/usr/share/pyshared/ase/visualize/fieldplotter.py is in python-ase 3.6.0.2515-1.1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 | """plotting fields defined on atoms during a simulation."""
from ase.visualize.primiplotter import PostScriptFile, PnmFile, GifFile, JpegFile, X11Window
from ase.visualize.primiplotter import PrimiPlotter as _PrimiPlotter
import numpy
import time
class FieldPlotter(_PrimiPlotter):
def __init__(self, atoms, datasource=None, verbose=0, timing=0,
interval=1, initframe=0):
_PrimiPlotter.__init__(self, atoms, verbose=verbose, timing=timing,
interval=interval, initframe=initframe)
self.datasource = datasource
self.dims = (100,100)
self.set_plot_plane("xy")
self.set_data_range("plot")
self.set_background(0.0)
self.set_red_yellow_colors()
def set_plot_plane(self, plane):
"""Set the plotting plane to xy, xz or yz (default: xy)"""
if plane in ("xy", "xz", "yz"):
self.plane = plane
else:
raise ValueError, "The argument to plotPlane must be 'xy', 'xz' or 'yz'."
def set_data_range(self, range1, range2=None):
"""Set the range of the data used when coloring.
This function sets the range of data values mapped unto colors
in the final plot.
Three possibilities:
'data': Autoscale using the data on visible atoms.
The range goes from the lowest to the highest
value present on the atoms. If only a few atoms
have extreme values, the entire color range may not
be used on the plot, as many values may be averaged
on each point in the plot.
'plot': Autoscale using the data on the plot. Unlike 'data'
this guarantees that the entire color range is used.
min, max: Use the range [min, max]
"""
if (range1 == "data" or range1 == "plot") and range2 == None:
self.autorange = range1
elif range2 != None:
self.autorange = None
self.range = (range1, range2)
else:
raise ValueError, "Illegal argument(s) to set_data_range"
def set_background(self, value):
"""Set the data value of the background. See also set_background_color
Set the value of the background (parts of the plot without atoms) to
a specific value, or to 'min' or 'max' representing the minimal or
maximal data values on the atoms.
Calling set_background cancels previous calls to set_background_color.
"""
self.background = value
self.backgroundcolor = None
def set_background_color(self, color):
"""Set the background color. See also set_background.
Set the background color. Use a single value in the range [0, 1[
for gray values, or a tuple of three such values as an RGB color.
Calling set_background_color cancels previous calls to set_background.
"""
self.background = None
self.backgroundcolor = color
def set_red_yellow_colors(self, reverse=False):
"""Set colors to Black-Red-Yellow-White (a.k.a. STM colors)"""
self.set_colors([(0.0, 0, 0, 0),
(0.33, 1, 0, 0),
(0.66, 1, 1, 0),
(1.0, 1, 1, 1)],
reverse)
def set_black_white_colors(self, reverse=False):
"""Set the color to Black-White (greyscale)"""
self.set_colors([(0.0, 0), (1.0, 1)], reverse)
def set_colors(self, colors, reverse=False):
colors = numpy.array(colors, numpy.float)
if len(colors.shape) != 2:
raise ValueError, "Colors must be a 2D array."
if reverse:
colors[:,0] = 1 - colors[:,0]
colors = numpy.array(colors[::-1,:])
#print colors
if colors[0,0] != 0.0 or colors[-1,0] != 1.0:
raise ValueError, "First row must define the value 0 and last row must define the value 1"
if colors.shape[1] == 2:
self.colormode = 1
elif colors.shape[1] == 4:
self.colormode = 3
else:
raise ValueError, "Color specification must be Nx2 (grey) or Nx4 (rgb) matrix."
self.colorfunction = InterpolatingFunction(colors[:,0], colors[:,1:])
def plot(self, data=None):
"""Create a plot now. Does not respect the interval timer.
This method makes a plot unconditionally. It does not look at
the interval variable, nor is this plot taken into account in
the counting done by the update() method if an interval
variable was specified.
If data is specified, it must be an array of numbers with the
same length as the atoms. That data will then be plotted. If
no data is given, the data source specified when creating the
plotter is used.
"""
if self.timing:
self._starttimer()
self.log("FieldPlotter: Starting plot at "
+ time.strftime("%a, %d %b %Y %H:%M:%S"))
if data is None:
data = self.datasource()
if len(data) != len(self.atoms):
raise ValueError, ("Data has wrong length: %d instead of %d."
% (len(data), len(self.atoms)))
invisible = self._getinvisible()
coords = self._rotate(self._getpositions())
radii = self._getradii()
if self.autoscale:
self._autoscale(coords,radii)
scale = self.scale * self.relativescale
coords = scale * coords
center = self._getcenter(coords)
offset = numpy.array(self.dims + (0.0,))/2.0 - center
coords = coords + offset
radii = radii * scale
self.log("Scale is %f and size is (%d, %d)"
% (scale, self.dims[0], self.dims[1]))
self.log("Physical size of plot is %f Angstrom times %f Angstrom"
% (self.dims[0] / scale, self.dims[1] / scale))
# Remove invisible atoms
selector = numpy.logical_not(invisible)
coords = numpy.compress(selector, coords, 0)
radii = numpy.compress(selector, radii)
data = numpy.compress(selector, data)
self.log("plotting data in the range [%f,%f]" %
(data.min(), data.max()))
# Now create the output array
sumarray = numpy.zeros(self.dims, numpy.float)
weight = numpy.zeros(self.dims)
# Loop over all atoms, and plot them
nmiss = 0
if self.plane == "xy":
xy = coords[:,:2]
elif self.plane == "xz":
xy = coords[:,::2]
elif self.plane == "yz":
xy = coords[:,1:]
else:
raise RuntimeError, "self.plane is bogus: "+str(self.plane)
assert xy.shape[1] == 2
self.log("plotting %d atoms on %d * %d (= %d) grid" %
(len(xy), sumarray.shape[0], sumarray.shape[1],
len(sumarray.flat)))
xy = xy.astype(numpy.int)
for i in xrange(len(xy)):
(x, y) = xy[i]
d = data[i]
if (x >= 0 and x < self.dims[0] and y >= 0 and y < self.dims[1]):
sumarray[x,y] += d
weight[x,y] += 1
else:
nmiss += 1
print "... %d atoms fell outside plot." % (nmiss,)
datamap = self._makedatamap(sumarray, weight, data.min(), data.max())
self.log("Range of data map: [%f, %f]" %
(datamap.min(), datamap.max()))
plot = self._makeplotmap(datamap, weight)
#self.log("Range of plot: [%f, %f]" %
# (min(plot.flat), max(plot.flat)))
examinplot = plot[:]
examinplot.shape = (plot.shape[0] * plot.shape[1],) + plot.shape[2:]
self.log("Range of plot: %s -> %s" %
(str(examinplot.min(0)), str(examinplot.max(0))))
del examinplot
for device in self.outputdevice:
device.inform_about_scale(scale)
device.plotArray(self.n, numpy.swapaxes(plot,0,1))
self.n = self.n + 1
self.log("FieldPlotter: Finished plotting at "
+ time.strftime("%a, %d %b %Y %H:%M:%S"))
self.log("\n\n")
def _makedatamap(self, sumarray, weight, minimum, maximum):
background = numpy.equal(weight, 0)
print "Number of background points:", sum(background.flat)
datamap = sumarray / numpy.where(background, 1, weight)
if self.background is not None:
if self.background == "min":
bg = minimum
elif self.background == "max":
bg = maximum
else:
bg = self.background
datamap = numpy.where(background, bg, datamap)
if self.autorange == "data":
datamap = (datamap - minimum) / (maximum - minimum)
self.log("Autorange using data. Data range is [%f, %f]"
% (minimum, maximum))
elif self.autorange == "plot":
ma = numpy.where(background, minimum, datamap).max()
mi = numpy.where(background, maximum, datamap).min()
datamap = (datamap - mi) / (ma - mi)
self.log("Autorange using plot. Data range is [%f, %f]"
% (mi, ma))
else:
assert self.autorange == None
datamap = (datamap - self.range[0]) / (self.range[1]
- self.range[0])
datamap = numpy.clip(datamap, 0.0, 1.0)
self.log("Data range specified by user: [%f, %f]" % self.range)
datamap = numpy.where(background, bg, datamap)
assert datamap.min() >= 0 and datamap.max() <= 1.0
return datamap
def _makeplotmap(self, datamap, weight):
plot = numpy.zeros(self.dims + (self.colormode,), numpy.float)
for i in range(self.dims[0]):
for j in range(self.dims[1]):
if self.backgroundcolor is not None and weight[i,j] == 0:
plot[i,j,:] = self.backgroundcolor
else:
x = datamap[i,j]
plot[i,j,:] = self.colorfunction(x)
return plot
class InterpolatingFunction:
def __init__(self, xpoints, ypoints):
if len(xpoints) != len(ypoints):
raise ValueError, "Length of x and y arrays should be the same."
idx = xpoints.argsort()
self.xpoints = xpoints[idx]
self.ypoints = ypoints[idx]
def __call__(self, x):
n = self.xpoints.searchsorted(x)
if n == 0:
return self.ypoints[0]
if n == len(self.xpoints):
return self.xpoints[-1]
x0 = self.xpoints[n-1]
x1 = self.xpoints[n]
y0 = self.ypoints[n-1]
y1 = self.ypoints[n]
return y0 + (y1 - y0) / (x1 - x0) * (x - x0)
|