def connection(filename): print "[displayGraphics] : setting paraview connection..." # Connection au servermanager if not sm.ActiveConnection: connection = sm.Connect() # Lecture du fichier .case print "[displayGraphics] : reading file..." reader=sm.sources.ensight(CaseFileName = filename,ByteOrder = 1) # Ajout du contenu du fichier au servermanager reader.UpdatePipelineInformation() # Creation de la fenetre de rendu return sm.CreateRenderView(),reader
# Test the vtkCellIntegrator
from paraview import smtesting
import os
import os.path
import sys
import paraview
paraview.compatibility.major = 3
paraview.compatibility.minor = 4
from paraview import servermanager
from paraview import util
smtesting.ProcessCommandLineArguments()
servermanager.Connect()
file1 = os.path.join(smtesting.DataDir, "Data/quadraticTetra01.vtu")
reader1 = servermanager.sources.XMLUnstructuredGridReader(FileName=file1)
reader1Output = servermanager.Fetch(reader1)
# General 3D cell
integVal = util.IntegrateCell(reader1Output, 0)
if integVal < 0.128 or integVal > 0.1285:
print "ERROR: incorrect result for cell 0 of 1st dataset"
sys.exit(1)
# General 2D cell
if util.IntegrateCell(reader1Output, 1) != 0.625:
print "ERROR: incorrect result for cell 1 of 1st dataset"
sys.exit(1)
def __enter__(self):
servermanager.Connect()
return servermanager
def render_paraview_scene(
pvd_file,
outfile,
field_name='m',
timesteps=None,
camera_position=[0, -200, +200],
camera_focal_point=[0, 0, 0],
camera_view_up=[0, 0, 1],
view_size=(800, 600),
magnification=1,
fit_view_to_scene=True,
color_by_axis=0,
colormap='coolwarm',
rescale_colormap_to_data_range=True,
show_colorbar=True,
colorbar_label_format='%-#5.2g',
add_glyphs=True,
glyph_type='cones',
glyph_scale_factor=1.0,
glyph_random_mode=True,
glyph_mask_points=True,
glyph_max_number_of_points=10000,
show_orientation_axes=False,
show_center_axes=False,
representation="Surface With Edges",
palette='screen',
use_parallel_projection=False,
trim_border=True,
rescale=None,
diffuse_color=None):
"""
Load a *.pvd file, render the scene in it and save the result to an image file.
*Returns*
An IPython.core.display.Image object containing the output image.
*Arguments*
pvd_file:
Input filename (must be in *.pvd format).
outfile:
Name of the output image file (may be None, which is the
default). The image type (e.g. PNG) is derived from the file
extension. If multiple timesteps are to be animated, the
output files will have additional suffixes of the form
'_N_TIMESTEP', where N represents the index of the timestep
(in the array passed as the argument `timesteps`) and TIMESTEP
is the actual timestep itself.
field_name:
The field to plot. Default: 'm' (= the normalised magnetisation).
Note that this field must of course have been saved in the .pvd
file.
timesteps:
The timesteps for which to render the scene. The default is
None, which means to animate all timesteps (and save them as a
sequence of images if `outfile` is specified). Other valid
values are either a single number or a list of numbers.
camera_position: 3-vector
camera_focal_point: 3-vector
camera_view_up: 3-vector
These variables control the position and orientation of the
camera.
view_size: pair of int
Controls the size of the view. This can be used to adjust the
size and aspect ratio of the visible scene (useful for example
if a colorbar is present). Default: (400, 400).
magnification: int
Magnification factor which controls the size of the saved image.
Note that due to limitations in Paraview this must be an integer.
fit_view_to_scene: True | False
If True (the default), the view is automatically adjusted so
that the entire scene is visible. In this case the exact
location of the camera is ignored and only its relative
position w.r.t. the focal point is taken into account.
color_by_axis: integer or string (allowed values: 0, 1, 2, -1,
or 'x', 'y', 'z', 'magnitude')
The vector components in the direction of this axis are used
to color the plot. If '-1' is given, the vector magnitudes
are used instead of any vector components.
colormap:
The colormap to use. Supported values:
{}.
rescale_colormap_to_data_range: True | False
If False (default: True), the colormap corresponds to the data
range [-1.0, +1.0]. If set to True, the colormap is rescaled
so that it corresponds to the minimum/maximum data values *over
all specified timesteps*.
show_colorbar: True | False
If True (the default), a colorbar is added to the plot.
colorbar_label_format: string
Controls how colorbar labels are formatted (e.g., how many
digits are displayed, etc.). This can be any formatting string
for floating point numbers as understood by Python's 'print'
statement. Default: '%-#5.2g'.
add_glyphs: True | False
If True (the default), little glyphs are added at the mesh
vertices to indicate the direction of the field vectors.
glyph_type: string
Type of glyphs to use. The only currently supported glyph type
is 'cones'.
glyph_scale_factor: float
Controls the glyph size. The default value of 1.0 corresponds
to a value automatically determined by the plotting function
which makes the glyphs visible but keeps them small enough so
that glyphs at different vertices don't overlap. This argument
can be used to tweak that size (e.g. '0.5' means to use half
the automatically determined size, and '3.0' three times that
size, etc.).
glyph_mask_points: True | False
If True (the default), limit the maximum number of glyphs to
the value indicated by glyph_max_number_of_points.
glyph_max_number_of_points: int
Specifies the maximum number of glyphs that should appear in
the output dataset if glyph_mask_points is True.
glyph_random_mode: True | False
If True (the default), the glyph positions are chosen
randomly. Otherwise the point IDs to which glyphs are attached
are evenly spaced. This setting only has an effect if
glyph_mask_points is True.
show_orientation_axes: False | True
If True (default: False), a set of three small axes is added
to the scene to indicate the directions of the coordinate axes.
show_center_axes: False | True
If True (default: False), a set of three axes is plotted at
the center of rotation.
representation: string
Controls the way in which the visual representation of bodies
in the scene. Allowed values:
{}
palette: 'print' | 'screen'
The color scheme to be used. The main difference is that
'print' uses white as the background color whereas 'screen'
uses dark grey.
use_parallel_projection: True | False
If False (the default), perspective projection is used to
render the scene. Otherwise parallel projection is used.
trim_border: True | False
If True (the default), any superfluous space around the scene
will be trimmed from the saved image. This requires imagemagick
to be installed.
rescale: float | None
Factor by which the output image will be rescaled. For example,
using 'rescale=0.4' will rescale the image by 40%.
diffuse_color: 3-tuple of RGB values
The solid color of the body. If given, this overrides any
colormap-related values.
"""
from paraview import servermanager
import paraview.simple as pv
# Paraview crashes if there is no X server running, so we check
# whether this is the case.
if not os.environ.has_key('DISPLAY'):
logger.warning(
"Could not detect a running X server (this may happen, for "
"example, if you are on a ssh connection without X forwarding; "
"use 'ssh -X' in this case). Aborting because Paraview is "
"likely to crash.")
if not representation in _representations:
raise ValueError("Unsupported representation: '{}'. Allowed values: "
"{}".format(representation, _representations))
if abs(magnification - int(magnification)) > 1e-6:
logger.warning("Due to limitations in Paraview, the 'magnification' "
"argument must be an integer (got: {}). Using nearest "
"integer value.".format(magnification))
magnification = int(round(magnification))
if not os.path.exists(pvd_file):
raise IOError("File does not exist: '{}'.".format(pvd_file))
servermanager.Disconnect()
servermanager.Connect()
reader = servermanager.sources.PVDReader(FileName=pvd_file)
reader.UpdatePipeline()
view = servermanager.CreateRenderView()
repr = servermanager.CreateRepresentation(reader, view)
repr.Representation = representation
view.CameraPosition = camera_position
view.CameraFocalPoint = camera_focal_point
view.CameraViewUp = camera_view_up
if fit_view_to_scene:
# N.B.: this email describes a more sophisticated (= proper?)
# way of doing this, but it's probably overkill for now:
#
# http://www.paraview.org/pipermail/paraview/2012-March/024352.html
#
view.ResetCamera()
view.OrientationAxesVisibility = (1 if show_orientation_axes else 0)
view.CenterAxesVisibility = (1 if show_center_axes else 0)
if palette == 'print':
view.Background = [1.0, 1.0, 1.0]
view.OrientationAxesLabelColor = [0.0, 0.0, 0.0]
repr.AmbientColor = [0.0, 0.0, 0.0]
elif palette == 'screen':
view.Background = [0.32, 0.34, 0.43]
view.OrientationAxesLabelColor = [1.0, 1.0, 1.0]
repr.AmbientColor = [1.0, 1.0, 1.0]
else:
raise ValueError("Palette argument must be either 'print' "
"or 'screen'. Got: {}".format(palette))
view.CameraParallelProjection = 1 if use_parallel_projection else 0
# Convert color_by_axis to integer and store the name separately
try:
color_by_axis = _axes[color_by_axis.lower()]
except AttributeError:
if not color_by_axis in [0, 1, 2, -1]:
raise ValueError("color_by_axis must have one of the values "
"[0, 1, 2, -1] or ['x', 'y', 'z', 'magnitude']. "
"Got: {}".format(color_by_axis))
color_by_axis_name = _axes_names[color_by_axis]
if timesteps is None:
timesteps = reader.TimestepValues
elif not isinstance(timesteps, (list, tuple, np.ndarray)):
if not isinstance(timesteps, numbers.Number):
raise TypeError(
"Argument 'timesteps' must be either None or a number or a list of numbers. Got: '{}'".format(timesteps))
timesteps = [timesteps]
data_range = (-1.0, 1.0)
if rescale_colormap_to_data_range:
dmin, dmax = np.infty, -np.infty
for t in timesteps:
reader.UpdatePipeline(t)
dataInfo = reader.GetDataInformation()
pointDataInfo = dataInfo.GetPointDataInformation()
arrayInfo = pointDataInfo.GetArrayInformation(field_name)
cur_data_range = arrayInfo.GetComponentRange(color_by_axis)
dmin = min(cur_data_range[0], dmin)
dmax = max(cur_data_range[1], dmax)
data_range = (dmin, dmax)
logger.debug("Rescaling colormap to data range: {}".format(data_range))
# Set the correct colormap and rescale it if necessary.
try:
cmap = _color_maps[colormap]
if colormap == 'blue_to_red_rainbow':
print(textwrap.dedent("""
Use of the 'rainbow' color map is discouraged as it has a number of distinct
disadvantages. Use at your own risk! For details see, e.g., [1], [2].
[1] K. Moreland, "Diverging Color Maps for Scientific Visualization"
http://www.sandia.gov/~kmorel/documents/ColorMaps/ColorMapsExpanded.pdf
[2] http://www.paraview.org/ParaView3/index.php/Default_Color_Map
"""))
except KeyError:
raise ValueError("Unsupported colormap: '{}'. Allowed values: "
"{}".format(colormap, _color_maps.keys()))
lut = servermanager.rendering.PVLookupTable()
lut.ColorSpace = cmap.color_space
rgb_points = cmap.rgb_points
dmin, dmax = data_range
cmin = rgb_points[0]
cmax = rgb_points[-4]
if cmin == cmax:
# workaround for the case that the data range only
# contains a single value
cmax += 1e-8
for i in xrange(0, len(rgb_points), 4):
rgb_points[i] = (rgb_points[i] - cmin) / \
(cmax - cmin) * (dmax - dmin) + dmin
lut.RGBPoints = rgb_points
lut.NanColor = cmap.nan_color
if color_by_axis in [0, 1, 2]:
lut.VectorMode = "Component"
lut.VectorComponent = color_by_axis
elif color_by_axis == -1:
lut.VectorMode = "Magnitude"
lut.VectorComponent = color_by_axis
if diffuse_color is not None:
print "diffuse_color: {} ({})".format(diffuse_color, type(diffuse_color))
repr.DiffuseColor = diffuse_color
else:
repr.LookupTable = lut
if field_name is not None:
repr.ColorArrayName = ("POINT_DATA", field_name)
if add_glyphs:
logger.debug("Adding cone glyphs.")
glyph = pv.servermanager.filters.Glyph(Input=reader)
# Try to determine an appropriate scale_factor automatically
import vtk.util.numpy_support as VN
grid = servermanager.Fetch(reader)
# Determine approximate mesh spacing
def mesh_spacing_for_cell(cell):
cell_bounds = np.array(cell.GetBounds()).reshape((3, 2))
return float(min(filter(lambda x: x != 0.0, cell_bounds[:, 1] - cell_bounds[:, 0])))
mesh_spacing = np.average(
[mesh_spacing_for_cell(grid.GetCell(i)) for i in range(grid.GetNumberOfCells())])
# Determine maximum field magnitude
m = VN.vtk_to_numpy(grid.GetPointData().GetArray(field_name))
max_field_magnitude = float(max(map(np.linalg.norm, m)))
glyph_scale_factor_internal = mesh_spacing / max_field_magnitude
logger.debug(
"Using automatically determined glyph_scale_factor_internal = {:.2g} "
"(determined from approximate mesh spacing {:.2g} and maximum "
"field magnitude {:.2g}). This may need manual tweaking in case "
"glyphs appear very large or very small.".format(
glyph_scale_factor_internal, mesh_spacing, max_field_magnitude))
glyph.ScaleFactor = glyph_scale_factor * glyph_scale_factor_internal
glyph.ScaleMode = 'vector'
glyph.Vectors = ['POINTS', field_name]
try:
# only relevant for animation IIUC, but can't hurt setting it
glyph.KeepRandomPoints = 1
except AttributeError:
# Older version of Paraview which doesn't support this setting.
# Ignoring for now.
pass
#glyph.MaskPoints = glyph_mask_points
#glyph.MaximumNumberofPoints = glyph_max_number_of_points
if glyph_type != 'cones':
glyph_type = 'cones'
logger.warning("Unsupported glyph type: '{}'. "
"Falling back to 'cones'.".format(glyph_type))
if glyph_type == 'cones':
cone = servermanager.sources.Cone()
cone.Resolution = 20
cone.Radius = 0.2
else:
# This should not happen as we're catching it above.
raise NotImplementedError()
glyph.SetPropertyWithName('Source', cone)
glyph_repr = servermanager.CreateRepresentation(glyph, view)
glyph_repr.LookupTable = lut
glyph_repr.ColorArrayName = ("POINT_DATA", 'GlyphVector')
if show_colorbar:
# XXX TODO: Remove the import of paraview.simple once I know why
from paraview.simple import CreateScalarBar
scalarbar = CreateScalarBar(
Title=field_name, ComponentTitle=color_by_axis_name.capitalize(),
Enabled=1, LabelFontSize=12, TitleFontSize=12)
scalarbar.LabelFormat = colorbar_label_format,
if palette == 'print':
scalarbar.LabelColor = [0.0, 0.0, 0.0] # black labels for print
else:
scalarbar.LabelColor = [1.0, 1.0, 1.0] # white labels for screen
view.Representations.append(scalarbar)
scalarbar.LookupTable = lut
reader.UpdatePipelineInformation()
if outfile is None:
_, outfile = tempfile.mkstemp(suffix='.png')
view.ViewSize = view_size
def write_image(outfilename):
_, suffix = os.path.splitext(outfilename)
if suffix == '.png':
view.WriteImage(outfilename, "vtkPNGWriter", magnification)
elif suffix in ['.jpg', '.jpeg']:
view.WriteImage(outfilename, "vtkJPEGWriter", magnification)
else:
raise ValueError("Output image must have extension '.png' or "
"'.jpg'. Got: {}".format(suffix))
if trim_border:
if palette == 'print':
bordercolor = '"rgb(255,255,255)"'
else:
# Strangely, we get a slightly different background
# color for PNG than for JPG.
bordercolor = '"rgb(82,87,110)"' if (
suffix == '.png') else '"rgb(82,87,109)"'
cmd = 'mogrify -bordercolor {} -border 1x1 -trim {}'.format(
bordercolor, outfilename)
try:
sp.check_output(cmd, stderr=sp.STDOUT, shell=True)
logger.debug("Trimming border from rendered scene.")
except OSError:
logger.warning(
"Using the 'trim' argument requires ImageMagick to be installed.")
except sp.CalledProcessError as ex:
logger.warning("Could not trim border from image. "
"The error message was: {}".format(ex.output))
if rescale:
rescale_factor = int(rescale * 100.0)
cmd = 'mogrify -resize {:d}% {}'.format(
rescale_factor, outfilename)
try:
sp.check_output(cmd, stderr=sp.STDOUT, shell=True)
logger.debug(
"Resizing output image by {:d}%".format(rescale_factor))
except OSError:
logger.warning(
"Using the 'rescale' argument requires ImageMagick to be installed.")
except sp.CalledProcessError as ex:
logger.warning(
"Could not rescale image. The error message was: {}".format(ex.output))
if len(timesteps) == 1:
# If a single timestep is rendered, we return the resulting image.
view.ViewTime = timesteps[0]
write_image(outfile)
res = IPython.core.display.Image(filename=outfile)
else:
# Otherwise we export a bunch of images with sequentially
# numbered suffixes.
#
# TODO: What should we return? Just the image for the first
# timestep as we currently do? Or can we somehow create
# a video and return that?
outbasename, outsuffix = os.path.splitext(outfile)
def generate_outfilename(i, t):
return outbasename + '_{:04d}_'.format(i) + str(t) + outsuffix
for (i, t) in enumerate(timesteps):
view.ViewTime = t
cur_outfilename = generate_outfilename(i, t)
logger.debug(
"Saving timestep {} to file '{}'.".format(t, cur_outfilename))
write_image(cur_outfilename)
res = IPython.core.display.Image(
filename=generate_outfilename(0, timesteps[0]))
servermanager.Disconnect()
return None
#!/usr/bin/python
# Import libraries
from paraview.simple import *
from paraview import servermanager
# Connect to server
if not servermanager.ActiveConnection:
connection = servermanager.Connect('x01y01.zbox.physik.uzh.ch', 11111)
# Load AstroViz plugin
servermanager.LoadPlugin('libAstroVizPlugin.dylib')
# Use the TipsyReader
tipsyfile = servermanager.sources.TipsyReader(FileName='b1.00300.d0-1000.std')
# Finding the center of mass
com = servermanager.filters.CenterOfMass(Input=tipsyfile)
com.SelectInputArray = ['POINTS', 'mass']
com.UpdatePipeline()
# Render the result
view = servermanager.CreateRenderView()
rep = servermanager.CreateRepresentation(com, view)
view.StillRender()
#########
# Make sure: PYTHONPATH includes the VTK/Wrapping/Python and the bin/ directories from your ParaView build:
#export PYTHONPATH=/Users/corbett/Documents/Projects/Work/Viz/ParaView/ParaView-3.6.1-Stable/ParaView-3.6.1-Stable_build/Utilities/VTKPythonWrapping/:/Users/corbett/Documents/Projects/Work/Viz/ParaView/ParaView-3.6.1-Stable/ParaView-3.6.1-Stable_build/bin/
###############
# Connect to Server
#
# Here we connect to the built in server.
# Can also connect to an external server e.g. with
# connection = servermanager.Connect('x01y01.zbox.physik.uzh.ch', 11111)
###############
if not servermanager.ActiveConnection:
from paraview import servermanager
from time import time
import timeit
import pickle
filename = "test.bench"
file = open(filename, 'w')
if not servermanager.ActiveConnection:
connection = servermanager.Connect()
coneSource = servermanager.sources.ConeSource()
shrinkFilter = servermanager.filters.ShrinkFilter(Input=coneSource)
view = servermanager.CreateRenderView()
rep = servermanager.CreateRepresentation(shrinkFilter, view)
view.ResetCamera()
total_time = 0.0
for i in range(0, 500):
t1 = time()
view.StillRender()
t2 = time()
elapsed = t2 - t1
total_time += elapsed
pickle.dump("elapsed time: " + str(elapsed), file)
print "total time: " + str(total_time)
print "avg time: " + str(total_time / 500)
