def f(mode, fname):
wf = functools.partial(lambda z, x: wavefs[mode - 1](x, z), 0)
(fig, ax) = plotting.setup_figure_standard(title=u'Argand diagram for n=%i mode' % mode,
xlabel=ur'$\mathrm{Re}\left((\psi\left(x\right)\right)$',
ylabel=ur'$\mathrm{Im}\left((\psi\left(x\right)\right)$')
plotting.plot_argand(ax, wf, waveguide.slab_gap)
plotting.save_figure(fig, fname)
def wrangle(data, target=sentiments):
'''Prepare data for classification'''
raw_len = len(data.index)
# Drop duplicate data
data.drop_duplicates('Sentence', inplace=True)
# Reverse one-hot encoding to obtain single target vector
data['Sentiment'] = data.loc[:, target].idxmax(axis=1)
data.drop(target, axis=1, inplace=True)
# Define length of sentences by splitting per space
data['NumWords'] = data['Sentence'].str.split().str.len()
# Remove special characters without meaning
#
# Pie chart to visualize fraction of useful data
print('{} instances of {} ({:.3f}%) remaining.'.format(
len(data.index), raw_len,
len(data.index) / raw_len * 100.0))
plt.figure()
labels = ['Good data', 'Bad data']
sizes = [len(data.index), raw_len - len(data.index)]
plt.pie(sizes, labels=labels, autopct='%1.1f%%')
plotting.save_figure(plt.gcf(), 'useful_data')
def evaluate_parameter(results, poi, tag,
xlabel=None, ylabel=None,
categorical=False, log=False, xticklabels=None):
# Replace NaN (or None) with 0 to allow for Python-style logical comparison
results = results.fillna(0)
# Create parameter index (without scores) to filter results with
optimum = results.loc[results[selected_score].idxmax()]
idx = optimum.index.drop(scores)
# Get sample of results where all parameters are identical except for the
# parameter of interest
sample = results[conjunction([results[n] == optimum[n] for n in idx.drop(poi)])]
if categorical:
c = sns.catplot(x=poi, y=selected_score, data=sample, kind='bar')
plt.ylabel(ylabel if ylabel else selected_score)
plt.xlabel(xlabel if xlabel else poi)
if xticklabels:
vals, labels = plt.xticks()
plt.xticks(vals,
[xticklabels(l.get_text()) for l in labels],
rotation=30)
plt.xlabel('')
plotting.save_figure(plt.gcf(), tag + '_' + poi)
else:
sample = sample.sort_values(by=poi)
plt.figure()
plt.errorbar(x=poi, y=selected_score, data=sample)
plt.fill_between(sample[poi],
sample[selected_score] - sample[selected_score_std],
sample[selected_score] + sample[selected_score_std],
alpha=0.2)
plt.ylabel(ylabel if ylabel else selected_score)
plt.xlabel(xlabel if xlabel else poi)
if log: plt.xscale('log')
plotting.save_figure(plt.gcf(), tag + '_' + poi)
def f(mode, fname):
wavefs_propagated = []
labels = []
for d in dists:
wavefs_propagated.append(functools.partial(lambda z,x: wavefs[mode-1](x,z), d))
labels.append('z=%.3fm' % d)
(fig, ax) = plotting.setup_figure_standard(title=u'Intensity for n=%i mode' % mode,
xlabel=u'Distance accross waveguide (m)',
ylabel=u'Wavefunction intensity (arbitrary units)')
#fig.hold(True)
colours = ['blue', 'orange', 'green', 'purple', 'grey', 'lime', 'cyan', 'yellow', 'black', 'navy', 'teal']
colours.reverse()
ax.hold(True)
d = []
for (wf, lbl) in zip(wavefs_propagated, labels):
d.append(plotting.plot_intensity(ax, wf, waveguide.slab_gap, colour=colours.pop(), label=lbl))
ax.hold(False)
if len(dists) != 1:
ax.legend(loc='upper left', prop={'size' : 10})
plotting.save_figure(fig, fname)
i = 1
for data in d:
sio.savemat(fname + str(i) + '.mat',data)
i += 1
def f(mode, fname):
wf = functools.partial(lambda z,x: wavefs[mode-1](x,z), 0)
(fig, ax) = plotting.setup_figure_standard(title=u'Poynting vector for n=%i mode' % mode,
xlabel=u'Distance accross waveguide (m)',
ylabel=u'Energy flow (right is +)')
d = plotting.plot_poynting_vector(ax, wf, waveguide.slab_gap)
plotting.save_figure(fig, fname)
sio.savemat(fname + '.mat', d)
def f(mode, fname):
wavef = lambda x: wavefs[mode-1](x, 0)
(fig, reax, imax) = plotting.setup_figure_topbottom(title=u'Wavefunction for n=%i mode' % mode,
#fig, reax) = plotting.setup_figure_standard(title=u'Wavefunction for n=%i mode' % mode,
xlabel=u'Distance accross waveguide (m)',
ylabel=u'Wavefunction (arbitrary units)')
d = plotting.plot_wavefunction(reax, imax, wavef, waveguide.slab_gap)
plotting.save_figure(fig, fname)
sio.savemat(fname + '.mat', d)
def tf_idf(data):
param_grid = [
{'vec__stop_words': [None, 'english'],
'vec__ngram_range': [(1, 1), (1, 2)],
'vec__min_df': [3, 4, 5],
'vec__max_df': [0.5, 0.75, 0.9],
# 'vec__sublinear_tf': [True, False],
'model': [LinearSVC],
'clf__max_iter': [10000],
'clf__dual': [False, True],
'clf__class_weight': ['balanced'],
'clf__random_state': [0]},
{'vec__stop_words': [None, 'english'],
'vec__ngram_range': [(1, 1), (1, 2)],
'vec__min_df': [1, 2, 3, 4],
'vec__max_df': [0.5, 0.75, 0.9, 0.95],
'vec__sublinear_tf': [True, False],
'model': [MultinomialNB, ComplementNB, BernoulliNB],
'clf__alpha': [0.001, 0.01, 0.1, 1.0]}
]
results = mlearning.cross_validate(
mlearning.tf_idf, ParameterGrid(param_grid),
data['Sentence'], data['Sentiment']
).sort_values(by='balanced_accuracy_score', ascending=False)
print('--- Tf-Idf results ---')
print(results
.drop(['accuracy_score', 'accuracy_score_std', 'confusion_matrix', 'parameters'], axis=1)
.rename({'balanced_accuracy_score': 'bas', 'balanced_accuracy_score_std': 'bas_std'}, axis=1)
.head(20))
mlearning.characterize_optimum(results, data, mlearning.tf_idf)
# Confusion matrix
optimum = results.loc[results['balanced_accuracy_score'].idxmax()]
plotting.confusion_heatmap(optimum['confusion_matrix'], sentiments)
plotting.save_figure(plt.gcf(), 'confusion_heatmap')
plotting.confusion_heatmap(optimum['confusion_matrix'], sentiments, True)
plotting.save_figure(plt.gcf(), 'confusion_heatmap_norm')
# Hyperparameter dependency
evaluate_parameter(results, 'clf__alpha', 'tfidf',
ylabel='Balanced accuracy score', xlabel='Alpha', log=True)
evaluate_parameter(results, 'vec__max_df', 'tfidf',
ylabel='Balanced accuracy score', xlabel='Max doc frequency')
evaluate_parameter(results, 'vec__min_df', 'tfidf',
ylabel='Balanced accuracy score', xlabel='Min doc frequency')
evaluate_parameter(results, 'vec__ngram_range', 'tfidf',
ylabel='Balanced accuracy score', xlabel='n-gram range', categorical=True)
evaluate_parameter(
results, 'model', 'tfidf', ylabel='Balanced accuracy score', categorical=True,
xticklabels=lambda x: str(x).split("'")[1].split('.')[-1])
def characterize(data):
'''Characterize the data visually'''
# Bias in terms of sentiment
plt.figure()
plotting.save_figure(sns.countplot(x='Sentiment', data=data),
'sentiment_balance')
# Overview of length of sentences
plt.figure()
plotting.save_figure(sns.distplot(data['NumWords']), 'num_words')
# Overview of length of sentences per sentiment
plt.figure()
colors = plotting.palette()
fig, axs = plt.subplots(1, 3, figsize=(15, 6), sharey=True)
for sentiment, ax in zip(sentiments, axs):
color = next(colors)
p = sns.distplot(data.loc[data['Sentiment'] == sentiment, 'NumWords'],
bins=8, ax=ax, color=color,
hist=True, kde=False, norm_hist=True,
fit=stats.lognorm, fit_kws={'color': color})
p.set_title(sentiment)
# plt.legend()
plotting.save_figure(plt.gcf(), 'num_words_sentiments')
# Overview of length of sentences per sentiment
plt.figure()
colors = plotting.palette()
for sentiment, ax in zip(sentiments, axs):
sns.distplot(data.loc[data['Sentiment'] == sentiment, 'NumWords'],
bins=8, label=sentiment,
hist=False, kde=False, norm_hist=True,
fit=stats.lognorm, fit_kws={'color': next(colors)})
plt.legend()
plotting.save_figure(plt.gcf(), 'num_words_fits')
# Visualize splitting of data
plt.figure()
fold = StratifiedKFold(n_splits=5)
plotting.cross_validation_indices(
fold, data['Sentence'],
data['Sentiment'].map(
{'Negative': 0, 'Neutral': 1, 'Positive': 2}).sort_values(),
ax, 5)
plotting.save_figure(plt.gcf(), 'stratified_cross_fold')
def sp_plot_wavefunction(waveguide, wavelength, angle, out_file, verbose=False):
"""
This method produces a wavefunction hitting the waveguide at angle, and splits it into the guided modes of the
waveguide. The resultant wavefunction is plotted
@param waveguide: The waveguide being illuminated
@type waveguide: PlanarWaveguide
@param wavelength: The wavelength of light illuminating the waveguide
@type wavelength: float
@param angle: The angle of the plane waves striking the waveguide, in radian
@type angle: float
@param out_file: The filename to write output to
@type out_file: str
@param verbose: Whether or not to give verbose output
@type verbose: bool
@return: None
"""
solver = modesolvers.ExpLossySolver(waveguide, wavelength)
kxs = solver.solve_transcendental(verbose=verbose)
wavefs = solver.get_wavefunctions(kxs, verbose=verbose)
k = 2*np.pi/wavelength
inkx = k*np.sin(angle)
inwave = lambda x: np.exp(1j*inkx*x)
splitter = splitters.ModeSplitter(inwave, wavefs)
cm = splitter.get_coupling_constants()
wffull = splitter.get_wavefunction(cm)
wf = functools.partial(lambda z,x: wffull(x,z), 0.005)
if verbose:
print 'Coupling coefficients: '
for (i,c) in enumerate(cm):
print '\t %i: Magnitude: %f, Phase: %f, Square: %f' % (i+1, np.abs(c), np.angle(c), np.abs(c)**2)
(fig, reax, imax) = plotting.setup_figure_topbottom(title=ur'Wavefunction for incidence angle $\theta=%f$ rad' % angle,
xlabel=u'Distance accross waveguide (m)',
ylabel=u'Wavefunction (arbitrary units)')
#plotting.plot_intensity(reax, wf, waveguide.slab_gap)
#phasef = lambda x: np.angle(wf(x))
plotting.plot_wavefunction(reax, imax, wf, waveguide.slab_gap)
plotting.save_figure(fig, unicode(out_file))
def f(mode, fname):
kx = kxs[mode-1]
inwave = lambda x: np.exp(1j*kx.real*x)
splitter = splitters.ModeSplitter(inwave, wavefs)
cm = splitter.get_coupling_constants()
wffull = splitter.get_wavefunction(cm)
wf = lambda x: wffull(x, 0.005)
#(fig, reax, imax) = plotting.setup_figure_topbottom(
# title=ur'Wavefunction for incidence angle on n=%i mode' % mode,
# xlabel=u'Distance accross waveguide (m)',
# ylabel=u'Wavefunction (arbitrary units)')
(fig, ax) = plotting.setup_figure_standard(
title=ur'Wavefunction for incidence angle on n=%i mode' % mode,
xlabel=u'Distance accross waveguide (m)',
ylabel=u'Wavefunction (arbitrary units)')
#plotting.plot_wavefunction(reax, imax, wf, waveguide.slab_gap)
plotting.plot_intensity(ax, wf, waveguide.slab_gap)
plotting.save_figure(fig, fname)
coefftable[mode-1, :] = np.abs(cm)**2
def sp_plot_coupled_map(waveguide, wavelength, angle, zlimits, out_file, verbose=False):
"""
This method produces a plot of the intensity map in the waveguide when it is illuminated by a plane wave
of arbitrary angle
@param waveguide: The waveguide being illuminated
@type waveguide: PlanarWaveguide
@param wavelength: The wavelength of light illuminating the waveguide
@type wavelength: float
@param angle: The angle of the plane waves striking the waveguide, in radian
@type angle: float
@param zlimits: The z-axis limits within which to plot the wavefunction
@type zlimits: tuple
@param out_file: The filename to write output to
@type out_file: str
@param verbose: Whether or not to give verbose output
@type verbose: bool
@return: None
"""
#do the decomposition
solver = modesolvers.ExpLossySolver(waveguide, wavelength)
kxs = solver.solve_transcendental(verbose=verbose)
wavefs = solver.get_wavefunctions(kxs, verbose=verbose)
k = 2 * np.pi / wavelength
inkx = k * np.sin(angle)
inwave = lambda x: np.exp(1j * inkx * x)
splitter = splitters.ModeSplitter(inwave, wavefs)
cm = splitter.get_coupling_constants()
wffull = splitter.get_wavefunction(cm)
(fig, ax) = plotting.setup_figure_standard(
title=ur'Intensity for incidence angle %f rads' % angle,
xlabel=u'Longitudinal distance accross waveguide (m)',
ylabel=u'Transverse distance accross waveguide (m)')
(X,Y,Z) = plotting.plot_intensity_map(ax, wffull, waveguide.slab_gap, zlimits)
plotting.save_figure(fig, unicode(out_file))
sio.savemat(out_file + '.mat', {'inten' : Z, 'X' : X, 'Y' : Y})
def sp_plot_total_coupling(waveguide, wavelength, out_file, verbose=False):
solver = modesolvers.ExpLossySolver(waveguide, wavelength)
kxs = solver.solve_transcendental(verbose=verbose)
wavefs = solver.get_wavefunctions(kxs, verbose=verbose)
tirang = waveguide.critical_angle(wavelength)
angs = np.linspace(0, tirang+0.1*tirang, num=100)
intensity = np.zeros(100)
k = 2*np.pi/wavelength
pb = progressbar.ProgressBar(widgets=['Calculating Integrals: ', progressbar.Percentage(), progressbar.Bar()],
maxval=len(angs))
if verbose:
pb.start()
for (n,a) in enumerate(angs):
kx = np.sin(a)*k
inwave = lambda x: np.exp(1j*kx*x)
splitter = splitters.ModeSplitter(inwave, wavefs)
cm = splitter.get_coupling_constants()
wffull = splitter.get_wavefunction(cm)
wf = lambda x: wffull(x, 0)
integrand = lambda x: np.abs(wf(x))**2
area = integrate.quad(integrand, -1e-6, 1e-6, limit=3000)[0]
intensity[n] = area
if verbose:
pb.update(n)
(fig, ax) = plotting.setup_figure_standard(title=u'Coupling efficiency of waveguide',
xlabel=u'Angle of radiation incidence (rads)',
ylabel=u'Intensity Coupled')
ax.plot(angs, intensity)
plotting.save_figure(fig, out_file)
sio.savemat(out_file + '.mat', {'angs' : angs, 'Intensity' : intensity})
linewidth=1.5,
zorder=5)
i += 1
nodes = [1, 8, 27, 64, 125, 216]
strnodes = ['$' + str(x) + '^3$' for x in [1, 2, 3, 4, 5, 7, 9]]
ax.set_xscale("log", nonposx='clip')
#ax.set_yscale("log", nonposy='clip')
#ax.set_yticks(nodes)
#ax.set_yticklabels(np.array(nodes, dtype=int))
ax.set_xticks(nodes)
ax.set_xticklabels(strnodes)
ax.plot(nodes, np.ones(len(nodes)), "k--", label="ideal")
ax.set_xlabel("Nodes")
ax.set_ylabel("Efficiency $E = \dfrac{t_1 N}{t_N}$")
plt.legend() #frameon=False)
myplt.set_font_sizes(ax)
myplt.make_grid(ax, True)
if blood:
myplt.save_figure(fig, 3, 'strong_scale_blood.pdf')
else:
myplt.save_figure(fig, 3, 'strong_scale_poiseuille.pdf')
plt.show()
def ic_exit_surface(waveguide, sourcecls, angle, nwaves, out_file, verbose=False):
"""
This method implements incoherent --exitsurface behaviour
It plots the intensity profile at the exit surface of the waveguide when it is illuminated with incoherent
radiation from source at an angle of angle
@param waveguide: The waveguide being illuminated
@type waveguide: PlanarWaveguide
@param source: The characteristics of the source that is generating the radiation hitting this waveguide
@type source: type
@param angle: The angle the waveguide is tilted at with respect to the source
@type angle: float
@param nwaves: The number of waves to contribute to the plot of the exit surface
@type nwaves: int
@param out_file: The file to write the resultant plot to
@type out_file: str
@param verbose: Whether or not to give verbose output
@type verbose: bool
@return: None
"""
source = sourcecls(angle)
intensity = np.zeros((250,500))
(X, Y) = np.meshgrid(np.linspace(0, waveguide.length, 500),
np.linspace(-waveguide.slab_gap, waveguide.slab_gap, 250))
#let's see if our source is monochromatic
prevval = source.get_wave()[1]
ismonochrome = True
for i in range(10):
newval = source.get_wave()[1]
ismonochrome = ismonochrome and prevval == newval
prevval = newval
if ismonochrome:
#get this done once
solver = modesolvers.ExpLossySolver(waveguide, prevval)
kx = solver.solve_transcendental(verbose=False) #this'll get real noisy otherwise
modewaves = solver.get_wavefunctions(kx, verbose=False)
pb = progressbar.ProgressBar(widgets=['Performing Incoherent Sum: ', progressbar.Percentage(), progressbar.Bar()],
maxval=nwaves)
if verbose:
pb.start()
#We need to iterate over many waves, summing their intensity
for i in xrange(nwaves):
#get a wave
(wave, wl) = source.get_wave()
inwave = lambda x: wave(x, 0) #only intersted in incidence surface
#solve modes for this wl
if not ismonochrome:
solver = modesolvers.ExpLossySolver(waveguide, wl)
kx = solver.solve_transcendental(verbose=False) #this'll get real noisy otherwise
modewaves = solver.get_wavefunctions(kx, verbose=False)
#split wave into modes
splitter = splitters.ModeSplitter(inwave, modewaves)
cm = splitter.get_coupling_constants()
wffull = splitter.get_wavefunction(cm)
#get wavefunction at exit surface
exitf = lambda x,y: np.abs(wffull(y, x))**2
#incoherently add exitf to the picture so far
exitvf = np.vectorize(exitf)
intensity = intensity + exitvf(X,Y)
if verbose:
pb.update(i+1)
#now plot it
(fig, ax) = plotting.setup_figure_standard(
title=ur'Exit surface intensity',
xlabel=u'Longitudinal distance across waveguide (m)',
ylabel=u'Transverse distance across waveguide (m)')
#ax.plot(x, intensity, color='blue', linestyle='-', marker=None, antialiased=True)
#ax.set_xlim((-waveguide.slab_gap, waveguide.slab_gap))
#ax.set_ylim((np.min(intensity) - 0.1 * np.abs(np.min(intensity)),
# np.max(intensity) + 0.1 * np.abs(np.max(intensity))))
#plotting.shade_waveguide(ax, waveguide.slab_gap)
#Consider only the rightmost 4/5ths when setting the intensity
#lowest = np.min(intensity[:, 360:])
#highest = np.max(intensity[:, 360:])
pcm = ax.pcolor(X, Y, intensity, cmap=matplotlib.cm.jet)
ax.get_figure().colorbar(pcm, ax=ax, use_gridspec=True)
ax.set_xlim((0, waveguide.length))
ax.set_ylim((-waveguide.slab_gap, waveguide.slab_gap))
#bang down a slightly different guideline for the waveguide limits
topr = Rectangle((0, waveguide.slab_gap / 2), waveguide.length, waveguide.slab_gap / 2, hatch='\\', fill=False)
bottomr = Rectangle((0, -waveguide.slab_gap), waveguide.length, waveguide.slab_gap / 2, hatch='/', fill=False)
ax.add_patch(topr)
ax.add_patch(bottomr)
plotting.save_figure(fig, unicode(out_file))
#save the data too
sio.savemat(out_file+'.mat', {'intensity':intensity, 'X':X, 'Y':Y})
plt.plot(nodes,
dump,
'--',
label='Data dump',
linewidth=1.5,
zorder=3,
color='C1')
plt.plot(nodes, total, '-', label='Total', linewidth=2.5, zorder=2, color='C2')
nodes = [x**3 for x in [1, 2, 3, 4, 5, 6]]
strnodes = ['$' + str(x) + '^3$' for x in [1, 2, 3, 4, 5, 6]]
ax.set_xscale("log", nonposx='clip')
#ax.set_yscale()
#ax.set_yscale("log", nonposy='clip')
#ax.set_yticks(nodes)
#ax.set_yticklabels(np.array(nodes, dtype=int))
ax.set_xticks(nodes)
ax.set_xticklabels(strnodes)
ax.set_xlabel("Nodes")
ax.set_ylabel("Time per step, \si{\milli\second}")
plt.legend() #frameon=False)
myplt.set_font_sizes(ax)
myplt.make_grid(ax)
myplt.save_figure(fig, 3, 'datadump.pdf')
#plt.show()
zorder=5)
i += 1
nodes = [x**3 for x in [1, 2, 3, 4, 5, 7, 9]]
strnodes = ['$' + str(x) + '^3$' for x in [1, 2, 3, 4, 5, 7, 9]]
ax.set_xscale("log", nonposx='clip')
#ax.set_yscale()
#ax.set_yscale("log", nonposy='clip')
#ax.set_yticks(nodes)
#ax.set_yticklabels(np.array(nodes, dtype=int))
ax.set_xticks(nodes)
ax.set_xticklabels(strnodes)
ax.plot(nodes, np.ones(len(nodes)), "k--", label="ideal")
ax.set_xlabel("Nodes")
ax.set_ylabel("Efficiency $E = \dfrac{t_1}{t_N}$")
plt.legend() #frameon=False)
myplt.set_font_sizes(ax)
myplt.make_grid(ax, True)
if blood:
myplt.save_figure(fig, 3, 'weak_scale_blood.pdf')
else:
myplt.save_figure(fig, 3, 'weak_scale_poiseuille.pdf')
plt.show()
def ic_far_field(waveguide, sourcecls, angle, nwaves, out_file, verbose=False):
"""Pretty shameless copy-paste of the exit_surface method follows"""
source = sourcecls(angle)
intensity = np.zeros(500)
xv = np.linspace(-waveguide.slab_gap, waveguide.slab_gap, 500)
#let's see if our source is monochromatic
prevval = source.get_wave()[1]
ismonochrome = True
for i in range(10):
newval = source.get_wave()[1]
ismonochrome = ismonochrome and prevval == newval
prevval = newval
if ismonochrome:
#get this done once
solver = modesolvers.ExpLossySolver(waveguide, prevval)
kx = solver.solve_transcendental(verbose=False) #this'll get real noisy otherwise
modewaves = solver.get_wavefunctions(kx, verbose=False)
pb = progressbar.ProgressBar(widgets=['Performing Incoherent Sum: ', progressbar.Percentage(), progressbar.Bar()],
maxval=nwaves)
if verbose:
pb.start()
#We need to iterate over many waves, summing their intensity
for i in xrange(nwaves):
#get a wave
(wave, wl) = source.get_wave()
inwave = lambda x: wave(x, 0) #only intersted in incidence surface
#solve modes for this wl
if not ismonochrome:
solver = modesolvers.ExpLossySolver(waveguide, wl)
kx = solver.solve_transcendental(verbose=False) #this'll get real noisy otherwise
modewaves = solver.get_wavefunctions(kx, verbose=False)
#split wave into modes
splitter = splitters.ModeSplitter(inwave, modewaves)
cm = splitter.get_coupling_constants()
wffull = splitter.get_wavefunction(cm)
#get wavefunction at exit surface
exitf = lambda x: wffull(x, waveguide.length)
#incoherently add exitf to the picture so far
exitvf = np.vectorize(exitf)
exitdata = exitvf(xv)
farfield = fft.fft(exitdata)
intensity = intensity + np.abs(fft.fftshift(farfield))**2
if verbose:
pb.update(i + 1)
#now plot it
(fig, ax) = plotting.setup_figure_standard(
title=ur'Propagated Intensity',
xlabel=u'Intensity at screen (arbitrary units)',
ylabel=u'Relative transverse distance on screen (m)')
ax.plot(xv, intensity, color='blue', linestyle='-', marker=None, antialiased=True)
ax.set_xlim((-waveguide.slab_gap, waveguide.slab_gap))
ax.set_ylim((np.min(intensity) - 0.1 * np.abs(np.min(intensity)),
np.max(intensity) + 0.1 * np.abs(np.max(intensity))))
plotting.shade_waveguide(ax, waveguide.slab_gap)
#Consider only the rightmost 4/5ths when setting the intensity
#lowest = np.min(intensity[:, 360:])
#highest = np.max(intensity[:, 360:])
#pcm = ax.pcolor(X, Y, intensity, cmap=matplotlib.cm.jet)
#ax.get_figure().colorbar(pcm, ax=ax, use_gridspec=True)
#ax.set_xlim((0, waveguide.length))
#ax.set_ylim((-waveguide.slab_gap, waveguide.slab_gap))
#bang down a slightly different guideline for the waveguide limits
#topr = Rectangle((0, waveguide.slab_gap / 2), waveguide.length, waveguide.slab_gap / 2, hatch='\\', fill=False)
#bottomr = Rectangle((0, -waveguide.slab_gap), waveguide.length, waveguide.slab_gap / 2, hatch='/', fill=False)
#ax.add_patch(topr)
#ax.add_patch(bottomr)
plotting.save_figure(fig, unicode(out_file))
#save the data too
sio.savemat(out_file + '.mat', {'intensity': intensity, 'xv' : xv})
