model, include_bonds=True)
for category in a1.get_annotation_categories():
assert a1.get_annotation(category).tolist() == \
a2.get_annotation(category).tolist()
assert a1.coord.flatten().tolist() == \
approx(a2.coord.flatten().tolist(), abs=1e-3)
assert a1.bonds == a2.bonds
def test_array_conversion(path, single_model):
model = 1 if single_model else None
mmtf_file = mmtf.MMTFFile()
mmtf_file.read(path)
a1 = mmtf.get_structure(mmtf_file, model=model, include_bonds=True)
mmtf_file = mmtf.MMTFFile()
mmtf.set_structure(mmtf_file, a1)
temp_file_name = biotite.temp_file("mmtf")
mmtf_file.write(temp_file_name)
mmtf_file = mmtf.MMTFFile()
mmtf_file.read(temp_file_name)
a2 = mmtf.get_structure(mmtf_file, model=model, include_bonds=True)
for category in a1.get_annotation_categories():
assert a1.get_annotation(category).tolist() == \
a2.get_annotation(category).tolist()
assert a1.coord.flatten().tolist() == \
approx(a2.coord.flatten().tolist(), abs=1e-3)
assert a1.bonds == a2.bonds
if a1.box is not None:
assert np.allclose(a1.box, a2.box)
# several multiples.
# The usage is similar to :class:`PDBxFile`: The :class:`MMTFFile` class
# decodes the file and makes it raw information accessible.
# Via :func:`get_structure()` the data can be loaded into an atom array
# (stack) and :func:`set_structure()` is used to save it back into a
# MMTF file.
import numpy as np
import biotite.structure.io.mmtf as mmtf
mmtf_file_path = rcsb.fetch("1l2y", "mmtf", biotite.temp_dir())
file = mmtf.MMTFFile()
file.read(mmtf_file_path)
stack = mmtf.get_structure(file)
array = mmtf.get_structure(file, model=1)
# Do some fancy stuff
mmtf.set_structure(file, array)
########################################################################
# A more low level access to MMTF files is also possible:
# An MMTF file is structured as dictionary, with each key being a
# strutural feature like the coordinates, the residue ID or the
# secondary structure. If a field is encoded the decoded
# :class:`ndarray` is returned, otherwise the dictionary value is
# directly returned.
# A list of all MMTF fields (keys) can be found in the
# `specification <https://github.com/rcsb/mmtf/blob/master/spec.md#fields>`_.
# The implementation of :class:`MMTFFile` decodes the encoded fields
# only when you need them, so no computation time is wasted on fields
# you are not interested in.
# Field is not encoded
