fem.post package

Submodules

fem.post.colormaps module

fem.post.create_disp_dat module

create_disp_dat.py - generate disp.dat binary from nodout ASCII

fem.post.create_disp_dat.correct_Enot(line)

Correct dropped ‘E’ in -??? scientific notation.

ls-dyna seems to drop the ‘E’ when the negative exponent is three digits, so check for those in the line data and change those to ‘E-100’ so that we can convert to floats

This is a helper function only called by parse_line_regex(), which is currently not being used by default.

Parameters:

line (str) – string of split raw string data

Returns:

line with corrected -??? -> -E100

fem.post.create_disp_dat.correct_neg(line)

Add space before negative coefficient numbers.

This is a helper function only called by parse_line_regex(), which is currently not being used by default.

Parameters:

line (str) – raw read line

Returns:

line with space(s) added before negative coefficients

Return type:

line (str)

fem.post.create_disp_dat.count_timesteps(outfile: str = 'nodout') → int

Count timesteps in to outfile.

Searches for ‘time’ in lines, and then removes 1 extra entry that occurs for t = 0. grep will be used on linux systems (way faster).

Parameters:

outfile (str) – usually ‘nodout’

Returns:

number of time steps counted - 1

Return type:

ts_count (int)

fem.post.create_disp_dat.create_dat(nodout='nodout', dispout='disp.dat', legacynodes=False)

Create binary data file from nodout ASCII file.

Parameters:

  • nodout (str) – nodout input filename

  • dispout (str) – binary output filename

  • legacynodes (Boolean) – (Default value = False)

fem.post.create_disp_dat.generate_header(data, outfile)

generate headers from data matrix of first time step

Parameters:

  • data (nparray) – data

  • outfile (str) – output filename to count times from

Returns:

Return type:

header (dict)

fem.post.create_disp_dat.main()
fem.post.create_disp_dat.open_dispout(dispout)

Open dispout file for writing.

Parameters:

dispout (str) – dispout filename (disp.dat.xz)

Returns:

file object

Return type:

dispout (obj)

fem.post.create_disp_dat.parse_cli()

parse command-line interface arguments

fem.post.create_disp_dat.parse_line(line)

Parse raw data line into vector of floats.

Parameters:

line (str) – raw data line from nodout

Returns:

[nodeID, xdisp, ydisp, zdisp]

Return type:

raw_data (list of floats)

fem.post.create_disp_dat.parse_line_regex(line)

Parse raw data line into list of floats using regex.

This regex approach works, but is very slow!! It also requires two helper functions to clean up malformed data written by ls-dyna (done on purpose, probably to save space).

Parameters:

line (str) – raw data line from nodout

Returns:

[nodeID, xdisp, ydisp, zdisp]

Return type:

raw_data (list of floats)

fem.post.create_disp_dat.process_timestep_data(data, outfile, writenode)

write data for the entire timestep to outfile

Parameters:

  • data (list) – list of all data, with each entry containing [nodeID, xdisp, ydisp, zdisp]

  • outfile (obj) – output file object

  • writenode (Boolean) – Boolean if the node IDs should be written to save ~25% of the disp.dat file size

Returns:

None

fem.post.create_disp_dat.write_headers(outfile, header)

Write binary header.

Write binary header information to reformat things on read downstream.

Parameters:

  • outfile (str) – output file object

  • header (dict) –

Returns:

None

fem.post.create_res_sim module

create_res_sim.py - extract data for display in different formats from disp.dat binary

fem.post.create_res_sim.create_zdisp(nodeidlist, disp_slice_z_only, zdisp)

create zdisp array from squeezed disp_slice at appropriate index

Parameters:

  • nodeidlist – first column of disp_slice with node IDs in row order

  • disp_slice_z_only – squeezed disp_slice of just zdisp

  • zdisp –

Returns:

rows corresponding to node ID

Return type:

zdisp (ndarray)

fem.post.create_res_sim.extract3Darfidata(dynadeck='dynadeck.dyn', disp_comp=2, disp_scale=- 10000.0, ressim='res_sim.h5', nodedyn='nodes.dyn', dispout='disp.dat', specific_times=None)

Extract 3D volume of specified displacement component.

Parameters:

  • dynadeck (str) – LS-DYNA3D input deck (used to get dt)

  • disp_comp (int) – displacement component to extract (0, 1, 2)

  • disp_scale (float) – displacement scaling factor (cm -> um)

  • ressim (str) – output file name [.mat, .h5, .pvd]

  • nodedyn (str) – node input file

  • dispout (str) – binary displacement data

  • specific_times (list) – optional list of specific time indices

fem.post.create_res_sim.extract_arfi_data(dispout, header, image_plane, disp_comp=2, disp_scale=- 10000.0, legacynodes=False, specific_times=None)

extract ARFI data from disp.dat

Parameters:

  • dispout (str) – name of disp.dat file

  • header (dict) – num_nodes, num_dims, num_timesteps

  • image_plane (ndarray) – matrix of image plane node IDs spatially sorted

  • disp_comp (int) – disp component index to extract (0, 1, 2 [default, z])

  • disp_scale (float) – cm -> um

  • legacynodes (Boolean) – node IDs repeated every timestep

  • specific_times (list) – optional list of specific time indices to extract

Returns:

Return type:

arfidata (ndarray)

Raises:

IndexError – unexpected arfidata dimensions

fem.post.create_res_sim.extract_dt(dyn_file)

extract time step (dt) from dyna input deck (assume comma-delimited)

Parameters:

dyn_file (str) – input.dyn filename

Raises:

FileNotFoundError –

Returns:

timestep interval

Return type:

dt (float)

fem.post.create_res_sim.extract_image_plane(snic, axes, plane_pos: float = 0.0, direction: int = 0)

extract 2D imaging plane node IDs

Extract a 2D matrix of the imaging plane node IDs based on the elevation position (mesh coordinates).

Parameters:

  • snic – sorted node IDs and coordinates

  • axes – spatial axes

  • plane_pos (float) – position of the plane to extract (in the specified plane_orientation)

  • direction (int) – orientation plane to extract from (0 = elev, 1 = lateral, 2 = axial)

Raises:

  • TypeError – deprecated ele_pos passed as keyword argument

  • ValueError – invalid direction axis specified

Returns:

image_plane (node IDs)

fem.post.create_res_sim.main()
fem.post.create_res_sim.open_dispout(dispout)

open dispout file for reading, potentially using lzma

Parameters:

dispout (pathlib / str) – dispout file

Raises:

  • FileNotFoundError – disp.dat[.xz] file cannot be found

  • ImportError – LZMA package not available

Returns:

open file object

Return type:

dispout (obj)

fem.post.create_res_sim.read_header(dispout, word_size_bytes: int = 4)

Read header (first 3 words) from dispout

Parameters:

  • dispout – disp.dat filename

  • word_size_bytes (int) – 4

Returns:

keys: num_nodes, num_dims, num_timesteps

Return type:

header (dict)

fem.post.create_res_sim.run(dynadeck, disp_comp=2, disp_scale=- 10000.0, ressim='res_sim.mat', nodedyn='nodes.dyn', dispout='disp.dat', legacynodes=False, plane_pos: float = 0.0, plane_orientation: int = 0, specific_times: Optional[list] = None)

helper function to run high-level, 2D plane extraction

look at using extract3Darfidata to get full, 3D datasets exported (e.g., to view in Paraview)

Parameters:

  • dynadeck (str) – main dyna input deck

  • disp_comp (int) – component of displacement to extract

  • disp_scale (float) – displacement scaling

  • ressim (str) – result filename to write

  • nodedyn (str) – node defintion input filename

  • dispout (str) – binary displacement input filename

  • legacynodes (Boolean) – node IDs written with each timestep in dispout

  • plane_pos (float) – position of the plane to extract (in the specified plane_orientation)

  • plane_orientation (int) – orientation plane to extract from (0 = elev, 1 = lateral, 2 = axial)

  • specific_times (list) – optional list of specific time indices to extract

fem.post.create_res_sim.save_res_sim(resfile, arfidata, axes, t, axis_scale=(- 10, 10, - 10), plane_pos=None, plane_orientation=None)

Save res_sim.[mat,h5,pvd] file with arfidata and relevant axes.

Data are saved as float32 (single) to save space.

Parameters:

  • resfile (str) – res_sim.[mat,h5,pvd] filename

  • arfidata (ndarray) – arfidata (3D or 4D (added elev dim, axes[0]))

  • axes (ndarrays tuple) – ele, lat, axial (mesh units)

  • t (ndarray) – time

  • axis_scale (floats tuple) – scale axes sign & mag

  • plane_pos (float) – position of the plane to extract (in the specified plane_orientation)

  • plane_orientation (int) – orientation plane to extract from (0 = elev, 1 = lateral, 2 = axial)

Raises:

  • ValueError – Cannot save 2D PVD timeseries data.

  • KeyError – Trying to save unknown output type.

fem.post.create_res_sim.saveh5(**kwargs)

Save HDF5 file with gzip compression.

Parameters:

  • arfidata (float) – 4D arfidata matrix

  • axial (float) – depth axis vector [mm]

  • lat (float) – lateral axis vector [mm]

  • elev (float) – elevation axis vector [mm]

  • t (float) – time vector (s)

  • resfile (str) – ‘res_sim.pvd’

fem.post.create_res_sim.savemat(**kwargs)

Save Matlab v5 file.

Parameters:

  • arfidata (float) – 4D arfidata matrix

  • axial (float) – depth axis vector [mm]

  • lat (float) – lateral axis vector [mm]

  • elev (float) – elevation axis vector [mm]

  • t (float) – time vector (s)

  • resfile (str) – ‘res_sim.pvd’

Raises:

TypeError – arfidata >4 GB

fem.post.create_res_sim.savepvd(ts_start=0, part=0, **kwargs)

Save Paraview PVD rectilinear (VTR) time series data.

Paraview data formats: https://www.paraview.org/Wiki/ParaView/Data_formats

Parameters:

  • ts_start (int) – override starting timestep index from 0 to this value

  • arfidata (float) – 4D arfidata matrix

  • axial (float) – depth axis vector [mm]

  • lat (float) – lateral axis vector [mm]

  • elev (float) – elevation axis vector [mm]

  • t (float) – time vector (s)

  • resfile (str) – ‘res_sim.pvd’

Raises:

  • ValueError – Not saving 3D time series data.

  • FileExistsError – PVD file directory cannot be created.

fem.post.h52vtr module

class fem.post.h52vtr.H52VTR(h5file=None, axisnames=('elev', 'lat', 'axial'), dataname='arfidata', vtrname='rectilinear')

Bases: object

Convert HDF5 file to rectilinear VTR file.

Parameters:

  • h5file (str) – input HDF5 filename

  • axisnames (str) – (‘elev’, ‘lat’, ‘axial’)

  • dataname (str) – ‘arfidata’

  • vtrname (str) – ‘rectilinear’

read_h5axes()

Read axes from HDF5 file.

read_h5data()

Read data in from HDF5 file.

write_vtr()

Write rectilinear VTR file.

fem.post.listedcmap2xml module

fem.post.listedcmap2xml.listedcmap2xml(cmap)

Convert ListedColormap to XML format for import into Paraview

Based on: https://www.paraview.org/Wiki/Colormaps

Parameters:

ListedColormap (ListedColormap) – Matplotlib-formated ListedColormap

fem.post.ressim module

class fem.post.ressim.ResSim(filename='res_sim.mat')

Bases: object

plot and animate res_sim.mat simulation data

filename

name of file to load in (MATv5 or HDF5)

Type:

str

arfidata

arfidata

Type:

float ndarray

axial

depth

Type:

float ndarray

lat

lateral

Type:

float ndarray

t

time

Type:

float ndarray

time_scale_factor

time scale factor (e.g., s -> ms)

Type:

float

disp_scale_factor

displacement scale factor

Type:

float

animate(i)
load()

load MATv5 or HDF5 data

Parameters:

filename (str) – input filename

Raises:

TypeError – MATv5 or HDF5 load exceptions

play(timerange, xlabel='Lateral (mm)', ylabel='Axial (mm)', show=True)

play an animation

Strongly recommend not stepping though each timesteps; use some skips!

Parameters:

  • timerange (range) – range generator of time steps to animate

  • xlabel (str) –

  • ylabel (str) –

  • show (Boolean) – show animation

Todo

  • Add save option

plot(timestep, show=True, save=False, savename='file', xlabel='Lateral (mm)', ylabel='Axial (mm)', title=None)

Plot arfidata at specified timestep.

Parameters:

  • timestep (int) –

  • show (Boolean) – show plot

  • save (Boolean) – save PNG (objectname_timestep.png)

  • savename (str) – saved PNG filename prefix

  • xlabel (str) –

  • ylabel (str) –

  • title (str) –

timeplot(axial, lat, xlabel='Time (ms)', ylabel='Displacement (\\mum)', title=None, show=True)

plot arfidata through time at specified ax and lat position (mm)

Parameters:

  • axial (ndarray) – axial depth (mm)

  • lat (ndarray) – lateral position (mm)

  • xlabel (str) –

  • ylabel (str) –

  • title (str) –

  • show (Boolean) – show the plot

Returns:

fem.post.savevtk module

class fem.post.savevtk.SaveVTK(data, origin, spacing)

Bases: object

Save structured 3D scalar and vector data in legacy ASCII VTK format.

Parameters:

  • data (ndarray) – [scalar] 3D numpy array [vector] tuple of 3D numpy arrays (X, Y, Z)

  • origin (float) – mesh origin tuple (x, y, z)

  • spacing (float) – node spacing tuple (x, y, z)

Raises:

IndexError – The shape of the data tuple ndarrays are not equal.

property data
save_scalar(filename, dataname='scalars', header_comment=None)

save 3D scalar data in VTK structure points format

Parameters:

  • filename (str) – output filename

  • dataname (str) – data name

  • header_comment (str) – optional header comment in output file

save_vector(filename, dataname='vectors', header_comment=None)

save 3D vector array in VTK structure points format

Parameters:

  • filename (str) – output filename

  • dataname (str) – data name

  • header_comment (str) – optional header comment in output file

Module contents