Skip to content

Commit

Permalink
support FFT ovf files from mumax
Browse files Browse the repository at this point in the history 
dtype=np.complex64 is very experimental.
Converts complex ovf files into dataset
  • Loading branch information
@timvgl
timvgl authored Nov 27, 2024
1 parent 64c81a4 commit fb95074
Showing 1 changed file with 100 additions and 39 deletions.
139 changes: 100 additions & 39 deletions src/mumaxXR/OvfEngine.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -559,6 +559,9 @@ def read_mesh_from_ovf(self, filename):
ynodes = int(footer_dict[b"ynodes"])
znodes = int(footer_dict[b"znodes"])

if self.dtype == np.complex64:
xnodes = int((xnodes / 2 + 1))


nodes = np.array([xnodes, ynodes, znodes])
xmin = float(footer_dict[b"xmin"])
Expand All @@ -582,6 +585,7 @@ def read_mesh_from_ovf(self, filename):
n_comp = int(footer_dict[b'valuedim'])
else:
n_comp = 1

return MumaxMesh(filename, nodes, world_min, world_max, tmax, n_comp, footer_dict)

def get_corresponding_files(self, filename, returnTData=False, type='', returnMeshData=True) -> Union[Tuple[list, np.ndarray], list, np.ndarray]:
Expand Down Expand Up @@ -702,28 +706,43 @@ def find_nearest(timeList: list, value: float) -> tuple:
def read_data_from_ovf(self, filename) -> np.ndarray:
filename = Path(filename)
with self.lock, open(filename, "rb") as file:
if not _binary == 4:
raise ValueError('Error: Unknown binary type assigned for reading in .ovf2 data')
if self.dtype == np.complex64:
_binary = 8
else:
_binary = 4

data_start_pos = 0

for i in range(0, 46):
line = file.readline()
if b'# Begin: Data' in line:
# the first binary number is a control number specified in the ovf format
file.read(_binary)
file.read(4)
data_start_pos = file.tell()
break
if (i == 45):
raise ValueError("Error: %s has no well formatted data segment." % filename)

file.seek(data_start_pos, 0)
size = int(self.mesh.n_comp * self.mesh.number_of_cells * _binary / 4)
data = np.fromfile(file, dtype=self.dtype, count=size)
if (self.sc == False):
data = data.reshape(self.mesh.nodes[2], self.mesh.nodes[1], self.mesh.nodes[0], self.mesh.n_comp)
if _binary == 8:
size = int(self.mesh.n_comp * self.mesh.number_of_cells * 2)
data = np.fromfile(file, dtype=np.float32, count=size)
else:
data = data.reshape(self.mesh.nodes[2], self.mesh.nodes[1], self.mesh.nodes[0])
size = int(self.mesh.n_comp * self.mesh.number_of_cells)
data = np.fromfile(file, dtype=self.dtype, count=size)

if _binary == 8:
if (self.sc == False):
data = data.reshape(self.mesh.nodes[2], self.mesh.nodes[1], int(self.mesh.nodes[0]), self.mesh.n_comp, 2)
else:
data = data.reshape(self.mesh.nodes[2], self.mesh.nodes[1], int(self.mesh.nodes[0]), 2)
data = data[...,0] + 1j*data[...,1]
else:
if (self.sc == False):
data = data.reshape(self.mesh.nodes[2], self.mesh.nodes[1], self.mesh.nodes[0], self.mesh.n_comp)
else:
data = data.reshape(self.mesh.nodes[2], self.mesh.nodes[1], self.mesh.nodes[0])

return data


Expand Down Expand Up @@ -782,37 +801,79 @@ def open_dataset(

dataSc = xr.core.indexing.LazilyIndexedArray(backend_array_sc)

notOneDims = ['x', 'y', 'z', 'comp']
defaultChunks = {}
if (backend_array.shape != ()):
for dim in backend_array.dims:
if (dim not in notOneDims):
defaultChunks[dim] = 1
defaultChunks['z'] = backend_array.coords[-4].size
defaultChunks['y'] = backend_array.coords[-3].size
defaultChunks['x'] = backend_array.coords[-2].size
defaultChunks['comp'] = backend_array.coords[-1].size

if (backend_array_sc.shape != ()):
defaultChunksSc = {}
for dim in backend_array_sc.dims:
if (dim not in notOneDims):
defaultChunksSc[dim] = 1
defaultChunksSc['z'] = backend_array_sc.coords[-3].size
defaultChunksSc['y'] = backend_array_sc.coords[-2].size
defaultChunksSc['x'] = backend_array_sc.coords[-1].size
if (backend_array.shape != ()):
var = xr.Variable(dims=backend_array.dims, data=data)
var.encoding["preferred_chunks"] = defaultChunks
if (backend_array_sc.shape != ()):
varSc = xr.Variable(dims=backend_array_sc.dims, data=dataSc)
varSc.encoding["preferred_chunks"] = defaultChunksSc
if (backend_array_sc.shape != () and backend_array.shape != ()):
dataset = xr.Dataset({'raw': var, 'rawSc': varSc}, coords=dict(zip(backend_array.dims + ['wavetypeSc'], backend_array.coords + [backend_array_sc.coords[backend_array_sc.dims.index('wavetypeSc')]])))
elif (backend_array_sc.shape != () and backend_array.shape == ()):
dataset = xr.Dataset({'rawSc': varSc}, coords=dict(zip(backend_array_sc.dims, backend_array_sc.coords)))
elif (backend_array_sc.shape == () and backend_array.shape != ()):
dataset = xr.Dataset({'raw': var}, coords=dict(zip(backend_array.dims, backend_array.coords)))
if dtype == np.complex64:
notOneDims = ['k_x', 'k_y', 'k_z', 'comp']
defaultChunks = {}
if (backend_array.shape != ()):
for dim in backend_array.dims:
if (dim not in notOneDims):
defaultChunks[dim] = 1
defaultChunks['k_z'] = backend_array.coords[-4].size
defaultChunks['k_y'] = backend_array.coords[-3].size
defaultChunks['k_x'] = backend_array.coords[-2].size
defaultChunks['comp'] = backend_array.coords[-1].size

if (backend_array_sc.shape != ()):
defaultChunksSc = {}
for dim in backend_array_sc.dims:
if (dim not in notOneDims):
defaultChunksSc[dim] = 1
defaultChunksSc['k_z'] = backend_array_sc.coords[-3].size
defaultChunksSc['k_y'] = backend_array_sc.coords[-2].size
defaultChunksSc['k_x'] = backend_array_sc.coords[-1].size
replaceDims = ['x', 'y', 'z']
if (backend_array.shape != ()):
for i in range(len(backend_array.dims)):
if backend_array.dims[i] in replaceDims:
backend_array.dims[i] = 'k_' + backend_array.dims[i]
var = xr.Variable(dims=backend_array.dims, data=data)
var.encoding["preferred_chunks"] = defaultChunks
if (backend_array_sc.shape != ()):
for i in range(len(backend_array_sc.dims)):
if backend_array_sc.dims[i] in replaceDims:
backend_array_sc.dims[i] = 'k_' + backend_array_sc.dims[i]
varSc = xr.Variable(dims=backend_array_sc.dims, data=dataSc)
varSc.encoding["preferred_chunks"] = defaultChunksSc
if (backend_array_sc.shape != () and backend_array.shape != ()):
print([backend_array_sc.coords[backend_array_sc.dims.index('wavetypeSc')]])
dataset = xr.Dataset({'raw': var, 'rawSc': varSc}, coords=dict(zip(backend_array.dims + ['wavetypeSc'], backend_array.coords + [backend_array_sc.coords[backend_array_sc.dims.index('wavetypeSc')]])))
elif (backend_array_sc.shape != () and backend_array.shape == ()):
dataset = xr.Dataset({'rawSc': varSc}, coords=dict(zip(backend_array_sc.dims, backend_array_sc.coords)))
elif (backend_array_sc.shape == () and backend_array.shape != ()):
dataset = xr.Dataset({'raw': var}, coords=dict(zip(backend_array.dims, backend_array.coords)))
else:
notOneDims = ['x', 'y', 'z', 'comp']
defaultChunks = {}
if (backend_array.shape != ()):
for dim in backend_array.dims:
if (dim not in notOneDims):
defaultChunks[dim] = 1
defaultChunks['z'] = backend_array.coords[-4].size
defaultChunks['y'] = backend_array.coords[-3].size
defaultChunks['x'] = backend_array.coords[-2].size
defaultChunks['comp'] = backend_array.coords[-1].size

if (backend_array_sc.shape != ()):
defaultChunksSc = {}
for dim in backend_array_sc.dims:
if (dim not in notOneDims):
defaultChunksSc[dim] = 1
defaultChunksSc['z'] = backend_array_sc.coords[-3].size
defaultChunksSc['y'] = backend_array_sc.coords[-2].size
defaultChunksSc['x'] = backend_array_sc.coords[-1].size
if (backend_array.shape != ()):
var = xr.Variable(dims=backend_array.dims, data=data)
var.encoding["preferred_chunks"] = defaultChunks
if (backend_array_sc.shape != ()):
varSc = xr.Variable(dims=backend_array_sc.dims, data=dataSc)
varSc.encoding["preferred_chunks"] = defaultChunksSc
if (backend_array_sc.shape != () and backend_array.shape != ()):
print([backend_array_sc.coords[backend_array_sc.dims.index('wavetypeSc')]])
dataset = xr.Dataset({'raw': var, 'rawSc': varSc}, coords=dict(zip(backend_array.dims + ['wavetypeSc'], backend_array.coords + [backend_array_sc.coords[backend_array_sc.dims.index('wavetypeSc')]])))
elif (backend_array_sc.shape != () and backend_array.shape == ()):
dataset = xr.Dataset({'rawSc': varSc}, coords=dict(zip(backend_array_sc.dims, backend_array_sc.coords)))
elif (backend_array_sc.shape == () and backend_array.shape != ()):
dataset = xr.Dataset({'raw': var}, coords=dict(zip(backend_array.dims, backend_array.coords)))
dataset.attrs["cellsize"] = backend_array.mesh.cellsize
dataset.attrs["nodes"] = backend_array.mesh.nodes
dataset.attrs["min_size"] = backend_array.mesh.world_min
Expand Down

0 comments on commit fb95074

Please sign in to comment.