def fit_result(obj, model, data, noise, mask=False, regrid=False):
halo = obj.halo
ra = halo.ra.value
dec = halo.dec.value
bmin = halo.bmin
bmaj = halo.bmaj
scale = 1.
xlabel = 'RA [Deg]'
ylabel = 'DEC [Deg]'
scale = 1.
#if mask:
image_mask = obj.image_mask
if regrid:
data = utils.regridding(obj.halo,data, decrease_fov=True)
model = utils.regridding(obj.halo,model)
#if mask:
image_mask = utils.regridding(obj.halo, obj.image_mask*u.Jy, mask= not obj.halo.cropped).value
noise = utils.findrms(data.value)*u.Jy
scale = (np.array((bmin.value,bmaj.value))/halo.pix_size).value
bmin = bmin/(scale[0]*halo.pix_size)
bmaj = bmaj/(scale[1]*halo.pix_size)
ra = np.arange(0,data.shape[1])#halo.ra.value
dec = np.arange(0,data.shape[0])#halo.dec.value
xlabel = 'Pixels'
ylabel = 'Pixels'
#plt.imshow(image_mask)
#plt.show()
fig, axes = plt.subplots(ncols=3, nrows=1, sharey=True)
for axi in axes.flat:
axi.xaxis.set_major_locator(plt.MaxNLocator(5))
axi.xaxis.set_major_formatter(ScalarFormatter(useOffset=False))
axi.yaxis.set_major_formatter(ScalarFormatter(useOffset=False))
fig.set_size_inches(3.2*5,5.1)
draw_sizebar(halo,axes[0], scale, regrid)
draw_ellipse(halo,axes[0], bmin, bmaj, regrid)
data = (data/halo.pix_area).to(uJyarcsec2).value
noise = (noise/halo.pix_area).to(uJyarcsec2).value
model = (model/halo.pix_area).to(uJyarcsec2).value
masked_data = np.copy(data)
#if mask:
if regrid:
masked_data[image_mask > obj.mask_treshold*image_mask.max()] =-10000.
else:
masked_data[image_mask==1]= -10000.
if regrid:
NORMres = mplc.Normalize(vmin=-2.*noise, vmax=1.*masked_data.max())
else: NORMres = mplc.Normalize(vmin=-2.*noise, vmax=1.*masked_data.max())
#Trying two different functions since names were changed in recent matplotlib 3.3 update.
try:
Normdiv = mplc.TwoSlopeNorm(vcenter=0., vmin=0.8*(data-model).min(), vmax=0.8*(data-model).max())
except:
Normdiv = mplc.DivergingNorm(vcenter=0., vmin=0.8*(data-model).min(), vmax=0.8*(data-model).max())
im1 = axes[0].imshow(masked_data,cmap='inferno', origin='lower',
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres)
LEVEL = np.array([1,2,4,8,16,32,64,128,256,512,1024,2048,4096])*noise
cont1 = axes[0].contour(model,colors='white', levels=LEVEL, alpha=0.6,
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres,linewidths=1.)
cont2 = axes[0].contour(masked_data,colors='lightgreen', levels=np.array([-9999.8]),
alpha=0.6, linestyles='-',extent=(ra.max(),ra.min(),dec.min(),dec.max()),
norm = NORMres,linewidths=1.5)
axes[0].annotate('$V(x,y)$',xy=(0.5, 1), xycoords='axes fraction',
fontsize=titlesize, xytext=(0, -9), textcoords='offset points',
ha='center', va='top', color='white')
axes[0].set_title("Radio data", fontsize=titlesize)
axes[0].set_xlabel(xlabel, fontsize=labelsize)
axes[0].set_ylabel(ylabel, fontsize=labelsize)
axes[0].grid(color='white', linestyle='-', alpha=0.25)
plt.tight_layout()
im2 = axes[1].imshow(model,cmap='inferno', origin='lower',
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres)
axes[1].annotate('$I(x,y)$',xy=(0.5, 1), xycoords='axes fraction',
fontsize=titlesize, xytext=(0, -9), textcoords='offset points',
ha='center', va='top', color='white')
axes[1].set_title(obj.modelName.replace('_',' ')+" model", fontsize=titlesize)
axes[1].set_xlabel(xlabel, fontsize=labelsize)
axes[1].grid(color='white', linestyle='-', alpha=0.25)
cbar = fig.colorbar(im2,ax=axes[1])
cbar.ax.set_ylabel('$\\mu$Jy arcsec$^{-2}$',fontsize=labelsize)
#cbar.formatter = ScalarFormatter(useMathText=False)
#cbar.formatter = ticker.LogFormatter(base=10.,labelOnlyBase=True)
#cbar.formatter = ticker.StrMethodFormatter('%.2f')
plt.tight_layout()
im3 = axes[2].imshow(data-model, cmap='PuOr_r', origin='lower',
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = Normdiv)
cont4 = axes[2].contour(masked_data,
colors='red', levels=np.array([-9999.8]), alpha=0.6, linestyles='-',
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres,linewidths=1.5)
try:
cont3 = axes[2].contour(model, alpha=0.7, colors='black', levels=[2*noise],
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm=NORMres)
axes[2].clabel(cont3, fontsize=12, inline=1, fmt='2$\\sigma_{\\mathrm{rms}}$',colors='black')
except: pass
axes[2].annotate('$V(x,y)-I(x,y)$',xy=(0.5, 1), xycoords='axes fraction',
fontsize=titlesize, xytext=(0, -9), textcoords='offset points',
ha='center', va='top', color='black')
axes[2].set_title("Residual image", fontsize=titlesize)
axes[2].set_xlabel(xlabel, fontsize=labelsize)
axes[2].grid(color='black', linestyle='-', alpha=0.25)
plt.tight_layout()
import matplotlib.ticker as ticker
cbar = fig.colorbar(im3,ax=axes[2])
cbar.ax.set_ylabel('$\\mu$Jy arcsec$^{-2}$',fontsize=labelsize)
#cbar.formatter = ScalarFormatter(useMathText=False)
#cbar.formatter = ticker.LogFormatter(base=10.,labelOnlyBase=True)
#cbar.formatter = ticker.StrMethodFormatter('%.2f')
if regrid:
plt.savefig(halo.plotPath +halo.file.replace('.fits','')+'_mcmc_model'+obj.filename_append+'_REGRID.pdf')
else:
plt.savefig(halo.plotPath +halo.file.replace('.fits','')+'_mcmc_model'+obj.filename_append+'.pdf')
#plt.show()
plt.clf()
plt.close(fig)
def make_plots(df, plot_folder, area_unit, unit_colors, data_type, plot_scope, name_dict, do_all):
max_value = df.values.max()
N = 12
last_file = None
# replace area unit numeric codes with actual names
# this is for USA counties only
if isinstance(name_dict, pd.DataFrame):
for c in df.columns:
if c not in name_dict.index.tolist():
# if FIPS codes are not in population data
# we don't draw represent them on the map anyway
# so let's delete them from the plots as well
df.drop(columns=[c], inplace=True)
continue
name = name_dict.loc[c, 'name']
df.rename(columns={c: name}, inplace=True)
unit_colors[name] = unit_colors.pop(c)
if do_all:
day_list = df.index.tolist()
else:
day_list = df.index.tolist()[-1:]
for day in day_list:
# make it look like YYYY/MM/DD
daystring = lfill_date(day)
plot_file = 'frame-' + daystring.replace('/', '') + '.png'
plot_full_path = os.path.join(plot_folder, plot_file)
# if the file exists, skip this frame
if os.path.exists(plot_full_path):
continue
plt.rcParams["figure.dpi"] = 192
fig = plt.figure(figsize=(10, 5.625))
plt.title(plot_scope + ' COVID-19 daily cases - ' + data_type + ' - linear scale - top ' + str(N) + ' ' + area_unit)
# x ticks get too crowded, limit their number
plt.gca().xaxis.set_major_locator(plt.MaxNLocator(nbins=7))
plt.gca().spines['right'].set_visible(False)
plt.gca().spines['top'].set_visible(False)
# extract one day from dataframe
# put area units in one column
# put values in the other column
daydf = df.loc[day].to_frame()
daydf.index.name = area_unit
daydf.reset_index(inplace=True)
topN = daydf.sort_values(by=[day], ascending=False).head(N)
topNunits = topN[area_unit].tolist()
all_champs = df.loc[:day, topNunits]
#max_value = all_champs.values.max()
plt.gca().set_ylim(bottom=0, top=max_value)
for u in topNunits:
champ = df.loc[:day, u].to_frame()
p = plt.plot(champ.index.tolist(), champ[u].tolist(), color=unit_colors[u])
leg = plt.legend(topNunits, loc='upper left', frameon=False)
for line, text in zip(leg.get_lines(), leg.get_texts()):
line.set_color(unit_colors[text.get_text()])
fig.subplots_adjust(left = 0.07, right = 0.99, bottom = 0.065, top = 0.94)
fig.savefig(plot_full_path)
last_file = plot_full_path
plt.close()
curr_snap = plot_scope + '_' + data_type + '_top.png'
curr_snap = curr_snap.replace(' ', '_')
curr_snap = curr_snap.lower()
if last_file != None:
shutil.copyfile(last_file, curr_snap)
np.random.seed(42)
x0 = [-4.0, 5.0]
s = SQUARE_FIGSIZE
s[0] *= 2
s[1] *= 0.8
fig, axes = pl.subplots(1, 3, sharex=True, sharey=True, figsize=s)
for n, ax in zip([0.0, -1.0, 2.0], axes):
chain, _ = run_mcmc(log_p_gauss,
np.array(x0),
nsteps=2e3,
prop_sigma=10**n)
ax.plot(chain[:, 0], chain[:, 1], "o-", color=COLORS["DATA"], ms=2)
ax.plot(x0[0], x0[1], "o", color=COLORS["MODEL_1"])
ax.set_xlim(-6.3, 6.3)
ax.set_ylim(-6.3, 6.3)
ax.set_xlabel("$x$")
ax.annotate(r"$\sigma_q = 10^{{{0:.0f}}}$".format(n), (1, 0),
xycoords="axes fraction",
xytext=(-5, 5),
textcoords="offset points",
ha="right",
va="bottom")
ax.yaxis.set_major_locator(pl.MaxNLocator(5))
ax.xaxis.set_major_locator(pl.MaxNLocator(5))
axes[0].set_ylabel("$y$")
savefig(fig, "MH_sigma.pdf")
cut_wav = signa[int(begs[i] * samp_rate):int(ends[i] * samp_rate)]
wavfile.write(
'{}/silaba_{}.wav'.format(templates_folder, syl_type),
samp_rate, cut_wav)
ncol = syl_classes[syl_type]
ax[0][ncol].plot(cut_wav)
ax[0][ncol].set_xticklabels([])
ax[0][ncol].set_yticklabels([])
ax[0][ncol].set_title(syl_type)
ax[1][ncol].specgram(cut_wav,
NFFT=NFFT,
Fs=samp_rate,
noverlap=int(NFFT * p_overlap),
cmap='Greys')
ax[1][ncol].set_yticklabels([])
ax[1][ncol].xaxis.set_major_locator(plt.MaxNLocator(2))
ax[1][ncol].set_ylim(min_freq, max_freq)
# conversion from times to 'bins' vector coordinates
# 'bins' = center times of spec
time_beg_coord = (np.abs(t_bins - begs[i])).argmin()
time_end_coord = (np.abs(t_bins - ends[i])).argmin()
for j in range(time_beg_coord, time_end_coord):
# Taking the frame from the spectrogram
training_frame = signa_relevant[j - half_window:j + 1 +
half_window]
training_set.append(training_frame)
frame_class.append(syl_id)
all_training_set += training_set
fig.tight_layout()
# %%
# Shuffle the training list (we have to shuffle things together)
solution = np.arange(len(gp))
solution[0] = v0[0]
solution[1] = v1[0]
solution[2] = v2[0]
solution[3:] = aa[:, 0]
gp.set_parameter_vector(solution)
# Make the maximum likelihood prediction
t = np.linspace(min(x), max(x), 10000)
mu, var = gp.predict(y, t, return_var=True)
std = np.sqrt(var)
# Plot the data
plt.figure()
color = "#ff7f0e"
plt.errorbar(x, y, yerr=yerr, fmt=".k", capsize=0)
plt.plot(t, mu, color=color)
plt.fill_between(t,
mu + std,
mu - std,
color=color,
alpha=0.3,
edgecolor="none")
plt.ylabel(r"$y$")
plt.xlabel(r"$t$")
plt.gca().yaxis.set_major_locator(plt.MaxNLocator(5))
plt.title("maximum likelihood prediction")
plt.savefig('HD85390-5-prediction.png')
os.chdir('..')
print " - omega_0 = %.10g" % omega_best
# do a fit to the first 4 Fourier components
mtf = MultiTermFit(omega_best, 4)
mtf.fit(t, y, dy)
phase_fit, y_fit, phased_t = mtf.predict(1000, return_phased_times=True)
# plot the phased data and best-fit curves
ax = fig.add_subplot(321 + i)
ax.errorbar(phased_t, y, dy, fmt='.k', ecolor='gray',
lw=1, ms=4, capsize=1.5)
ax.plot(phase_fit, y_fit, '-b', lw=2)
ax.set_xlim(0, 1)
ax.set_ylim(plt.ylim()[::-1])
ax.yaxis.set_major_locator(plt.MaxNLocator(4))
ax.text(0.03, 0.04, "ID = %i" % ids[i], ha='left', va='bottom',
transform=ax.transAxes)
ax.text(0.03, 0.96, "P = %.2f hr" % (2 * np.pi / omega_best * 24.),
ha='left', va='top',
transform=ax.transAxes)
ylim = ax.get_ylim()
ax.set_ylim(ylim[0], ylim[0] + 1.1 * (ylim[1] - ylim[0]))
if i < 4:
ax.xaxis.set_major_formatter(plt.NullFormatter())
if i % 2 == 0:
ax.set_ylabel('mag')
def create(self,):
if(any([item.updated for item in self.parent.custom_metrics])):
if(len(self.parent.custom_metrics)!=self.parent.n_custom_plots):
warnings.warn("Data provided does not match the number of custom plots")
self.parent.total_plots=len(self.parent.custom_metrics)
self.parent.n_custom_plots=len(self.parent.custom_metrics)
else:
self.parent.total_plots=self.parent.n_custom_plots
if((self.parent.top_rows)*(self.parent.top_cols)<self.parent.total_plots):
warnings.warn("Total number of plots does not match the number of rows, number of rows has been increased to \
account for the difference")
while(((self.parent.top_rows)*(self.parent.top_cols)<self.parent.total_plots)):
self.parent.top_rows+=1
n_splits= self.parent.n_splits
# len(cm_df_overall.index)//2
self.parent.figure = plt.figure(figsize=self.parent.figsize)
self.parent.main_grid = gridspec.GridSpec(self.parent.nrows,
self.parent.ncols,
hspace=self.main_grid_hspace,
wspace=self.main_grid_wspace)
self.parent.top_cell = self.parent.main_grid[0,0:]
self.parent.bottom_cell = self.parent.main_grid[1,0:]
self.parent.inner_grid_top = gridspec.GridSpecFromSubplotSpec(
(self.parent.top_rows*self.parent.plot_width)+(self.parent.top_rows-1),
(self.parent.top_cols*self.parent.plot_height)+(self.parent.top_cols-1),
self.parent.top_cell,hspace=1
)
self.parent.bottom_cell = self.parent.main_grid[1,0:]
self.parent.inner_grid_bottom = gridspec.GridSpecFromSubplotSpec(n_splits,3, self.parent.bottom_cell)
self.parent.top_axes=[]
for j,k in enumerate(range(0,self.parent.top_rows*self.parent.plot_width,self.parent.plot_width)):
for l,m in enumerate(range(0,self.parent.top_cols*self.parent.plot_height,self.parent.plot_height)):
temp=plt.subplot(self.parent.inner_grid_top[k+j:k+self.parent.plot_height+j,
l+m:(l-1)+m+self.parent.plot_width])
self.parent.top_axes.append(temp)
self.parent._avg_axes=[]
for i,item in enumerate(self.parent.top_axes):
if(i<self.parent.total_plots):
if(self.parent.custom_metrics[i].average):
ax_avg=item.twiny().twinx()
self.parent._avg_axes.append(ax_avg)
# self.parent.middle_cell = self.parent.main_grid[1,0:]
# self.parent.inner_grid_middle = gridspec.GridSpecFromSubplotSpec(n_splits,3, self.parent.middle_cell)
# middle_axes=[]
#
#adds the main plots
for i in range(0,self.parent.n_custom_plots):
try:
if(self.parent.custom_metrics[i].updated):
custom_data=self.parent.custom_metrics[i].window()
if(custom_data.empty):
custom_data=pd.DataFrame([0,0,0,0],columns=['No Data Available Yet'])
#This can be optimized later
# top_axes[i].clear()
self.parent.top_axes[i].plot(custom_data.iloc[-1*self.parent.custom_metrics[i].w_size:,:])
self.parent.top_axes[i].legend(self.parent.custom_metrics[i].window().columns)
self.parent.top_axes[i].set_title(self.parent.custom_metrics[i].name)
self.parent.top_axes[i].set_ylabel('')
self.parent.top_axes[i].set_xlabel('')
if(self.parent.custom_metrics[i].xaxis_int):
self.parent.top_axes[i].xaxis.set_major_locator(MaxNLocator(integer=True))
#limit the number of ticks
if(self.parent.custom_metrics[i].n_ticks):
_k,_l=self.parent.custom_metrics[i].n_ticks
self.parent.top_axes[i].xaxis.set_major_locator(plt.MaxNLocator(_k))
self.parent.top_axes[i].yaxis.set_major_locator(plt.MaxNLocator(_l))
# self.parent.top_axes[i].set_xticks(self.parent.top_axes[i].get_xticks()[::2])
if(self.parent.custom_metrics[i].show_grid):
item.grid()
if(self.parent.custom_metrics[i].average):
self.parent._avg_axes[i].clear()
avg=self.parent.custom_metrics[i].means
self.parent._avg_axes[i].plot(avg,linestyle='--',alpha=0.6)
except Exception as error:
print(error, 'Happened while adding main plots.')
# Adds table
split_df=[]
bottom_axes=[]
self.tail()
if(self.parent.show_table):
if(not self.parent.main_results.empty):
for i,item in enumerate(self.chunks_df(self.parent.main_results.round(6).T,
math.ceil(len(self.parent.main_results.columns)/2),'r')):
if((i+1)%2==0 and (i>0)):
loc='bottom right'
else:
loc='bottom'
if(not item.empty):
split_df.append(item)
temp=plt.subplot(self.parent.inner_grid_bottom[i])
temp.axis('off')
temp.axis('tight')
table=temp.table(cellText=item.values,
loc=loc,
cellLoc='left',
rowLabels=item.index,
# colLabels=item.columns,
)
bottom_axes.append(table)
table.scale(1, 2)
self.parent.split_df=split_df
self.parent.bottom_axes=bottom_axes
#moves legends to the bottom of the plots
for i,item in enumerate(self.parent.top_axes[:self.parent.n_custom_plots]):
box = item.get_position()
item.set_position([box.x0, box.y0 + box.height * 0.1,box.width, box.height * 0.9])
lines=item.get_lines()
if(custom_data.empty):
custom_data=pd.DataFrame([0,0,0,0],columns=['No Data Available Yet'])
lines[0].get_xydata()
item.legend(self.parent.custom_metrics[i].window().columns,loc='upper center',
bbox_to_anchor=(0.5, -0.1),fancybox=True, shadow=True, ncol=5)
#add scientific formatting
# item.yaxis.set_major_formatter(mtick.FormatStrFormatter('%.2e'))
#Removes empty plots
for i,ax in enumerate(self.parent.top_axes):
lines=ax.get_lines()
try:
lines[0].get_xydata()
except Exception as error:
self.parent.top_axes[i].axis('off')
plt.show()
for item in self.parent.custom_metrics:
if(item.updated):
item.updated=False
# plt.close(self.parent.figure)
clear_output(wait=True)
gc.collect()
self.parent.counter+=1
def getMLPlotExt(y_ext, y_predict_ext, X_ext, flag,
nFeatures):
dffExt = pd.read_csv(myConfig.featurePathExt)
dffExt = dffExt.copy().dropna(axis=0, how='any').reset_index()
getTrendPlot1(dffExt, y_predict_ext, 'fracNa')
# Pb bandgap graph
flagComp = False
if flagComp == True:
#ctuple = [c[k] for k in key]
compounds = []
for index, row in dffExt.iterrows():
#row['counts']
counter = row['counts']
counter = counter.split('(')[1].split(')')[0]
counter = literal_eval(counter)
A_t = ''
B_t = ''
C_t = ''
for i in counter:
if i == 'Cs' or i == 'Rb' or i == 'Na' or i == 'K':
A = i
Ai = str(counter[i])
if counter[i] == 1: Ai = ''
A_t += A + '$_{' + Ai + '}$'
if i == 'Sn' or i == 'Ge' or i == 'Pb':
B = i
Bi = str(counter[i])
if counter[i] == 1: Bi = ''
B_t += B + '$_{' + Bi + '}$'
if i == 'Br' or i == 'Cl' or i == 'I':
C = i
Ci = str(counter[i])
if counter[i] == 1: Ci = ''
C_t += C + '$_{' + Ci + '}$'
compounds += [A_t + B_t + C_t]
fig, ax1 = plt.subplots()
plt.title('Pb Bandgap Predictions by Composition')
plt.ylabel('Bandgap (eV)')
x = range(len(compounds))
plt.xticks(x, compounds, rotation = 90)
#plt.xlabel('CsSnI$_{3}$')
fig.set_size_inches(8, 5)
#ax1 = plt.axes()
ax1.scatter(x, y_predict_ext, marker = 'o', color = 'xkcd:blue', label='Predicted')
ax1.scatter(x, y_ext, marker = 'o', color = 'xkcd:red', label = 'DFT')
plt.legend()
ax1.plot(x[1:7], y_predict_ext[1:7], linestyle = '--', color = 'xkcd:blue')
ax1.plot(x[1:7], y_ext[1:7], linestyle = '--', color = 'xkcd:red')
ax1.tick_params(direction='in', top=True, right=True)
ax1.yaxis.set_major_locator(plt.MaxNLocator(5))
plt.tight_layout()
path = '/Users/Jared/Dropbox/Master Thesis/code/codeOutputs/'
plt.savefig(path + 'predictPb1.png', dpi = 400, bbox_inches="tight")
#plt.show()
#print(dffExt['counts'])
#cols = [x for x in list(X_ext) if "frac" in x]
#print(cols)
flagComp = False
if flagComp == True:
savePath = '/Users/Jared/Dropbox/Master Thesis/code/codeOutputs/'
#plt.scatter(y_ext, y_predict_ext, alpha=0.8, color='#00ccff',
# marker='o', s=80)
my_dpi = 500
fig = plt.figure(figsize=(5, 5), dpi=my_dpi)
ymin = 1.08*min(y_ext) if min(y_ext) <=0 else 0.92*min(y_ext)
ymax = 1.08*max(y_ext) if max(y_ext) >=0 else 0.92*max(y_ext)
xmax = ymax
print(len(y_ext))
plt.ylabel('$E_{g}$ Prediction (eV)')
plt.xlabel('$E_{g}$ (eV)')
plt.title('Bandgap Prediction', y=1.04)
bandgapCs = np.array([0.213, ])
#plt.ylabel('$\Delta H_{f}$ Prediction (eV/atom)')
#plt.xlabel('$\Delta H_{f}$ (eV/atom)')
#plt.title('$\Delta H_{f}$ Prediction', y=1.04)
#plt.ylabel('ML Prediction (eV)')
#plt.xlabel('Bandgap (eV)')
#err = round(sqrt(mean_squared_error(y_ext, y_predict_ext)),3)
#plt.title('Bandgap Prediction')
plt.ylim(ymin, ymax)
plt.xlim(ymin, xmax)
#plt.legend(['Train', 'Test'])
plt.ylim(ymin, ymax)
plt.xlim(ymin, xmax)
ax = plt.axes()
xy = [ymin, ymax]
#ax.grid()
#plt.xticks(np.arange(ymin, ymax, 0.25))
#plt.yticks(np.arange(round(ymin, 2),
# round(ymax + 0.1, 2) + 0.25, 0.25))
#ax.xaxis.set_major_locator(plt.MaxNLocator(12))
#ax.yaxis.set_major_locator(plt.MaxNLocator(12))
ax.tick_params(direction='in', top=True, right=True)
ax.xaxis.set_major_locator(plt.MaxNLocator(6))
ax.yaxis.set_major_locator(plt.MaxNLocator(6))
p0 = plt.plot(xy, xy, 'k', zorder=1)
p1 = plt.scatter(y_ext, y_predict_ext, color='#00ccff',
marker='o', s=80, label='Train', zorder=2)
#plt.legend(['Train', 'Test'])
#plt.legend(handles=[p1])
#xy = np.arange(0, 3.2, .005)
#plt.plot(xy, xy, 'k', alpha = 0.75)
#plt.show()
plt.savefig(savePath + 'paper_predict_noCsnoRb_form.png', dpi=500, bbox_inches="tight")
def test_contourf_symmetric_locator():
# github issue 7271
z = np.arange(12).reshape((3, 4))
locator = plt.MaxNLocator(nbins=4, symmetric=True)
cs = plt.contourf(z, locator=locator)
assert_array_almost_equal(cs.levels, np.linspace(-12, 12, 5))
def plotElements(self):
'''Plot elements and convective regions'''
# Initialize parameters
xLimit = None
yLimit = None
lines = []
# Plot preprocess
if self.ax is None:
self.ax = self.fig.add_subplot(1, 1, 1)
self.ax.set_xlabel("M/M$_\odot$")
self.ax.set_ylabel("Mass fraction")
self.ax.set_yscale("log")
# Manage plot attributes beforehand
if self.loadedModels:
xLimit = self.pltAtrb["xrange"]
if xLimit is None:
xLimit = [self.plotMasses[0], self.plotMasses[-1]]
yLimit = self.pltAtrb["yrange"]
if yLimit is None:
yLimit = [1e-24, 1]
# Plot convective regions
if self.showConve:
for region in self.plotConvRegions:
self.ax.fill_between(region,
yLimit[0],
yLimit[1],
facecolor="none",
hatch="/",
edgecolor="k")
for ii in range(len(self.pltAtrb["elements"])):
name = self.pltAtrb["elements"][ii]
capel = name[0].upper() + name[1:]
if self.dashElements:
lines.append(
self.ax.plot(self.plotMasses,
self.plotListsOfData[ii],
"--",
label=capel,
lw=2))
else:
lines.append(
self.ax.plot(self.plotMasses,
self.plotListsOfData[ii],
label=capel,
lw=2))
# Plot mesh
if self.showMesh:
constLevel = [
math.sqrt(yLimit[0] * yLimit[1]) for x in self.plotBordMass
]
lines.append(
self.ax.plot(self.plotBordMass,
constLevel,
"r.",
label="Mesh",
lw=2))
# Plot grid
if self.showGrid:
nHz = int(math.log10(yLimit[1] / yLimit[0]))
for ii in range(nHz):
val = 10**(int(math.log10(yLimit[0])) + ii)
self.ax.plot(xLimit, (val, val), "r-")
val = xLimit[0] + ii * (xLimit[1] - xLimit[0]) / (nHz - 1)
self.ax.plot((val, val), yLimit, "r-")
# Plot temperature
if self.showTemp:
if self.ax2 is None:
self.ax2 = self.ax.twinx()
self.ax2.set_ylabel("Temperature (K)")
self.ax2.set_yscale("log")
self.ax2.set_ylim([1e5, 1e9])
lines.append(
self.ax2.plot(self.plotMasses,
self.plotTemp,
"k--",
label="Temperature",
lw=2))
# Plot neutron density
elif self.showRho:
if self.ax2 is None:
self.ax2 = self.ax.twinx()
self.ax2.set_ylabel("Neutron density (n/cm$^3$)")
self.ax2.set_yscale("log")
lines.append(
self.ax2.plot(self.plotMasses,
self.plotRho,
"k--",
label="Neutron density",
lw=2))
# Unpack lines
lins = []
for line in lines:
lins += line
lines = lins
labs = [l.get_label() for l in lines]
self.ax.legend(lines, labs, prop={"size": 10})
# Limits
self.ax.set_xlim(xLimit)
self.ax.set_ylim(yLimit)
self.ax.xaxis.set_major_locator(plt.MaxNLocator(nbins=5))
self.ax.xaxis.set_major_formatter(ScalarFormatter(useOffset=False))
def getMLPlot(train_y, test_y, y_predict_train, y_predict_test, flag,
nFeatures):
flag = True
train_y = 1000*train_y
test_y = 1000*test_y
y_predict_train = 1000*y_predict_train
y_predict_test = 1000*y_predict_test
if flag == True:
savePath = '/Users/Jared/Dropbox/Master Thesis/code/codeOutputs/'
my_dpi = 500
fig = plt.figure(figsize=(5, 5), dpi=my_dpi)
#ymax = 3.2
#xmax = ymax
ymin = 1.08*min(train_y) if min(train_y) <=0 else 0.92*min(train_y)
ymax = 1.08*max(train_y) if max(train_y) >=0 else 0.92*max(train_y)
xmax = ymax
plt.ylabel('$E_{g}$ Prediction (eV)')
plt.xlabel('$E_{g}$ (eV)')
plt.title('Bandgap Prediction', y=1.04)
plt.ylabel('$\Delta E_{hull}$ Prediction (meV/atom)')
plt.xlabel('$\Delta E_{hull}$ (meV/atom)')
plt.title('$\Delta E_{hull}$ Prediction', y=1.04)
#plt.ylabel('$\Delta H_{f}$ Prediction (eV/atom)')
#plt.xlabel('$\Delta H_{f}$ (eV/atom)')
#plt.title('$\Delta H_{f}$ Prediction', y=1.04)
plt.ylim(ymin, ymax)
plt.xlim(ymin, xmax)
xy = [ymin, ymax] #np.arange(ymin, ymax, .0005)
ax = plt.axes()
#ax.grid()
#plt.xticks(np.arange(ymin, ymax, 0.25))
#plt.yticks(np.arange(round(ymin, 2),
# round(ymax + 0.1, 2) + 0.25, 0.25))
ax.xaxis.set_major_locator(plt.MaxNLocator(6))
ax.yaxis.set_major_locator(plt.MaxNLocator(6))
ax.tick_params(direction='in', top=True, right=True)
p0 = plt.plot(xy, xy, 'k', zorder=1)
p1 = plt.scatter(train_y, y_predict_train, color='#00ccff',
marker='o', s=80, label='Train', zorder=2)
p2 = plt.scatter(test_y, y_predict_test, color = '#ffb31a',
marker='o', s=80, label = 'Test', zorder =3)
#plt.legend(['Train', 'Test'])
plt.legend(handles=[p1, p2])
#plt.plot(xy, xy, 'k', alpha = 1.0)
plt.savefig(savePath + 'paper_ML_example_d2_junk.png',
bbox_inches="tight")
flag = False
#plt.savefig(savePath + 'ML_example_d1.png')
plt.show()
for i in range(10, 21):
for f in range(1, 13):
#print(len(f, 2000i, sep = '.')
month = str(f)
year = str(i)
date = month + "." + year
dates.append(date)
fig, ax = plt.subplots()
plt.tight_layout()
ax.plot(dates, monthlyDev, color=('xkcd:emerald'))
ax.axhline(y=0, color='orange', linestyle='solid')
ax.xaxis.set_major_locator(plt.MaxNLocator(12))
#ax.set_xticklabels(dates, rotation = 90)
plt.xlabel("Month and Year")
plt.ylabel("Monthly Deviation (NDVI)")
plt.title("Monthly deviation of NDVI - Sichuan Province from 2010 to 2021")
plt.style.use("seaborn-dark")
for param in ['figure.facecolor', 'axes.facecolor', 'savefig.facecolor']:
plt.rcParams[param] = '#212946' # bluish dark grey
for param in ['text.color', 'axes.labelcolor', 'xtick.color', 'ytick.color']:
plt.rcParams[param] = '0.9' # very light grey
ax.grid(
color='#2A3459') # bluish dark grey, but slightly lighter than background
plt.tight_layout()
def summarize_results(model_path, dest_path=None, fig_formats=['pdf']):
if dest_path is None:
dest_path = op.split(model_path)[0]
results = np.load(model_path)
y_pred = results['y_pred']
y = results['y']
fs = results['samplerate']
model = results['estimator'].item()
cv_results = results['cv_results'].item()
if 'frequency' in results:
freqs = results['frequency']
else:
feature_extractor = STFTFeatureExtractor(nfft=2 * 256,
shift=1,
samplerate=fs,
f_lower=0,
f_upper=10.,
winlen=2 * 256)
freqs = feature_extractor.get_frequencies()
if 'n_channels' in results:
n_channels = results['n_channels']
else:
n_channels = 3
all_labels = np.unique(MOUSE_GROUPS_SIMPLE.keys())
# ----- show in-sample predictions -----
fig, ax = plt.subplots()
t = np.arange(y.shape[0]) / float(fs)
ax.plot(t, y, 'r-', lw=1, label='Label')
ax.plot(t, y_pred, '-', color=3 * [.25], lw=.5, label='Predicted')
ax.set_xlabel('Time (s)')
ax.set_xlim(t[0], t[-1])
ax.legend(loc='best', fontsize=8)
ax.set_yticks(all_labels)
ax.set_yticklabels([MOUSE_GROUPS_SIMPLE[k] for k in all_labels])
llutil.set_font_axes(ax, add_size=2)
llutil.simple_xy_axes(ax)
fig.set_size_inches(10, 3)
fig.tight_layout()
if dest_path is not None:
for ff in fig_formats:
fig.savefig(op.join(dest_path, 'prediction_insample.' + ff),
format=ff)
# ----- show separating hyperplanes -----
n_classes = len(model.classes_)
colors = [
llutil.get_nice_color('blue'),
llutil.get_nice_color('red'),
llutil.get_nice_color('gray')
]
if n_classes > 2:
# one hyperplane per class for OVR classifier
W = model.coef_
else:
# single hyperplance for binary classifier
W = np.vstack((model.coef_, model.coef_))
fig, axarr = plt.subplots(nrows=n_classes,
ncols=1,
sharex=True,
sharey=True)
for i, cls in enumerate(model.classes_):
ax = axarr[i]
w = W[i, :]
n = w.shape[0] / n_channels
for j in range(n_channels):
ax.plot(freqs, w[j * n:(j + 1) * n], '-', color=colors[j], lw=1.5)
ax.set_xlabel('Frequency (Hz)')
ax.set_ylabel('Weight')
ax.set_title('class {} ({})'.format(int(cls),
MOUSE_GROUPS_SIMPLE[cls]))
llutil.set_font_axes(ax, add_size=2)
llutil.simple_xy_axes(ax)
ax.yaxis.set_major_locator(plt.MaxNLocator(3))
ax.set_ylim(W.min(), W.max())
fig.set_size_inches(5, 1.1 * n_classes)
fig.tight_layout()
if dest_path is not None:
for ff in fig_formats:
fig.savefig(op.join(dest_path, 'classifier_weights.' + ff),
format=ff)
# ----- error rates -----
fig, ax = plt.subplots()
xticklabels = [
MOUSE_GROUPS_SIMPLE[k].replace(' ', '\n') for k in all_labels
]
tpr = np.mean(cv_results['tpr'], axis=1)
fpr = np.mean(cv_results['fpr'], axis=1)
ax.bar(all_labels, tpr, width=0.35, color=3 * [.2], label='TPR')
ax.bar(all_labels + 0.35, fpr, width=0.35, color=3 * [.75], label='FPR')
ax.set_xticks(all_labels + .35)
ax.set_yticks([0, .25, .5, .75, 1])
# ax.set_xticklabels([groups[k] for k in all_labels])
ax.set_xticklabels(xticklabels)
ax.set_ylabel('Rate')
ax.set_xlim(0, all_labels.max() + 1)
ax.set_ylim(0, 1.2)
ax.axhline(1, linestyle='-', color=3 * [0], linewidth=1.5)
ax.axhline(.9, linestyle='--', color=3 * [.5])
ax.axhline(.1, linestyle='--', color=3 * [.5])
ax.legend(loc='best', fontsize=8)
llutil.set_font_axes(ax, add_size=3)
llutil.simple_xy_axes(ax)
fig.set_size_inches(7, 3)
fig.tight_layout()
if dest_path is not None:
for ff in fig_formats:
fig.savefig(op.join(dest_path, 'error_rates.' + ff), format=ff)
plt.show()
def train_classifier(rec_paths,
run_cv=True,
samplerate=100.,
dest_path=None,
algorithm='SVM',
use_cluster=False,
f_upper=10.):
# ----- parameters -----
n_folds = 4
feature_extractor = STFTFeatureExtractor(nfft=2 * 256,
shift=1,
samplerate=samplerate,
f_lower=0,
f_upper=f_upper,
winlen=2 * 256)
if dest_path is None:
dest_path = op.join(op.expanduser('~'), 'research', 'data',
'experiments', 'BehavioralScoring')
llutil.makedirs_save(dest_path)
fig_formats = ['pdf', 'png']
# ----- load and convert data -----
X, y, fs, n_channels, groups = load_data(rec_paths,
feature_extractor,
samplerate=samplerate)
assert fs == samplerate
freqs = feature_extractor.get_frequencies()
all_labels = np.asarray(groups.keys())
n_labels = len(all_labels)
print "total data size: {} observations x {} features".format(
X.shape[0], X.shape[1])
# ----- set up estimation/validation methods -----
if algorithm.upper() == 'SVM':
if use_cluster:
tmp_path = op.join(dest_path, 'tmp')
model = ParallelLinearSVC(tmp_path=tmp_path, local=False)
else:
cv = cross_validation.StratifiedKFold(y,
n_folds=n_folds,
shuffle=True,
random_state=0)
minor_version = get_sklearn_minor_version()
if minor_version <= 16:
class_weight = 'auto'
else:
class_weight = 'balanced'
C_values = 2.**np.linspace(-3, 5, 10)
model = LinearSVM(n_folds=n_folds,
verbose=False,
class_weight=class_weight,
dual=False,
penalty='l2',
max_iter=1000,
C_values=C_values)
elif algorithm.upper() == 'LDA':
priors = []
for i, label in enumerate(all_labels):
n = np.sum(y == label)
if n > 0:
priors.append(n / float(y.shape[0]))
model = lda.LDA(priors=np.asarray(priors))
# ----- cross-validation -----
if run_cv:
# cross-validation
cv = cross_validation.StratifiedKFold(y,
n_folds=n_folds,
shuffle=True,
random_state=0)
cv_results = {
'score': np.zeros((n_folds, )),
'tpr': np.zeros((n_labels, n_folds)),
'fpr': np.zeros((
n_labels,
n_folds,
))
}
for i, (train_ind, test_ind) in enumerate(cv):
print("Fold {}/{}".format(i + 1, n_folds))
model.fit(X[train_ind, :], y[train_ind])
cv_results['score'][i] = model.score(X[test_ind], y[test_ind])
y_hat = model.predict(X[test_ind, :])
tpr, fpr = compute_error_rates(y[test_ind], y_hat, all_labels)
cv_results['tpr'][:, i] = tpr
cv_results['fpr'][:, i] = fpr
print "Cross-validation results:"
for k in ['score', 'tpr', 'fpr']:
print "{}: {:.2f} += {:.2f}".format(k, np.mean(cv_results[k]),
np.std(cv_results[k]))
else:
cv_results = None
# ----- fit model using all data (and do in-sample prediction) -----
model.fit(X, y)
# clf = model.best_estimator_
y_pred = model.predict(X)
if isinstance(model, LinearSVM):
print "Best misclassification cost:", model.best_param_
tp_rate, fp_rate = compute_error_rates(y, y_pred, all_labels)
print "in-sample prediction error rate:", np.mean(y_pred != y)
for i, g in enumerate(all_labels):
print "label {} ({}): TPR={:.2f}, FPR={:.2f}".format(
g, groups[g], tp_rate[i], fp_rate[i])
# ----- save results -----
result_file = op.join(dest_path, 'classification_results.npz')
np.savez(result_file,
rec_paths=rec_paths,
run_cv=run_cv,
samplerate=samplerate,
cv_results=cv_results,
tp_rate=tp_rate,
fp_rate=fp_rate,
estimator=model,
groups=groups,
y=y.astype(np.uint8),
y_pred=y_pred.astype(np.uint8),
frequency=freqs,
n_channels=n_channels)
# ----- show in-sample predictions -----
fig, axarr = plt.subplots(nrows=2, ncols=1, sharex=True)
ax = axarr[0]
t = np.arange(y.shape[0]) / float(fs)
ax.plot(t, y, 'r-', lw=2, label='label')
ax.plot(t, y_pred, '--', color=3 * [.25], lw=1, label='pred')
ax.set_xlabel('Time (s)')
ax.set_ylabel('Label')
ax.set_xlim(t[0], t[-1])
ax.legend(loc='best')
ax = axarr[1]
ax.plot(t, y_pred != y, 'k-', lw=1)
ax.set_xlabel('Time (s)')
ax.set_ylabel('Misclassified')
ax.set_xlim(t[0], t[-1])
ax.set_ylim(0, 1.2)
for ax in axarr.flat:
# Hide the right and top spines
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
# Only show ticks on the left and bottom spines
ax.yaxis.set_ticks_position('left')
ax.xaxis.set_ticks_position('bottom')
ax.tick_params(axis='both', which='major', labelsize=8)
ax.xaxis.label.set_fontsize(9)
ax.xaxis.label.set_fontname('Arial')
ax.yaxis.label.set_fontsize(9)
ax.xaxis.label.set_fontname('Arial')
ax.xaxis.set_major_locator(plt.MaxNLocator(4))
ax.yaxis.set_major_locator(plt.MaxNLocator(4))
fig.set_size_inches(7, 3.5)
fig.subplots_adjust()
if dest_path is not None:
for ff in fig_formats:
fig.savefig(op.join(dest_path, 'prediction_insample.' + ff),
format=ff)
# ----- show separating hyperplanes -----
n_classes = len(model.classes_)
colors = [
llutil.get_nice_color('blue'),
llutil.get_nice_color('red'),
llutil.get_nice_color('gray')
]
if n_classes > 2:
# one hyperplane per class for OVR classifier
W = model.coef_
else:
# single hyperplance for binary classifier
W = np.vstack((model.coef_, model.coef_))
fig, axarr = plt.subplots(nrows=n_classes,
ncols=1,
sharex=True,
sharey=True)
for i, cls in enumerate(model.classes_):
ax = axarr[i]
w = W[i, :]
n = w.shape[0] / n_channels
for j in range(n_channels):
ax.plot(freqs, w[j * n:(j + 1) * n], '-', color=colors[j])
ax.set_xlabel('Frequency (Hz)')
ax.set_ylabel('Weight')
ax.set_title('class {} ({})'.format(int(cls), groups[cls]))
llutil.set_font_axes(ax, add_size=2)
llutil.simple_xy_axes(ax)
fig.set_size_inches(7, 2 * n_classes)
fig.tight_layout()
if dest_path is not None:
for ff in fig_formats:
fig.savefig(op.join(dest_path, 'classifier_weights.' + ff),
format=ff)
# ----- error rates -----
n_rows = 1 + int(run_cv)
fig, axarr = plt.subplots(nrows=n_rows, ncols=1, sharex=True, sharey=True)
axarr = np.atleast_1d(axarr)
xticklabels = [groups[k].replace(' ', '\n') for k in all_labels]
for i in range(n_rows):
if i == 0:
tpr = tp_rate
fpr = fp_rate
title = 'in-sample'
else:
tpr = np.mean(cv_results['tpr'], axis=1)
fpr = np.mean(cv_results['fpr'], axis=1)
title = 'cross-validated'
ax = axarr[i]
ax.set_title(title)
ax.bar(all_labels, tpr, width=0.35, color=3 * [.25], label='TPR')
ax.bar(all_labels + 0.35,
fpr,
width=0.35,
color=3 * [.75],
label='FPR')
ax.set_xticks(all_labels + .35)
ax.set_xticklabels(xticklabels)
ax.set_ylabel('Rate')
ax.set_xlim(0, all_labels.max() + 1)
ax.set_ylim(0, 1.2)
ax.axhline(1, linestyle='-', color=3 * [0], linewidth=1.5)
ax.axhline(.9, linestyle='--', color=3 * [.5])
ax.axhline(.1, linestyle='--', color=3 * [.5])
ax.legend(loc='best')
llutil.set_font_axes(ax, add_size=2)
llutil.simple_xy_axes(ax)
fig.set_size_inches(7, 5)
fig.tight_layout()
if dest_path is not None:
for ff in fig_formats:
fig.savefig(op.join(dest_path, 'error_rates.' + ff), format=ff)
plt.show()
def plot_mcmc(
samples, labels=None, priors=None, ptrue=None, precision=None, nbins=30, s=1.0
):
"""Plots a Giant Triangle Confusogram
Parameters
----------
samples: 2-D array, shape (N, ndim)
Samples from ndim variables to be plotted in the GTC
labels: list of strings, optional
List of names for each variable (size ndim)
priors: list of callables, optional
List of prior functions for the variables distributions (size ndim)
ptrue: list of floats, optional
List of true estimates for each parameter
precision: list of ints, optional
List of decimal places to write down for each parameter. Defaults to 2
nbins: int, optional
Number of bins to be used in 1D and 2D histograms. Defaults to 30
s: float, optional
Standard deviation of Gaussian filter applied to smooth histograms. Defaults to 1.0
"""
p = map(
lambda v: (v[1], v[1] - v[0], v[2] - v[1]),
zip(*np.percentile(samples, [16, 50, 84], axis=0)),
)
p = list(p)
ndim = samples.shape[-1]
fig = plt.figure(figsize=(8, 8))
fig.subplots_adjust(left=0.05, right=0.95, top=0.95, bottom=0.05)
# TODO: pyplot style context
grid = plt.GridSpec(ndim, ndim, wspace=0.0, hspace=0.0)
handles = []
if precision is None:
precision = 2 * np.ones(ndim, dtype=int)
if ptrue is None:
ptrue = np.array([None for _ in range(ndim)])
# PLOT 1D
for i in range(ndim):
ax = fig.add_subplot(grid[i, i])
H, edges = np.histogram(samples[:, i], bins=nbins, density=True)
centers = (edges[1:] + edges[:-1]) / 2
data = ndimage.gaussian_filter1d((centers, H), sigma=s)
data[1] /= data[1].sum()
(l1,) = ax.plot(data[0], data[1], "b-", lw=1, label="posterior")
if priors is not None:
pr = priors[i](centers)
pr /= pr.sum()
(l2,) = ax.plot(centers, pr, "k-", lw=1, label="prior")
l3 = ax.axvline(p[i][0], color="k", ls="--", label="median")
mask = np.logical_and(
centers - p[i][0] <= p[i][2], p[i][0] - centers <= p[i][1]
)
ax.fill_between(
centers[mask], np.zeros(mask.sum()), data[1][mask], color="b", alpha=0.3
)
if ptrue[i] is not None:
l4 = ax.axvline(ptrue[i], color="gray", lw=1.5, label="true")
if i < ndim - 1:
ax.set_xticks([])
else:
ax.tick_params(rotation=45)
ax.set_yticks([])
ax.get_xaxis().set_major_locator(plt.MaxNLocator(3))
ax.set_ylim(0)
if labels is not None:
ax.set_title(
"{0} = {1:.{4}f}$^{{+{2:.{4}f}}}_{{-{3:.{4}f}}}$".format(
labels[i], p[i][0], p[i][2], p[i][1], precision[i]
)
)
handles.append(l1)
try:
handles.append(l2)
except UnboundLocalError:
pass
try:
handles.append(l3)
except UnboundLocalError:
pass
try:
handles.append(l4)
except UnboundLocalError:
pass
# PLOT 2D
nbins_flat = np.linspace(0, nbins ** 2, nbins ** 2)
for i in range(ndim):
for j in range(i):
ax = fig.add_subplot(grid[i, j])
H, xi, yi = np.histogram2d(samples[:, j], samples[:, i], bins=nbins)
extents = [xi[0], xi[-1], yi[0], yi[-1]]
H /= H.sum()
H_order = np.sort(H.flat)
H_cumul = np.cumsum(H_order)
tmp = np.interp([0.0455, 0.3173, 1.0], H_cumul, nbins_flat)
chainlevels = np.interp(tmp, nbins_flat, H_order)
data = ndimage.gaussian_filter(H.T, sigma=s)
xbins = (xi[1:] + xi[:-1]) / 2
ybins = (yi[1:] + yi[:-1]) / 2
ax.contourf(
xbins,
ybins,
data,
levels=chainlevels,
colors=["#1f77b4", "#52aae7", "#85ddff"],
alpha=0.3,
)
ax.contour(data, chainlevels, extent=extents, colors="b")
ax.get_xaxis().set_major_locator(plt.MaxNLocator(3))
ax.get_yaxis().set_major_locator(plt.MaxNLocator(3))
if ptrue[i] is not None:
ax.axhline(ptrue[i], color="gray", lw=1.5)
if ptrue[j] is not None:
ax.axvline(ptrue[j], color="gray", lw=1.5)
if i < ndim - 1:
ax.set_xticks([])
else:
ax.tick_params(rotation=45)
if j > 0:
ax.set_yticks([])
else:
ax.tick_params(rotation=45)
fig.legend(handles=handles)
return fig
def plot_wind(data, specs, outputpath):
'''
routine plots vertical profiles of wind speed and direction
and saves plot as .png
INPUT:
- data: dictionnary with data (eg filled by read_ncfile())
- specs: dictionnary with filename specifications (filled
by read_ncfile())
- outputpath: path where png will be stored in.
OUTPUT: .png file stored in outputpath
'''
logging.info('now plotting wind speed and direction sounding.........')
# define outputname of .png-file:
variable = 'wind'
outputname = '{platform}_{instrument}{direction}_{variable}_{date}_{tempres}.png'.format(
platform=specs['platform_short'],
instrument=specs['type'].replace(' ', '').replace('_', ''),
direction=specs['direction'],
variable=variable,
date=specs['date'] + '_' + specs['time'],
tempres=specs['tempres'].replace(' ', ''))
fig, ax = plt.subplots(1, 2, sharey=True, figsize=(8, 6))
# plot the data into subpanels:
ax[0].plot(data['windSpeed'], data['altitude'], '.-k', markersize=1)
ax[1].plot(data['windDirection'], data['altitude'], '.-k', markersize=1)
# general cosmetics:
for i in range(2):
ax[i].spines['top'].set_visible(False)
ax[i].spines['right'].set_visible(False)
ax[i].spines['left'].set_visible(False)
ax[i].grid(axis='y', linestyle='-', color='gray')
ax[i].set_ylim(0, ax[i].get_ylim()[-1])
ax[i].xaxis.set_minor_locator(AutoMinorLocator())
ax[i].yaxis.set_minor_locator(AutoMinorLocator())
ax[i].xaxis.set_major_locator(plt.MaxNLocator(4))
ax[i].tick_params(top=False, right=False)
# axis labels:
ax[i].set_ylabel('Altitude [m]', fontsize=14)
# switch off minor ticks for top axis:
ax[i].tick_params(axis='x', which='minor', top=False)
# make labels larger for all ticks:
ax[i].tick_params(axis='both', labelsize=14)
# switch off some ticks and labels and spines manually.
ax[0].spines['left'].set_visible(True)
ax[1].tick_params(left=False)
ax[0].tick_params(right=False, which='minor', axis='y')
ax[1].spines['right'].set_visible(True)
ax[1].yaxis.set_ticks_position('right')
ax[1].yaxis.set_label_position('right')
# set wind direction axis to valid range:
ax[1].set_xlim(0, 360)
ax[0].set_xlabel('Wind Speed [m s$^{-1}$]', fontsize=14)
ax[1].set_xlabel('Wind Direction [$^\circ$]', fontsize=14)
plt.subplots_adjust(top=0.9, right=0.85, left=0.15)
fig.suptitle(
'%s, %s %sUTC' %
(specs['location'], specs['date'], data['time_of_launch_HHmmss'][:-2]),
fontsize=18)
fig.savefig(outputpath + outputname)
logging.info('Wind profile saved at {}'.format(outputpath + outputname))
def plotter(fdict):
""" Go """
import matplotlib
matplotlib.use('agg')
import matplotlib.pyplot as plt
from matplotlib.ticker import FormatStrFormatter
pgconn = psycopg2.connect(database='postgis', host='iemdb', user='******')
station = fdict.get('station', 'DMX')[:4]
phenomena = fdict.get('phenomena', 'SV')
significance = fdict.get('significance', 'W')
split = fdict.get('split', 'jan1')
nt = NetworkTable('WFO')
if split == 'jan1':
sql = """SELECT extract(year from issue)::int as year,
min(issue at time zone 'UTC') as min_issue,
max(issue at time zone 'UTC') as max_issue,
count(distinct eventid)
from warnings where wfo = %s and phenomena = %s and significance = %s
and issue is not null
GROUP by year ORDER by year ASC"""
else:
sql = """SELECT
extract(year from issue - '6 months'::interval)::int as year,
min(issue at time zone 'UTC') as min_issue,
max(issue at time zone 'UTC') as max_issue,
count(distinct eventid)
from warnings where wfo = %s and phenomena = %s and significance = %s
and issue is not null
GROUP by year ORDER by year ASC"""
df = read_sql(sql,
pgconn,
params=(station, phenomena, significance),
index_col=None)
# Since many VTEC events start in 2005, we should not trust any
# data that has its first year in 2005
if df['year'].min() == 2005:
df = df[df['year'] > 2005]
def myfunc(row):
year = row[0]
valid = row[1]
if year == valid.year:
return int(valid.strftime("%j"))
else:
days = (datetime.date(year + 1, 1, 1) -
datetime.date(year, 1, 1)).days
return int(valid.strftime("%j")) + days
df['startdoy'] = df[['year', 'min_issue']].apply(myfunc, axis=1)
df['enddoy'] = df[['year', 'max_issue']].apply(myfunc, axis=1)
df.set_index('year', inplace=True)
ends = df['enddoy'].values
starts = df['startdoy'].values
years = df.index.values
fig = plt.Figure()
ax = plt.axes([0.1, 0.1, 0.7, 0.8])
ax.barh(years - 0.4, (ends - starts), left=starts, fc='blue')
ax.axvline(np.average(starts[:-1]), lw=2, color='red')
ax.axvline(np.average(ends[:-1]), lw=2, color='red')
ax.set_xlabel(("Avg Start Date: %s, End Date: %s") %
((datetime.date(2000, 1, 1) + datetime.timedelta(
days=int(np.average(starts[:-1])))).strftime("%-d %b"),
(datetime.date(2000, 1, 1) + datetime.timedelta(
days=int(np.average(ends[:-1])))).strftime("%-d %b")),
color='red')
ax.set_title(("[%s] NWS %s\nPeriod between First and Last %s %s") %
(station, nt.sts[station]['name'], vtec._phenDict[phenomena],
vtec._sigDict[significance]))
ax.grid()
days = [1, 32, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335]
days = days + [x + 365 for x in days]
ax.set_xticks(days)
ax.set_xticklabels(calendar.month_abbr[1:] + calendar.month_abbr[1:])
ax.set_xlim(df['startdoy'].min() - 10, df['enddoy'].max() + 10)
ax.set_ylabel("Year")
ax.set_ylim(years[0] - 0.5, years[-1] + 0.5)
xFormatter = FormatStrFormatter('%d')
ax.yaxis.set_major_formatter(xFormatter)
ax = plt.axes([0.82, 0.1, 0.13, 0.8])
ax.barh(years - 0.4, df['count'], fc='blue')
ax.set_ylim(years[0] - 0.5, years[-1] + 0.5)
plt.setp(ax.get_yticklabels(), visible=False)
ax.grid(True)
ax.set_xlabel("# Events")
ax.yaxis.set_major_formatter(xFormatter)
xloc = plt.MaxNLocator(3)
ax.xaxis.set_major_locator(xloc)
return fig, df
def plot_ptrh(data, specs, outputpath):
'''
routine plots vertical profiles of temperature, pressure, rel humidity and
saves plot as .png
INPUT:
- data: dictionnary with data (eg filled by read_ncfile())
- specs: dictionnary with filename specifications
(filled by read_ncfile())
- outputpath: path where png will be stored in.
OUTPUT: .png file stored in outputpath
'''
logging.info(
'now plotting pressure, temperature, rel humidity sounding.........')
# define outputname of .png-file:
variable = 'ptrelh'
outputname = '{platform}_{instrument}{direction}_{variable}_{date}_{tempres}.png'.format(
platform=specs['platform_short'],
instrument=specs['type'].replace(' ', '').replace('_', ''),
direction=specs['direction'],
variable=variable,
date=specs['date'] + '_' + specs['time'],
tempres=specs['tempres'].replace(' ', ''))
fig, ax = plt.subplots(1, 3, sharey=True, figsize=(8, 6))
# plot temperature, pressure, humidity in three panels:
ax[0].plot(data['temperature'], data['altitude'], '.-k', markersize=1)
ax[1].plot(data['pressure'], data['altitude'], '.-k', markersize=1)
ax[2].plot(data['humidity'], data['altitude'], '.-k', markersize=1)
# do some cosmetics regarding the layout, axislabels, etc.:
for i in range(3):
# switch off some spines:
ax[i].spines['top'].set_visible(False)
ax[i].spines['right'].set_visible(False)
ax[i].spines['left'].set_visible(False)
ax[i].grid(axis='y', linestyle='-', color='gray')
# set height axis to start at 0m:
ax[i].set_ylim(0, ax[i].get_ylim()[-1])
# major minor ticks:
ax[i].xaxis.set_minor_locator(AutoMinorLocator())
ax[i].yaxis.set_minor_locator(AutoMinorLocator())
ax[i].xaxis.set_major_locator(plt.MaxNLocator(4))
# switch off major ticks top and right axis:
ax[i].tick_params(top=False, right=False)
# switch off minor ticks for top axis:
ax[i].tick_params(axis='x', which='minor', top=False)
# make labels larger for all ticks:
ax[i].tick_params(axis='both', labelsize=14)
ax[0].spines['left'].set_visible(True)
ax[0].tick_params(axis='y', right=False, which='minor')
ax[1].tick_params(left=False)
ax[1].tick_params(axis='y', right=False, left=False, which='minor')
ax[2].tick_params(left=False)
ax[2].spines['right'].set_visible(True)
ax[2].yaxis.set_ticks_position('right')
ax[2].yaxis.set_label_position('right')
# set the relh panel always to values between 0 and 100:
ax[2].set_xlim(0, 100)
# and the pressure to max 1100 hPa:
ax[1].set_xlim(ax[1].get_xlim()[0], 1100)
# axis labels:
ax[0].set_ylabel('Altitude [m]', fontsize=14)
ax[2].set_ylabel('Altitude [m]', fontsize=14)
ax[0].set_xlabel('Temperature [$^\circ$C]', fontsize=14)
ax[1].set_xlabel('Pressure [hPa]', fontsize=14)
ax[2].set_xlabel('Rel Humidity [%]', fontsize=14)
plt.subplots_adjust(top=0.9, right=0.85, left=0.15)
fig.suptitle(
'%s, %s %sUTC' %
(specs['location'], specs['date'], data['time_of_launch_HHmmss'][:-2]),
fontsize=18)
fig.savefig(outputpath + outputname)
logging.info('{} profiles saved at {}'.format(variable,
outputpath + outputname))
def generate_compare_plot(self, tran, ambig, min_read, max_read,
master_filepath_dict, lite, offset_dict,
ribocoverage, organism, normalize, short_code,
background_col, hili_start, hili_stop, comp_uag_col,
comp_uga_col, comp_uaa_col, title_size,
subheading_size, axis_label_size, marker_size,
cds_marker_size, cds_marker_colour, legend_size,
transcriptome):
labels = []
start_visible = []
line_collections = []
all_stops = ["TAG", "TAA", "TGA"]
returnstr = "Position,"
y_max = 50
if normalize == True:
y_max = 0
connection = sqlite3.connect(
'/home/DATA/www/tripsviz/tripsviz/trips.sqlite')
connection.text_factory = str
cursor = connection.cursor()
cursor.execute(
"SELECT owner FROM organisms WHERE organism_name = '{}' and transcriptome_list = '{}';"
.format(organism, transcriptome))
owner = (cursor.fetchone())[0]
if owner == 1:
if os.path.isfile("{0}/{1}/{2}/{2}.{3}.sqlite".format(
config.SCRIPT_LOC, config.ANNOTATION_DIR, organism,
transcriptome)):
transhelve = sqlite3.connect("{0}/{1}/{2}/{2}.{3}.sqlite".format(
config.SCRIPT_LOC, config.ANNOTATION_DIR, organism,
transcriptome))
else:
return_str = "Cannot find annotation file {}.{}.sqlite".format(
organism, transcriptome)
return {
'current': 400,
'total': 100,
'status': 'return_str',
'result': return_str
}
else:
transhelve = sqlite3.connect(
"{0}transcriptomes/{1}/{2}/{3}/{2}_{3}.sqlite".format(
config.UPLOADS_DIR, owner, organism, transcriptome))
cursor = transhelve.cursor()
cursor.execute(
"SELECT * from transcripts WHERE transcript = '{}'".format(tran))
result = cursor.fetchone()
traninfo = {
"transcript": result[0],
"gene": result[1],
"length": result[2],
"cds_start": result[3],
"cds_stop": result[4],
"seq": result[5],
"strand": result[6],
"stop_list": result[7].split(","),
"start_list": result[8].split(","),
"exon_junctions": result[9].split(","),
"tran_type": result[10],
"principal": result[11]
}
traninfo["stop_list"] = [int(x) for x in traninfo["stop_list"]]
traninfo["start_list"] = [int(x) for x in traninfo["start_list"]]
if str(traninfo["exon_junctions"][0]) != "":
traninfo["exon_junctions"] = [
int(x) for x in traninfo["exon_junctions"]
]
else:
traninfo["exon_junctions"] = []
transhelve.close()
gene = traninfo["gene"]
tranlen = traninfo["length"]
cds_start = traninfo["cds_start"]
cds_stop = traninfo["cds_stop"]
strand = traninfo["strand"]
if cds_start == 'NULL' or cds_start == None:
cds_start = 0
if cds_stop == 'NULL' or cds_stop == None:
cds_stop = 0
all_starts = traninfo["start_list"]
all_stops = {"TAG": [], "TAA": [], "TGA": []}
seq = traninfo["seq"].upper()
for i in range(0, len(seq)):
if seq[i:i + 3] in all_stops:
all_stops[seq[i:i + 3]].append(i + 1)
start_stop_dict = {
1: {
"starts": [0],
"stops": {
"TGA": [0],
"TAG": [0],
"TAA": [0]
}
},
2: {
"starts": [0],
"stops": {
"TGA": [0],
"TAG": [0],
"TAA": [0]
}
},
3: {
"starts": [0],
"stops": {
"TGA": [0],
"TAG": [0],
"TAA": [0]
}
}
}
for start in all_starts:
rem = ((start - 1) % 3) + 1
start_stop_dict[rem]["starts"].append(start - 1)
for stop in all_stops:
for stop_pos in all_stops[stop]:
rem = ((stop_pos - 1) % 3) + 1
start_stop_dict[rem]["stops"][stop].append(stop_pos - 1)
fig = plt.figure(figsize=(23, 12))
ax_main = plt.subplot2grid((30, 1), (0, 0), rowspan=22)
if normalize != True:
label = 'Read count'
else:
label = 'Normalized read count'
ax_main.set_ylabel(label, fontsize=axis_label_size, labelpad=30)
label = 'Position (nucleotides)'
ax_main.set_xlabel(label, fontsize=axis_label_size, labelpad=10)
#if normalize is true work out the factors for each colour
if normalize == True:
all_mapped_reads = []
for color in master_filepath_dict:
all_mapped_reads.append(
master_filepath_dict[color]["mapped_reads"])
min_reads = float(min(all_mapped_reads))
for color in master_filepath_dict:
factor = min_reads / float(
master_filepath_dict[color]["mapped_reads"])
master_filepath_dict[color]["factor"] = factor
# So items can be plotted alphabetically
unsorted_list = []
for color in master_filepath_dict:
input_list = [
color, master_filepath_dict[color]["file_names"],
master_filepath_dict[color]["file_descs"],
master_filepath_dict[color]["file_ids"],
master_filepath_dict[color]["filepaths"],
master_filepath_dict[color]["file_type"],
master_filepath_dict[color]["minread"],
master_filepath_dict[color]["maxread"]
]
if "factor" in master_filepath_dict[color]:
input_list.append(master_filepath_dict[color]["factor"])
unsorted_list.append(input_list)
sorted_list = sorted(unsorted_list, key=lambda x: x[1][0])
returndict = {}
for item in sorted_list:
# needed to make get_reads accept file_paths
file_paths = {"riboseq": {}}
for i in range(0, len(item[3])):
file_paths["riboseq"][item[3][i]] = item[4][i]
file_names = item[1][0]
file_descs = item[2]
if item[5] == "riboseq":
filename_reads, seqvar_dict = get_reads(ambig, item[6], item[7],
tran, file_paths, tranlen,
ribocoverage, organism,
False, False, "fiveprime",
"riboseq", 1)
else:
filename_reads, seqvar_dict = get_reads(ambig, item[6], item[7],
tran, file_paths, tranlen,
True, organism, False,
False, "fiveprime",
"riboseq", 1)
if normalize == False:
try:
max_val = max(filename_reads.values()) * 1.1
if max_val > y_max:
y_max = max_val
except Exception as e:
pass
labels.append(file_names)
start_visible.append(True)
plot_filename = ax_main.plot(filename_reads.keys(),
filename_reads.values(),
alpha=1,
label=labels,
zorder=1,
color=item[0],
linewidth=3)
line_collections.append(plot_filename)
returndict[file_names] = {}
for pos in filename_reads:
returndict[file_names][pos] = filename_reads[pos]
else:
normalized_reads = {}
for pos in filename_reads:
normalized_reads[pos] = filename_reads[pos] * item[8]
try:
max_val = max(normalized_reads.values()) * 1.1
if max_val > y_max:
y_max = max_val
except Exception as e:
pass
labels.append(file_names)
start_visible.append(True)
plot_filename = ax_main.plot(normalized_reads.keys(),
normalized_reads.values(),
alpha=1,
label=labels,
zorder=1,
color=item[0],
linewidth=3)
line_collections.append(plot_filename)
returndict[file_names] = {}
for pos in filename_reads:
returndict[file_names][pos] = normalized_reads[pos]
for plot_filename in returndict:
returnstr += "{},".format(plot_filename)
returnstr += "\n"
for i in range(0, tranlen):
returnstr += "{},".format(i)
for plot_filename in returndict:
returnstr += "{},".format(returndict[plot_filename][i])
returnstr += "\n"
ax_main.set_ylim(0, y_max)
# draw cds start
#plt.plot((cds_start,cds_start), (0, y_max), cds_marker_colour,linestyle = ':',linewidth=cds_marker_size)
# draw cds end
#plt.plot((cds_stop, cds_stop), (0, y_max), cds_marker_colour,linestyle = ':',linewidth=cds_marker_size)
cds_markers = ax_main.plot((cds_start, cds_start), (0, y_max * 0.97),
color=cds_marker_colour,
linestyle='solid',
linewidth=cds_marker_size)
ax_main.text(cds_start,
y_max * 0.97,
"CDS start",
fontsize=18,
color="black",
ha="center")
#ax_main.annotate('axes fraction',xy=(3, 1), xycoords='data',xytext=(0.8, 0.95), textcoords='axes fraction',arrowprops=dict(facecolor='black', shrink=0.05),horizontalalignment='right', verticalalignment='top')
#trans = blended_transform_factory(ax_main.transData, ax_main.transAxes)
#ax_main.annotate('CDS RELATIVE START',(100,100),transform=trans)
#tform = blended_transform_factory(ax_main.transData, ax_main.transAxes)
#r=10
#ax_main.text(cds_start, 0.9, "CDS START OR WHATEVER", fontsize='xx-large', color='r', transform=tform)
cds_markers += ax_main.plot((cds_stop + 1, cds_stop + 1),
(0, y_max * 0.97),
color=cds_marker_colour,
linestyle='solid',
linewidth=cds_marker_size)
ax_main.text(cds_stop,
y_max * 0.97,
"CDS stop",
fontsize=18,
color="black",
ha="center")
line_collections.append(cds_markers)
start_visible.append(True)
labels.append("CDS Markers")
ax_f1 = plt.subplot2grid((30, 1), (27, 0), rowspan=1, sharex=ax_main)
ax_f1.set_axis_bgcolor('lightgray')
ax_f2 = plt.subplot2grid((30, 1), (28, 0), rowspan=1, sharex=ax_main)
ax_f2.set_axis_bgcolor('lightgray')
ax_f6 = plt.subplot2grid((30, 1), (29, 0), rowspan=1, sharex=ax_main)
ax_f6.set_axis_bgcolor('lightgray')
ax_f6.set_xlabel('Transcript: {} Length: {} nt'.format(tran, tranlen),
fontsize=subheading_size,
labelpad=10)
for axis, frame in ((ax_f1, 1), (ax_f2, 2), (ax_f6, 3)):
color = color_dict['frames'][frame - 1]
axis.set_xlim(0, tranlen)
starts = [(item, 1) for item in start_stop_dict[frame]['starts']]
axis.broken_barh(starts, (0.5, 1),
color='white',
zorder=5,
linewidth=2)
stops = [(item, 1) for item in start_stop_dict[frame]['stops']]
uag_stops = [(item, 1)
for item in start_stop_dict[frame]['stops']['TAG']]
uaa_stops = [(item, 1)
for item in start_stop_dict[frame]['stops']['TAA']]
uga_stops = [(item, 1)
for item in start_stop_dict[frame]['stops']['TGA']]
axis.broken_barh(uag_stops, (0, 1),
color=comp_uag_col,
zorder=2,
linewidth=2)
axis.broken_barh(uaa_stops, (0, 1),
color=comp_uaa_col,
zorder=2,
linewidth=2)
axis.broken_barh(uga_stops, (0, 1),
color=comp_uga_col,
zorder=2,
linewidth=2)
axis.set_ylabel('{}'.format(frame),
rotation='horizontal',
labelpad=10,
verticalalignment='center')
axis.set_ylim(0, 1)
axis.tick_params(top=False,
left=False,
right=False,
bottom=False,
labeltop=False,
labelleft=False,
labelright=False,
labelbottom=False)
ax_f6.axes.get_yaxis().set_ticks([])
ax_f2.axes.get_yaxis().set_ticks([])
ax_f1.axes.get_yaxis().set_ticks([])
title_str = '{} ({})'.format(gene, short_code)
plt.title(title_str, fontsize=title_size, y=36)
if not (hili_start == 0 and hili_stop == 0):
hili_start = int(hili_start)
hili_stop = int(hili_stop)
hili = ax_main.fill_between([hili_start, hili_stop], [y_max, y_max],
zorder=0,
alpha=0.75,
color="#fffbaf")
labels.append("Highligter")
start_visible.append(True)
line_collections.append(hili)
leg_offset = (legend_size - 17) * 5
if leg_offset < 0:
leg_offset = 0
leg_offset += 230
ilp = plugins.InteractiveLegendPlugin(line_collections,
labels,
alpha_unsel=0,
alpha_sel=0.85,
xoffset=leg_offset,
yoffset=20,
start_visible=start_visible,
fontsize=legend_size)
plugins.connect(fig, ilp, plugins.TopToolbar(yoffset=100),
plugins.DownloadProfile(returnstr=returnstr),
plugins.DownloadPNG(returnstr=title_str))
ax_main.set_axis_bgcolor(background_col)
# This changes the size of the tick markers, works on both firefox and chrome.
ax_main.tick_params('both', labelsize=marker_size)
ax_main.xaxis.set_major_locator(plt.MaxNLocator(3))
ax_main.yaxis.set_major_locator(plt.MaxNLocator(3))
ax_main.grid(color="white", linewidth=20, linestyle="solid")
graph = "<div style='padding-left: 55px;padding-top: 22px;'> <a href='https://trips.ucc.ie/short/{0}' target='_blank' ><button class='button centerbutton' type='submit'><b>Direct link to this plot</b></button></a> </div>".format(
short_code)
tot_prog = 100
graph += mpld3.fig_to_html(fig)
return {
'current': 400,
'total': tot_prog,
'status': 'Complete',
'result': graph
}
def generate_plot(self, tran, ambig, min_read, max_read, lite, ribocoverage,
organism, readscore, noisered, primetype, minfiles, nucseq,
user_hili_starts, user_hili_stops, uga_diff, file_paths_dict,
short_code, color_readlen_dist, background_col, uga_col,
uag_col, uaa_col, advanced, seqhili, seq_rules, title_size,
subheading_size, axis_label_size, marker_size, transcriptome,
trips_uploads_location, cds_marker_size, cds_marker_colour,
legend_size, ribo_linewidth, secondary_readscore, pcr,
mismatches, hili_start, hili_stop):
#self.update_state(state='PROGRESS',meta={'current': 0, 'total': 100,'status': "Generate plot called"})
if lite == "n" and ribocoverage == True:
return_str = "Error: Cannot display Ribo-Seq Coverage when 'Line Graph' is turned off"
return {
'current': 100,
'total': 100,
'status': 'Complete',
'result': return_str
}
labels = [
"Frame 1 profiles", "Frame 2 profiles", "Frame 3 profiles", "RNA",
"Exon Junctions"
]
start_visible = [True, True, True, True, False]
if mismatches == True:
labels.append("Mismatches A")
labels.append("Mismatches T")
labels.append("Mismatches G")
labels.append("Mismatches C")
start_visible.append(False)
start_visible.append(False)
start_visible.append(False)
start_visible.append(False)
start_visible.append(True)
labels.append("CDS markers")
#This is a list of booleans that decide if the interactive legends boxes are filled in or not.Needs to be same length as labels
stop_codons = ["TAG", "TAA", "TGA"]
frame_orfs = {1: [], 2: [], 3: []}
connection = sqlite3.connect('{}/trips.sqlite'.format(config.SCRIPT_LOC))
connection.text_factory = str
cursor = connection.cursor()
cursor.execute(
"SELECT owner FROM organisms WHERE organism_name = '{}' and transcriptome_list = '{}';"
.format(organism, transcriptome))
owner = (cursor.fetchone())[0]
if owner == 1:
if os.path.isfile("{0}/{1}/{2}/{2}.{3}.sqlite".format(
config.SCRIPT_LOC, config.ANNOTATION_DIR, organism,
transcriptome)):
transhelve = sqlite3.connect("{0}/{1}/{2}/{2}.{3}.sqlite".format(
config.SCRIPT_LOC, config.ANNOTATION_DIR, organism,
transcriptome))
else:
return_str = "Cannot find annotation file {}.{}.sqlite".format(
organism, transcriptome)
return {
'current': 100,
'total': 100,
'status': 'Complete',
'result': return_str
}
else:
transhelve = sqlite3.connect(
"{0}transcriptomes/{1}/{2}/{3}/{2}_{3}.sqlite".format(
trips_uploads_location, owner, organism, transcriptome))
connection.close()
cursor = transhelve.cursor()
cursor.execute(
"SELECT * from transcripts WHERE transcript = '{}'".format(tran))
result = cursor.fetchone()
traninfo = {
"transcript": result[0],
"gene": result[1],
"length": result[2],
"cds_start": result[3],
"cds_stop": result[4],
"seq": result[5],
"strand": result[6],
"stop_list": result[7].split(","),
"start_list": result[8].split(","),
"exon_junctions": result[9].split(","),
"tran_type": result[10],
"principal": result[11]
}
try:
traninfo["stop_list"] = [int(x) for x in traninfo["stop_list"]]
except:
traninfo["stop_list"] = []
try:
traninfo["start_list"] = [int(x) for x in traninfo["start_list"]]
except:
traninfo["start_list"] = []
if str(traninfo["exon_junctions"][0]) != "":
traninfo["exon_junctions"] = [
int(x) for x in traninfo["exon_junctions"]
]
else:
traninfo["exon_junctions"] = []
all_cds_regions = []
# Check if the 'coding_regions' table exists
cursor.execute(
"SELECT name FROM sqlite_master WHERE type='table' AND name='coding_regions';"
)
result = cursor.fetchone()
if result != None:
cursor.execute(
"SELECT * from coding_regions WHERE transcript = '{}'".format(
tran))
result = cursor.fetchall()
for row in result:
all_cds_regions.append((row[1], row[2]))
transhelve.close()
gene = traninfo["gene"]
tranlen = traninfo["length"]
cds_start = traninfo["cds_start"]
cds_stop = traninfo["cds_stop"]
if cds_start == "NULL" or cds_start == None:
cds_start = 0
if cds_stop == "NULL" or cds_stop == None:
cds_stop = 0
all_starts = traninfo["start_list"]
all_stops = {"TAG": [], "TAA": [], "TGA": []}
exon_junctions = traninfo["exon_junctions"]
seq = traninfo["seq"].upper()
for i in range(0, len(seq)):
if seq[i:i + 3] in stop_codons:
all_stops[seq[i:i + 3]].append(i + 1)
# Error occurs if one of the frames is empty for any given start/stop, so we initialise with -5 as this won't be seen by user and will prevent the error
start_stop_dict = {
1: {
"starts": [-5],
"stops": {
"TGA": [-5],
"TAG": [-5],
"TAA": [-5]
}
},
2: {
"starts": [-5],
"stops": {
"TGA": [-5],
"TAG": [-5],
"TAA": [-5]
}
},
3: {
"starts": [-5],
"stops": {
"TGA": [-5],
"TAG": [-5],
"TAA": [-5]
}
}
}
for start in all_starts:
rem = ((start - 1) % 3) + 1
start_stop_dict[rem]["starts"].append(start)
for stop in all_stops:
for stop_pos in all_stops[stop]:
rem = ((stop_pos - 1) % 3) + 1
start_stop_dict[rem]["stops"][stop].append(stop_pos)
#find all open reading frames
for frame in [1, 2, 3]:
for start in start_stop_dict[frame]["starts"]:
best_stop_pos = 10000000
for stop in start_stop_dict[frame]["stops"]:
for stop_pos in start_stop_dict[frame]["stops"][stop]:
if stop_pos > start and stop_pos < best_stop_pos:
best_stop_pos = stop_pos
if best_stop_pos != 10000000:
frame_orfs[frame].append((start, best_stop_pos))
#self.update_state(state='PROGRESS',meta={'current': 100, 'total': 100,'status': "Fetching RNA-Seq Reads"})
all_rna_reads, rna_seqvar_dict = get_reads(ambig,
min_read,
max_read,
tran,
file_paths_dict,
tranlen,
True,
organism,
False,
noisered,
primetype,
"rnaseq",
readscore,
pcr,
get_mismatches=mismatches,
self_obj=self)
#self.update_state(state='PROGRESS',meta={'current': 100, 'total': 100,'status': "Fetching Ribo-Seq Reads"})
all_subcodon_reads, ribo_seqvar_dict = get_reads(ambig,
min_read,
max_read,
tran,
file_paths_dict,
tranlen,
ribocoverage,
organism,
True,
noisered,
primetype,
"riboseq",
readscore,
secondary_readscore,
pcr,
get_mismatches=mismatches,
self_obj=self)
seq_var_dict = merge_dicts(ribo_seqvar_dict, rna_seqvar_dict)
try:
rnamax = max(all_rna_reads.values())
except:
rnamax = 0
try:
subcodonmax = max(all_subcodon_reads.values())
except:
subcodonmax = 0
y_max = max(1, rnamax, subcodonmax) * 1.1
fig = plt.figure(figsize=(22, 13))
ax_main = plt.subplot2grid((30, 1), (0, 0), rowspan=22)
ax_main.spines['bottom'].set_visible(False)
for s in ['bottom', 'left', 'top', 'right']:
ax_main.spines[s].set_linewidth(15)
ax_main.spines[s].set_color("red")
alt_seq_type_vars = []
# Plot any alternative sequence types if there are any
#self.update_state(state='PROGRESS',meta={'current': 100, 'total': 100,'status': "Fetching alternative Seq Types"})
for seq_type in file_paths_dict:
if seq_type != "riboseq" and seq_type != "rnaseq":
if file_paths_dict[seq_type] == {}:
continue
if seq_rules[seq_type]["frame_breakdown"] == 1:
frame_breakdown = True
else:
frame_breakdown = False
alt_sequence_reads, empty_seqvar_dict = get_reads(ambig,
min_read,
max_read,
tran,
file_paths_dict,
tranlen,
True,
organism,
frame_breakdown,
noisered,
primetype,
seq_type,
readscore,
self_obj=self)
if frame_breakdown == False:
alt_seq_plot = ax_main.plot(alt_sequence_reads.keys(),
alt_sequence_reads.values(),
alpha=1,
label=seq_type,
zorder=2,
color='#5c5c5c',
linewidth=2)
labels.append(seq_type)
start_visible.append(True)
alt_seq_type_vars.append(alt_seq_plot)
else:
alt_frame_counts = {
0: collections.OrderedDict(),
1: collections.OrderedDict(),
2: collections.OrderedDict()
}
for key in alt_sequence_reads:
start = key
rem = start % 3
if rem == 1: # frame 1
frame = 2
elif rem == 2: # frame 2
frame = 0
elif rem == 0: # frame 3
frame = 1
alt_frame_counts[frame][key] = alt_sequence_reads[key]
frame0_altseqplot = ax_main.plot(alt_frame_counts[0].keys(),
alt_frame_counts[0].values(),
alpha=0.75,
label=seq_type + "frame0",
zorder=2,
color="#FF4A45",
linewidth=2)
frame1_altseqplot = ax_main.plot(alt_frame_counts[1].keys(),
alt_frame_counts[1].values(),
alpha=0.75,
label=seq_type + "frame1",
zorder=2,
color="#64FC44",
linewidth=2)
frame2_altseqplot = ax_main.plot(alt_frame_counts[2].keys(),
alt_frame_counts[2].values(),
alpha=0.75,
label=seq_type + "frame2*",
zorder=2,
color="#5687F9",
linewidth=2)
labels.append(seq_type + "frame 1")
labels.append(seq_type + "frame 2")
labels.append(seq_type + "frame 3")
start_visible.append(True)
start_visible.append(True)
start_visible.append(True)
alt_seq_type_vars.append(frame0_altseqplot)
alt_seq_type_vars.append(frame1_altseqplot)
alt_seq_type_vars.append(frame2_altseqplot)
if max(alt_sequence_reads.values()) > y_max:
y_max = max(alt_sequence_reads.values())
label = 'Reads'
ax_main.set_ylabel(label, fontsize=axis_label_size, labelpad=30)
label = 'Position (nucleotides)'
ax_main.set_xlabel(label, fontsize=axis_label_size, labelpad=5)
ax_main.set_ylim(0, y_max)
if lite == "n":
rna_bars = ax_main.bar(all_rna_reads.keys(),
all_rna_reads.values(),
alpha=1,
label=labels,
zorder=1,
color='lightgray',
linewidth=0,
width=1)
else:
rna_bars = ax_main.plot(all_rna_reads.keys(),
all_rna_reads.values(),
alpha=1,
label=labels,
zorder=1,
color='#a7adb7',
linewidth=4)
cds_markers = ax_main.plot((cds_start, cds_start), (0, y_max * 0.97),
color=cds_marker_colour,
linestyle='solid',
linewidth=cds_marker_size)
ax_main.text(cds_start,
y_max * 0.97,
"CDS start",
fontsize=18,
color="black",
ha="center")
#ax_main.annotate('axes fraction',xy=(3, 1), xycoords='data',xytext=(0.8, 0.95), textcoords='axes fraction',arrowprops=dict(facecolor='black', shrink=0.05),horizontalalignment='right', verticalalignment='top')
#trans = blended_transform_factory(ax_main.transData, ax_main.transAxes)
#ax_main.annotate('CDS RELATIVE START',(100,100),transform=trans)
#tform = blended_transform_factory(ax_main.transData, ax_main.transAxes)
#r=10
#ax_main.text(cds_start, 0.9, "CDS START OR WHATEVER", fontsize='xx-large', color='r', transform=tform)
cds_markers += ax_main.plot((cds_stop + 1, cds_stop + 1),
(0, y_max * 0.97),
color=cds_marker_colour,
linestyle='solid',
linewidth=cds_marker_size)
ax_main.text(cds_stop,
y_max * 0.97,
"CDS stop",
fontsize=18,
color="black",
ha="center")
ax_cds = plt.subplot2grid((31, 1), (26, 0), rowspan=1, sharex=ax_main)
ax_cds.set_axis_bgcolor("white")
ax_cds.set_ylabel('Merged CDS',
labelpad=4,
verticalalignment='center',
horizontalalignment="right",
rotation="horizontal",
color="black",
fontsize=(axis_label_size / 1.5))
ax_f1 = plt.subplot2grid((31, 1), (27, 0), rowspan=1, sharex=ax_main)
ax_f1.set_axis_bgcolor(color_dict['frames'][0])
ax_f2 = plt.subplot2grid((31, 1), (28, 0), rowspan=1, sharex=ax_main)
ax_f2.set_axis_bgcolor(color_dict['frames'][1])
ax_f3 = plt.subplot2grid((31, 1), (29, 0), rowspan=1, sharex=ax_main)
ax_f3.set_axis_bgcolor(color_dict['frames'][2])
ax_nucseq = plt.subplot2grid((31, 1), (30, 0), rowspan=1, sharex=ax_main)
ax_nucseq.set_xlabel('Transcript: {} Length: {} nt'.format(tran, tranlen),
fontsize=subheading_size)
for tup in all_cds_regions:
ax_cds.fill_between([tup[0], tup[1]], [1, 1],
zorder=0,
alpha=1,
color="#001285")
#plot a dummy exon junction at postion -1, needed in cases there are no exon junctions, this wont be seen
allexons = ax_main.plot((-1, -1), (0, 1),
alpha=0.01,
color='black',
linestyle='-.',
linewidth=2)
for exon in exon_junctions:
allexons += ax_main.plot((exon, exon), (0, y_max),
alpha=0.01,
color='black',
linestyle='-.',
linewidth=3)
#dictionary for each frame in which the keys are the posistions and the values are the counts
frame_counts = {
0: collections.OrderedDict(),
1: collections.OrderedDict(),
2: collections.OrderedDict()
}
for key in all_subcodon_reads:
rem = key % 3
if rem == 1: # frame 1
frame = 0
elif rem == 2: # frame 2
frame = 1
elif rem == 0: # frame 3
frame = 2
frame_counts[frame][key] = all_subcodon_reads[key]
if lite == "n":
frame_counts[frame][key + 1] = 0
frame_counts[frame][key + 2] = 0
if lite == "n":
frame0subpro = ax_main.bar(frame_counts[0].keys(),
frame_counts[0].values(),
alpha=0.75,
label=labels,
zorder=2,
color="#FF4A45",
edgecolor="#FF4A45",
width=1,
linewidth=4)
frame1subpro = ax_main.bar(frame_counts[1].keys(),
frame_counts[1].values(),
alpha=0.75,
label=labels,
zorder=2,
color="#64FC44",
edgecolor="#64FC44",
width=1,
linewidth=4)
frame2subpro = ax_main.bar(frame_counts[2].keys(),
frame_counts[2].values(),
alpha=0.75,
label=labels,
zorder=2,
color="#5687F9",
edgecolor="#5687F9",
width=1,
linewidth=4)
else:
frame0subpro = ax_main.plot(frame_counts[0].keys(),
frame_counts[0].values(),
alpha=0.75,
label=labels,
zorder=2,
color="#FF4A45",
linewidth=ribo_linewidth)
frame1subpro = ax_main.plot(frame_counts[1].keys(),
frame_counts[1].values(),
alpha=0.75,
label=labels,
zorder=2,
color="#64FC44",
linewidth=ribo_linewidth)
frame2subpro = ax_main.plot(frame_counts[2].keys(),
frame_counts[2].values(),
alpha=0.75,
label=labels,
zorder=2,
color="#5687F9",
linewidth=ribo_linewidth)
if mismatches == True:
a_mismatches = ax_main.plot(seq_var_dict["A"].keys(),
seq_var_dict["A"].values(),
alpha=0.01,
label=labels,
zorder=2,
color="purple",
linewidth=2)
t_mismatches = ax_main.plot(seq_var_dict["T"].keys(),
seq_var_dict["T"].values(),
alpha=0.01,
label=labels,
zorder=2,
color="yellow",
linewidth=2)
g_mismatches = ax_main.plot(seq_var_dict["G"].keys(),
seq_var_dict["G"].values(),
alpha=0.01,
label=labels,
zorder=2,
color="orange",
linewidth=2)
c_mismatches = ax_main.plot(seq_var_dict["C"].keys(),
seq_var_dict["C"].values(),
alpha=0.01,
label=labels,
zorder=2,
color="pink",
linewidth=2)
xy = 0
if nucseq == True:
ax_nucseq.set_axis_bgcolor(background_col)
mrnaseq = seq.replace("T", "U")
color_list = ["#FF4A45", "#64FC44", "#5687F9"]
char_frame = 0
for char in mrnaseq:
ax_nucseq.text((xy + 1) - 0.1,
0.2,
mrnaseq[xy],
fontsize=20,
color=color_list[char_frame % 3])
xy += 1
char_frame += 1
# If the user passed a list of sequences to highlight, find and plot them here.
if seqhili != ['']:
near_cog_starts, signalhtml = get_user_defined_seqs(seq, seqhili)
for slip in near_cog_starts[0]:
try:
hili_sequences += ax_f1.plot((slip, slip), (0, 0.5),
alpha=1,
label=labels,
zorder=4,
color='black',
linewidth=5)
except Exception as e:
hili_sequences = ax_f1.plot((slip, slip), (0, 0.5),
alpha=1,
label=labels,
zorder=4,
color='black',
linewidth=5)
for slip in near_cog_starts[1]:
try:
hili_sequences += ax_f2.plot((slip, slip), (0, 0.5),
alpha=1,
label=labels,
zorder=4,
color='black',
linewidth=5)
except:
hili_sequences = ax_f2.plot((slip, slip), (0, 0.5),
alpha=1,
label=labels,
zorder=4,
color='black',
linewidth=5)
for slip in near_cog_starts[2]:
try:
hili_sequences += ax_f3.plot((slip, slip), (0, 0.5),
alpha=1,
label=labels,
zorder=4,
color='black',
linewidth=5)
except:
hili_sequences = ax_f3.plot((slip, slip), (0, 0.5),
alpha=1,
label=labels,
zorder=4,
color='black',
linewidth=5)
#Plot sequence identifiers which will create a popup telling user what the subsequence is (useful if they have passed multiple subsequences)
frame1_subsequences = ax_f1.plot(near_cog_starts[0],
[0.25] * len(near_cog_starts[0]),
'o',
color='b',
mec='k',
ms=12,
mew=1,
alpha=0,
zorder=4)
frame2_subsequences = ax_f2.plot(near_cog_starts[1],
[0.25] * len(near_cog_starts[1]),
'o',
color='b',
mec='k',
ms=12,
mew=1,
alpha=0,
zorder=4)
frame3_subsequences = ax_f3.plot(near_cog_starts[2],
[0.25] * len(near_cog_starts[2]),
'o',
color='b',
mec='k',
ms=12,
mew=1,
alpha=0,
zorder=4)
#Attach the labels to the subsequences plotted above
signaltooltip1 = PointHTMLTooltip(frame1_subsequences[0],
signalhtml[0],
voffset=10,
hoffset=10,
css=point_tooltip_css)
signaltooltip2 = PointHTMLTooltip(frame2_subsequences[0],
signalhtml[1],
voffset=10,
hoffset=10,
css=point_tooltip_css)
signaltooltip3 = PointHTMLTooltip(frame3_subsequences[0],
signalhtml[2],
voffset=10,
hoffset=10,
css=point_tooltip_css)
for axisname in (ax_f1, ax_f2, ax_f3, ax_nucseq, ax_cds):
axisname.tick_params(top=False,
bottom=False,
labelleft=False,
labelright=False,
labelbottom=False)
for label in ax_main.xaxis.get_majorticklabels():
label.set_fontsize(36)
for axis, frame in ((ax_f1, 1), (ax_f2, 2), (ax_f3, 3)):
axis.set_xlim(1, tranlen)
starts = [(item, 1) for item in start_stop_dict[frame]['starts']]
uag_stops = [(item, 1)
for item in start_stop_dict[frame]['stops']['TAG']]
uaa_stops = [(item, 1)
for item in start_stop_dict[frame]['stops']['TAA']]
uga_stops = [(item, 1)
for item in start_stop_dict[frame]['stops']['TGA']]
#Plot start positions
axis.broken_barh(starts, (0.30, 1),
color="white",
zorder=2,
linewidth=7)
#Plot stop positions
axis.broken_barh(uag_stops, (0, 1),
color=uag_col,
zorder=2,
linewidth=4)
axis.broken_barh(uaa_stops, (0, 1),
color=uaa_col,
zorder=2,
linewidth=4)
axis.broken_barh(uga_stops, (0, 1),
color=uga_col,
zorder=2,
linewidth=4)
axis.set_ylim(0, 1)
axis.set_ylabel('Frame {}'.format(frame),
labelpad=4,
verticalalignment='center',
horizontalalignment="right",
rotation="horizontal",
color="black",
fontsize=(axis_label_size / 1.5))
title_str = '{} ({})'.format(gene, short_code)
plt.title(title_str, fontsize=title_size, y=38)
line_collections = [
frame0subpro, frame1subpro, frame2subpro, rna_bars, allexons
]
if mismatches == True:
line_collections.append(a_mismatches)
line_collections.append(t_mismatches)
line_collections.append(g_mismatches)
line_collections.append(c_mismatches)
line_collections.append(cds_markers)
if not (hili_start == 0 and hili_stop == 0):
hili_start = int(hili_start)
hili_stop = int(hili_stop)
hili = ax_main.fill_between([hili_start, hili_stop], [y_max, y_max],
zorder=0,
alpha=0.75,
color="#fffbaf")
labels.append("Highligted region")
start_visible.append(True)
line_collections.append(hili)
for alt_plot in alt_seq_type_vars:
line_collections.append(alt_plot)
if 'hili_sequences' in locals():
labels.append("Highligted sequences")
start_visible.append(True)
line_collections.append(hili_sequences)
if user_hili_starts != [] and user_hili_stops != []:
for i in range(0, len(user_hili_starts)):
user_hili_start = int(user_hili_starts[i])
user_hili_stop = int(user_hili_stops[i])
try:
hili += ax_main.fill_between([user_hili_start, user_hili_stop],
[y_max, y_max],
alpha=0.75,
color="#fffbaf")
except:
hili = ax_main.fill_between([user_hili_start, user_hili_stop],
[y_max, y_max],
alpha=0.75,
color="#fffbaf")
labels.append("Highligter")
start_visible.append(True)
line_collections.append(hili)
leg_offset = (legend_size - 17) * 5
if leg_offset < 0:
leg_offset = 0
ilp = InteractiveLegendPlugin(line_collections,
labels,
alpha_unsel=0,
alpha_sel=0.85,
start_visible=start_visible,
fontsize=legend_size,
xoffset=leg_offset)
htmllabels = {1: [], 2: [], 3: []}
all_start_points = {1: [], 2: [], 3: []}
try:
con_scores = SqliteDict(
"{0}/{1}/homo_sapiens/score_dict.sqlite".format(
config.SCRIPT_LOC, config.ANNOTATION_DIR))
except Exception as e:
con_scores = []
for frame in [1, 2, 3]:
orf_list = frame_orfs[frame]
for tup in orf_list:
orf_ribo = 0.0
outframe_ribo = 0.0
orf_rna = 0.0001
start = tup[0]
try:
context = (seq[start - 7:start + 4].upper()).replace("T", "U")
except Exception as e:
con_score = "?"
if len(context) != 11 or context[6:9] != "AUG":
con_score = "?"
else:
try:
con_score = con_scores[context.upper()]
except Exception as e:
con_score = "?"
all_start_points[frame].append(start - 1)
stop = tup[1]
other_ribo = 0.0
otherother_ribo = 0.0
for i in range(start + 2, stop, 3):
for subframe in [0, 1, 2]:
if i in frame_counts[subframe]:
orf_ribo += frame_counts[subframe][i]
for i in range(start, stop, 3):
for subframe in [0, 1, 2]:
if i in frame_counts[subframe]:
outframe_ribo += frame_counts[subframe][i]
for i in range(start + 1, stop, 3):
for subframe in [0, 1, 2]:
if i in frame_counts[subframe]:
outframe_ribo += frame_counts[subframe][i]
for i in range(start, stop + 1):
if i in all_rna_reads:
orf_rna += all_rna_reads[i]
orf_te = float(orf_ribo) / float(orf_rna)
orf_len = int(stop - start)
try:
in_out_ratio = orf_ribo / outframe_ribo
except:
in_out_ratio = "Null"
datadict = {
'inframe ribo': [orf_ribo],
'outframe ribo': [outframe_ribo],
'in/out ratio': [in_out_ratio],
'rna': [orf_rna],
'te': [orf_te],
'len': [orf_len],
'context_score': [str(con_score) + "/150"]
}
df = pd.DataFrame(datadict,
columns=([
"inframe ribo", "outframe ribo",
"in/out ratio", "rna", "te", "len",
"context_score"
]))
label = df.ix[[0], :].T
label.columns = ["Start pos: {}".format(start - 1)]
htmllabels[frame].append(str(label.to_html()))
points1 = ax_f1.plot(all_start_points[1],
[0.75] * len(all_start_points[1]),
'o',
color='b',
mec='k',
ms=13,
mew=1,
alpha=0,
zorder=3)
points2 = ax_f2.plot(all_start_points[2],
[0.75] * len(all_start_points[2]),
'o',
color='b',
mec='k',
ms=13,
mew=1,
alpha=0,
zorder=3)
points3 = ax_f3.plot(all_start_points[3],
[0.75] * len(all_start_points[3]),
'o',
color='b',
mec='k',
ms=13,
mew=1,
alpha=0,
zorder=3)
tooltip1 = PointHTMLTooltip(points1[0],
htmllabels[1],
voffset=10,
hoffset=10,
css=point_tooltip_css)
tooltip2 = PointHTMLTooltip(points2[0],
htmllabels[2],
voffset=10,
hoffset=10,
css=point_tooltip_css)
tooltip3 = PointHTMLTooltip(points3[0],
htmllabels[3],
voffset=10,
hoffset=10,
css=point_tooltip_css)
ax_f3.axes.get_yaxis().set_ticks([])
ax_f2.axes.get_yaxis().set_ticks([])
ax_f1.axes.get_yaxis().set_ticks([])
returnstr = "Position,Sequence,Frame 1,Frame 2,Frame 3,RNA-Seq\n"
for i in range(0, len(seq)):
f1_count = 0
f2_count = 0
f3_count = 0
rna_count = 0
if i + 1 in frame_counts[0]:
f1_count = frame_counts[0][i + 1]
elif i + 1 in frame_counts[1]:
f2_count = frame_counts[1][i + 1]
elif i + 1 in frame_counts[2]:
f3_count = frame_counts[2][i + 1]
if i + 1 in all_rna_reads:
rna_count = all_rna_reads[i + 1]
returnstr += "{},{},{},{},{},{}\n".format(i + 1, seq[i], f1_count,
f2_count, f3_count,
rna_count)
if seqhili == ['']:
plugins.connect(fig, ilp, tooltip1, tooltip2, tooltip3,
TopToolbar(yoffset=180, xoffset=-70),
DownloadProfile(returnstr=returnstr),
DownloadPNG(returnstr=title_str))
else:
plugins.connect(fig, ilp, tooltip1, tooltip2, tooltip3,
signaltooltip1, signaltooltip2, signaltooltip3,
TopToolbar(yoffset=100),
DownloadProfile(returnstr=returnstr),
DownloadPNG(returnstr=title_str))
ax_main.set_axis_bgcolor(background_col)
# This changes the size of the tick markers, works on both firefox and chrome.
ax_main.tick_params('both', labelsize=marker_size)
ax_main.xaxis.set_major_locator(plt.MaxNLocator(3))
ax_main.yaxis.set_major_locator(plt.MaxNLocator(3))
ax_main.grid(True, color="white", linewidth=30, linestyle="solid")
#Without this style tag the markers sizes will appear correct on browser but be original size when downloaded via png
graph = "<style>.mpld3-xaxis {{font-size: {0}px;}} .mpld3-yaxis {{font-size: {0}px;}}</style>".format(
marker_size)
graph += "<div style='padding-left: 55px;padding-top: 22px;'> <a href='https://trips.ucc.ie/short/{0}' target='_blank' ><button class='button centerbutton' type='submit'><b>Direct link to this plot</b></button></a> </div>".format(
short_code)
graph += mpld3.fig_to_html(fig)
tot_prog = 100
return {
'current': 400,
'total': tot_prog,
'status': 'Complete',
'result': graph
}
ax[0][n].set_title(syl_types[n])
fu, tu, Sxx = signal.spectrogram(temp,
samp,
nperseg=NN,
noverlap=NN * overlap,
window=signal.get_window(
('gaussian', sigma), NN),
scaling='spectrum')
Sxx = np.clip(Sxx, a_min=np.amax(Sxx) * 0.000001, a_max=np.amax(Sxx))
ax[1][n].pcolormesh(tu,
fu,
np.log(Sxx),
cmap=plt.get_cmap('Greys'),
rasterized=True)
ax[1][n].set_ylim(0, 8000)
ax[1][n].yaxis.set_major_locator(plt.MaxNLocator(2))
ax[1][n].xaxis.set_major_locator(plt.MaxNLocator(2))
fig.tight_layout()
# Ahora 0 va a ser silencio
# syl_types.append('s')
num_syl = len(syl_types)
glyphs_to = [x for x in syl_types]
orden = 1
combinations = list(it.product(glyphs_to, repeat=orden))
glyphs_from = [''.join(list(x)) for x in combinations]
num_glyphs_to = len(glyphs_to)
num_glyphs_from = len(glyphs_from)
# ------------------#
# Transition matrix, numbers
n_tr_matrix = np.zeros((num_files, num_glyphs_from, num_glyphs_to))
n_tr_matrix_norep = np.zeros((num_files, num_glyphs_from, num_glyphs_to))
list_files = np.array_split(runs, 5)
list_dist = np.array_split(dist, 5)
fig, ax = plt.subplots(2,
3,
subplot_kw={'projection': 'polar'},
figsize=(60, 60))
for i in range(len(ax)):
for r, j, d in zip(list_files[i], range(3), list_dist[i]):
df, _ = access_file(direc + r)
angle, fit = angle_theta_binning(df, col)
#ax[i, j].scatter(angle['theta'], angle['signal'], s=0.01)
ax[i, j].plot(fit['theta'], fit['signal'], color='red')
ax[i, j].set_rlabel_position(90)
ax[i, j].set_theta_zero_location("N") # theta=0 at the top
ax[i, j].set_theta_direction(-1)
ax[i, j].yaxis.set_major_locator(plt.MaxNLocator(3))
if col == 'Couple':
ax[i, j].set_rticks([7.2, 7.4, 7.6]) # Less radial ticks
elif col == 'Force':
ax[i, j].set_rticks([169, 172, 176]) # Less radial ticks
ax[i, j].tick_params(axis='y', which='major', labelsize=9)
ax[i, j].set_title('d = {:.2f}'.format(float(d) / 724))
plt.tight_layout()
plt.savefig(out_direc + col + '_1.pdf')
fig, ax = plt.subplots(2,
3,
subplot_kw={'projection': 'polar'},
figsize=(60, 60))
counts = Counter(match)
print counts
x_axis = []
y_axis = []
for eachPixcel in counts:
#print eachPixcel
x_axis.append(eachPixcel)
y_axis.append(counts[eachPixcel])
#print counts[eachPixcel]
fig = plt.figure()
axis1 = fig.add_subplot(211)
axis2 = fig.add_subplot(212)
#axis1.legend("Given Image")
#axis2.legend("OCR Prediction")
#plt.ylim(400)
x_points = plt.MaxNLocator(12)
axis2.xaxis.set_major_locator(x_points)
axis1.imshow(test_iar, label="Given Image")
axis2.bar(x_axis, y_axis, label="OCR Prediction")
axis2.set_ylim(350)
axis1.legend()
axis2.legend()
plt.show()
#a =one[1].split("],")
#print one[0]
#print len(eachPixcel)
# Ingredient 3 - Define your label axis
# Ingrediente 3 - Defina os rotulos dos eixos
plt.xlabel("Amigos")
plt.ylabel("ECDF")
plt.margins(0.02)
# Ingredient 5 - Define your pattern from x(or y) axis
# Ingrediente 5 - Defina o padrao dos numeros eixo x(ou y)
def major_formatter(a, pos):
return "%.f" % (a)
ax = plt.axes()
# Ingredient 6 - Define how many elements show in y axis
# Ingrediente 6 - Defina Quantos elementos aparecerao no eixo y
ax.yaxis.set_major_locator(plt.MaxNLocator(9))
# Ingredient 7 - Define how many elements show in x axis
# Ingrediente 7 - Defina Quantos elementos aparecerao no eixo x
ax.xaxis.set_major_locator(plt.MaxNLocator(11))
# Ingredient 8 - Add the Ingredient 5 in x axis
# Ingrediente 8 - Adiciona o Ingrediente 5 no eixo x
ax.xaxis.set_major_formatter(plt.FuncFormatter(major_formatter))
plt.legend()
####################################################
# Receita para dar um Zoom no recheio do Bolo ##
# Recipe to make zoom inside the cake ##
####################################################
def plotErrorRate(res):
halpha = 6562.80
hbeta = 4861.325
o2 = 3728.5
lya = 1215.67
mg2 = 2798.75
c3 = 1908.73
o3d = 5006.91
c4 = 1549.06
qop4 = res[res[:, 0] >= 3]
thresh = 0.01
autozNum = 1.0 * (np.abs(qop4[:, 3] - qop4[:, 1]) < thresh).sum()
marzNum = 1.0 * (np.abs(qop4[:, 2] - qop4[:, 1]) < thresh).sum()
runzxNum = 1.0 * (np.abs(qop4[:, 4] - qop4[:, 1]) < thresh).sum()
runzeNum = 1.0 * (np.abs(qop4[:, 5] - qop4[:, 1]) < thresh).sum()
autoz = ((1 + qop4[:, 3]) /
(1 + qop4[:, 1]))[np.abs(qop4[:, 3] - qop4[:, 1]) > thresh]
marz = ((1 + qop4[:, 2]) /
(1 + qop4[:, 1]))[np.abs(qop4[:, 2] - qop4[:, 1]) > thresh]
runzx = ((1 + qop4[:, 4]) /
(1 + qop4[:, 1]))[np.abs(qop4[:, 4] - qop4[:, 1]) > thresh]
runze = ((1 + qop4[:, 5]) /
(1 + qop4[:, 1]))[np.abs(qop4[:, 5] - qop4[:, 1]) > thresh]
msr = (100.0 * (1.0 - (1.0 * marz.size / qop4[:, 0].size)))
asr = (100.0 * (1.0 - (1.0 * autoz.size / qop4[:, 0].size)))
rxsr = (100.0 * (1.0 - (1.0 * runzx.size / qop4[:, 0].size)))
resr = (100.0 * (1.0 - (1.0 * runze.size / qop4[:, 0].size)))
print("marz %0.3f%% success rate, %0.3f%% fail rate" % (msr, 100 - msr))
print("autoz %0.3f%% success rate, %0.3f%% fail rate" % (asr, 100 - asr))
print("runzx %0.3f%% success rate, %0.3f%% fail rate" % (rxsr, 100 - rxsr))
print("runze %0.3f%% success rate, %0.3f%% fail rate" % (resr, 100 - resr))
fig = plt.figure(figsize=(7, 7), dpi=300)
matplotlib.rcParams.update({'font.size': 12})
#matplotlib.rcParams['axes.labelsize'] = 20
rc('text', usetex=False)
matplotlib.rcParams['xtick.labelsize'] = 12
matplotlib.rcParams['ytick.labelsize'] = 12
gs = gridspec.GridSpec(4, 1)
gs.update(wspace=0.0, hspace=0.0)
ax0 = fig.add_subplot(gs[0])
ax1 = fig.add_subplot(gs[1], sharex=ax0)
ax2 = fig.add_subplot(gs[2], sharex=ax0)
ax3 = fig.add_subplot(gs[3], sharex=ax0)
bins = np.linspace(0.2, 1.5, 200)
width = 1 * (bins[1] - bins[0])
center = (bins[:-1] + bins[1:]) / 2
hist0 = 100 * np.histogram(runzx, bins=bins)[0] / autozNum
hist1 = 100 * np.histogram(runze, bins=bins)[0] / marzNum
hist2 = 100 * np.histogram(autoz, bins=bins)[0] / runzxNum
hist3 = 100 * np.histogram(marz, bins=bins)[0] / runzeNum
maxes = max([max(hist0), max(hist1), max(hist2), max(hist3)])
ax0.plot([halpha / o3d, halpha / o3d], [1.1 * maxes, 0],
':',
color='#555555')
ax1.plot([halpha / o3d, halpha / o3d], [1.1 * maxes, 0],
':',
color='#555555')
ax2.plot([halpha / o3d, halpha / o3d], [1.1 * maxes, 0],
':',
color='#555555')
ax3.plot([halpha / o3d, halpha / o3d], [1.1 * maxes, 0],
':',
color='#555555')
ax2.text(1.23, 2, r'$\mathrm{H}\alpha\mathrm{/O[III]}$', fontsize=14)
ax0.plot([o2 / o3d, o2 / o3d], [1.1 * maxes, 0], ':', color='#555555')
ax1.plot([o2 / o3d, o2 / o3d], [1.1 * maxes, 0], ':', color='#555555')
ax2.plot([o2 / o3d, o2 / o3d], [1.1 * maxes, 0], ':', color='#555555')
ax3.plot([o2 / o3d, o2 / o3d], [1.1 * maxes, 0], ':', color='#555555')
ax0.text(0.63, 0.92, r'$\mathrm{O[II]/O[III]}$', fontsize=14)
ax0.plot([mg2 / halpha, mg2 / halpha], [1.1 * maxes, 0],
':',
color='#555555')
ax1.plot([mg2 / halpha, mg2 / halpha], [1.1 * maxes, 0],
':',
color='#555555')
ax2.plot([mg2 / halpha, mg2 / halpha], [1.1 * maxes, 0],
':',
color='#555555')
ax3.plot([mg2 / halpha, mg2 / halpha], [1.1 * maxes, 0],
':',
color='#555555')
ax0.text(0.31, 0.92, r'$\mathrm{MgII/H}\alpha$', fontsize=14)
ax0.plot([o2 / halpha, o2 / halpha], [1.1 * maxes, 0],
':',
color='#555555')
ax1.plot([o2 / halpha, o2 / halpha], [1.1 * maxes, 0],
':',
color='#555555')
ax2.plot([o2 / halpha, o2 / halpha], [1.1 * maxes, 0],
':',
color='#555555')
ax3.plot([o2 / halpha, o2 / halpha], [1.1 * maxes, 0],
':',
color='#555555')
ax3.text(0.455, 1.5, r'$\mathrm{O[II]/H}\alpha$', fontsize=14)
a0 = ax0.bar(center,
hist0,
align='center',
width=width,
edgecolor="none",
facecolor="#E53935",
label="Runz xcor")
a1 = ax1.bar(center,
hist1,
align='center',
width=width,
edgecolor="none",
facecolor="#AB47BC",
label="Runz ELM")
a2 = ax2.bar(center,
hist2,
align='center',
width=width,
edgecolor="none",
facecolor="#4CAF50",
label="Autoz")
a3 = ax3.bar(center,
hist3,
align='center',
width=width,
edgecolor="none",
facecolor="#2196F3",
label="Marz")
'''
ax0.plot(center, hist0, linewidth=2, color="#E53935")
ax1.plot(center, hist1, linewidth=2, color="#AB47BC")
ax2.plot(center, hist2, linewidth=2, color="#4CAF50")
ax3.plot(center, hist3, linewidth=2, color="#2196F3")
'''
ax0.yaxis.set_major_locator(plt.MaxNLocator(4))
ax1.yaxis.set_major_locator(plt.MaxNLocator(4))
ax2.yaxis.set_major_locator(plt.MaxNLocator(4))
ax3.yaxis.set_major_locator(plt.MaxNLocator(4))
ax0.yaxis.get_major_ticks()[0].label1.set_visible(False)
ax1.yaxis.get_major_ticks()[0].label1.set_visible(False)
ax2.yaxis.get_major_ticks()[0].label1.set_visible(False)
ax3.set_xlabel(r"$(1 + z_A)/(1 + z_M)$", fontsize=18)
ax0.get_xaxis().set_visible(False)
ax1.get_xaxis().set_visible(False)
ax2.get_xaxis().set_visible(False)
xmax = 1.6
ax0.set_ylim(0, 1.1 * max(hist0))
ax1.set_ylim(0, 1.1 * max(hist1))
ax2.set_ylim(0, 1.1 * max(hist2))
ax3.set_ylim(0, 1.1 * max(hist3))
ax0.set_xlim(0.2, xmax)
ax1.set_xlim(0.2, xmax)
ax2.set_xlim(0.2, xmax)
ax3.set_xlim(0.2, xmax)
ax0.text(0.98 * xmax,
0.9 * max(hist0),
'Runz xcor',
fontsize=14,
horizontalalignment='right')
ax1.text(0.98 * xmax,
0.9 * max(hist1),
'Runz ELM',
fontsize=14,
horizontalalignment='right')
ax2.text(0.98 * xmax,
0.9 * max(hist2),
'Autoz',
fontsize=14,
horizontalalignment='right')
ax3.text(0.98 * xmax,
0.9 * max(hist3),
'Marz',
fontsize=14,
horizontalalignment='right')
figtext(0.03,
0.7,
r"probability to misidentify [%]",
fontdict={'fontsize': 15},
rotation=90)
#fig.savefig("errorRateqop3.png", bbox_inches='tight', dpi=600, transparent=True)
fig.savefig("errorRateqop3.pdf", bbox_inches='tight', transparent=True)
def plot(self, t, dynamic=False):
"""Produces a quick plot with the internal states at time t.
Parameters
----------
t : float
Dimensional time (in 'time_units') at which to plot.
"""
import matplotlib.pyplot as plt
import matplotlib.gridspec as gridspec
from matplotlib import cm, colors
t_in_seconds = t * self.time_scaling_factor
self.fig = plt.figure(figsize=self.figsize)
self.gridspec = gridspec.GridSpec(self.n_rows, self.n_cols)
self.plots = {}
self.time_lines = {}
self.colorbars = {}
self.axes = []
# initialize empty handles, to be created only if the appropriate plots are made
solution_handles = []
for k, (key, variable_lists) in enumerate(self.variables.items()):
ax = self.fig.add_subplot(self.gridspec[k])
self.axes.append(ax)
x_min, x_max, y_min, y_max = self.axis_limits[key]
ax.set_xlim(x_min, x_max)
if y_min is not None and y_max is not None:
ax.set_ylim(y_min, y_max)
ax.xaxis.set_major_locator(plt.MaxNLocator(3))
self.plots[key] = defaultdict(dict)
variable_handles = []
# Set labels for the first subplot only (avoid repetition)
if variable_lists[0][0].dimensions == 0:
# 0D plot: plot as a function of time, indicating time t with a line
ax.set_xlabel("Time [{}]".format(self.time_unit))
for i, variable_list in enumerate(variable_lists):
for j, variable in enumerate(variable_list):
if len(variable_list) == 1:
# single variable -> use linestyle to differentiate model
linestyle = self.linestyles[i]
else:
# multiple variables -> use linestyle to differentiate
# variables (color differentiates models)
linestyle = self.linestyles[j]
full_t = self.ts_seconds[i]
(self.plots[key][i][j], ) = ax.plot(
full_t / self.time_scaling_factor,
variable(full_t, warn=False),
# color=self.colors[i],
linestyle=linestyle,
)
variable_handles.append(self.plots[key][0][j])
solution_handles.append(self.plots[key][i][0])
y_min, y_max = ax.get_ylim()
ax.set_ylim(y_min, y_max)
(self.time_lines[key], ) = ax.plot(
[
t_in_seconds / self.time_scaling_factor,
t_in_seconds / self.time_scaling_factor,
],
[y_min, y_max],
"k--",
lw=1.5,
)
elif variable_lists[0][0].dimensions == 1:
# 1D plot: plot as a function of x at time t
# Read dictionary of spatial variables
spatial_vars = self.spatial_variable_dict[key]
spatial_var_name = list(spatial_vars.keys())[0]
ax.set_xlabel(
"{} [{}]".format(spatial_var_name, self.spatial_unit), )
for i, variable_list in enumerate(variable_lists):
for j, variable in enumerate(variable_list):
if len(variable_list) == 1:
# single variable -> use linestyle to differentiate model
linestyle = self.linestyles[i]
else:
# multiple variables -> use linestyle to differentiate
# variables (color differentiates models)
linestyle = self.linestyles[j]
(self.plots[key][i][j], ) = ax.plot(
self.first_dimensional_spatial_variable[key],
variable(t_in_seconds, **spatial_vars, warn=False),
# color=self.colors[i],
linestyle=linestyle,
zorder=10,
)
variable_handles.append(self.plots[key][0][j])
solution_handles.append(self.plots[key][i][0])
# add lines for boundaries between subdomains
for boundary in variable_lists[0][0].internal_boundaries:
boundary_scaled = boundary * self.spatial_factor
ax.axvline(boundary_scaled, color="0.5", lw=1, zorder=0)
elif variable_lists[0][0].dimensions == 2:
# Read dictionary of spatial variables
spatial_vars = self.spatial_variable_dict[key]
# there can only be one entry in the variable list
variable = variable_lists[0][0]
# different order based on whether the domains are x-r, x-z or y-z
if self.is_x_r[key] is True:
x_name = list(spatial_vars.keys())[1][0]
y_name = list(spatial_vars.keys())[0][0]
x = self.second_dimensional_spatial_variable[key]
y = self.first_dimensional_spatial_variable[key]
var = variable(t_in_seconds, **spatial_vars, warn=False)
else:
x_name = list(spatial_vars.keys())[0][0]
y_name = list(spatial_vars.keys())[1][0]
x = self.first_dimensional_spatial_variable[key]
y = self.second_dimensional_spatial_variable[key]
# need to transpose if domain is x-z
if self.is_y_z[key] is True:
var = variable(t_in_seconds,
**spatial_vars,
warn=False)
else:
var = variable(t_in_seconds,
**spatial_vars,
warn=False).T
ax.set_xlabel("{} [{}]".format(x_name, self.spatial_unit))
ax.set_ylabel("{} [{}]".format(y_name, self.spatial_unit))
vmin, vmax = self.variable_limits[key]
# store the plot and the var data (for testing) as cant access
# z data from QuadMesh or QuadContourSet object
if self.is_y_z[key] is True:
self.plots[key][0][0] = ax.pcolormesh(
x,
y,
var,
vmin=vmin,
vmax=vmax,
)
else:
self.plots[key][0][0] = ax.contourf(x,
y,
var,
levels=100,
vmin=vmin,
vmax=vmax)
self.plots[key][0][1] = var
if vmin is None and vmax is None:
vmin = ax_min(var)
vmax = ax_max(var)
self.colorbars[key] = self.fig.colorbar(
cm.ScalarMappable(colors.Normalize(vmin=vmin, vmax=vmax)),
ax=ax,
)
# Set either y label or legend entries
if len(key) == 1:
title = split_long_string(key[0])
ax.set_title(title)
else:
ax.legend(
variable_handles,
[split_long_string(s, 6) for s in key],
bbox_to_anchor=(0.5, 1),
loc="lower center",
)
# Set global legend
if len(self.labels) > 1:
fig_legend = self.fig.legend(solution_handles,
self.labels,
loc="lower right")
# Get the position of the top of the legend in relative figure units
# There may be a better way ...
try:
legend_top_inches = fig_legend.get_window_extent(
renderer=self.fig.canvas.get_renderer()).get_points()[1, 1]
fig_height_inches = (self.fig.get_size_inches() *
self.fig.dpi)[1]
legend_top = legend_top_inches / fig_height_inches
except AttributeError: # pragma: no cover
# When testing the examples we set the matplotlib backend to "Template"
# which means that the above code doesn't work. Since this is just for
# that particular test we can just skip it
legend_top = 0
else:
legend_top = 0
# Fix layout
if dynamic:
slider_top = 0.05
else:
slider_top = 0
bottom = max(legend_top, slider_top)
self.gridspec.tight_layout(self.fig, rect=[0, bottom, 1, 1])
def createContourPlot(self, title, ax, x, y, z, rangey, rangez, startTime,
endTime, sdt_count):
tmin = time.mktime(startTime.timetuple())
tmax = time.mktime(endTime.timetuple())
tgrid_max = 1000 # Reasonable maximum width for time-depth-flot plot is about 1000 pixels
dgrid_max = 200 # Height of time-depth-flot plot area is 200 pixels
dinc = 0.5 # Average vertical resolution of AUV Dorado
nlevels = 255 # Number of color filled contour levels
zmin = rangez[0]
zmax = rangez[1]
dmin = rangey[0]
dmax = rangey[1]
scale_factor = 1
# 2 points define a line, take half the number of simpledepthtime points
sdt_count = int(max(sdt_count, 2) / 2)
if sdt_count > tgrid_max:
sdt_count = tgrid_max
xi = np.linspace(tmin, tmax, sdt_count)
#print 'xi = %s' % xi
# Make depth spacing dinc m, limit to time-depth-flot resolution (dgrid_max)
y_count = int((dmax - dmin) / dinc)
if y_count > dgrid_max:
y_count = dgrid_max
yi = np.linspace(dmin, dmax, y_count)
#print 'yi = %s' %yi
try:
scale_factor = float(tmax - tmin) / (dmax - dmin)
except ZeroDivisionError as e:
logger.warn(
'Not setting scale_factor. Scatter plots will still work.')
contour_flag = False
scale_factor = 1
else:
logger.warn('self.scale_factor = %f', scale_factor)
xi = xi / scale_factor
xg = [xe / scale_factor for xe in x]
contour_flag = True
zi = []
cs = None
# Register the signal function handler
signal.signal(signal.SIGALRM, self.handler)
# Define a timeout of 90 seconds for gridding functions
signal.alarm(90)
if not self.data:
logger.warn('no data found to plot')
signal.alarm(0)
raise Exception('no data')
if contour_flag:
try:
logger.warn(
'Gridding data with sdt_count = %d, and y_count = %d',
sdt_count, y_count)
zi = self.gridData(xg, y, z, xi, yi)
signal.alarm(0)
except KeyError:
logger.warn('Got KeyError. Could not grid the data')
contour_flag = False
scale_factor = 1
try:
# use RBF
logger.warn('Trying radial basis function')
xi, yi, zi = self.gridDataRbf(tmin, tmax, dmin, dmax, xg,
y, z)
contour_flag = True
signal.alarm(0)
except Exception as e:
logger.warn('Could not grid the data' + str(e))
except Exception as e:
logger.warn('Could not grid the data' + str(e))
contour_flag = False
try:
# use RBF
logger.warn('Trying radial basis function')
xi, yi, zi = self.gridDataRbf(tmin, tmax, dmin, dmax, xg,
y, z)
contour_flag = True
signal.alarm(0)
except Exception as e:
logger.warn('Could not grid the data' + str(e))
try:
if scale_factor > 1 and contour_flag:
ax.set_xlim(tmin / scale_factor, tmax / scale_factor)
else:
ax.set_xlim(tmin, tmax)
self.scale_factor = scale_factor
ax.set_ylim([dmax, dmin])
ax.set_ylabel('depth (m)', fontsize=8)
ax.tick_params(axis='both', which='major', labelsize=8)
ax.tick_params(axis='both', which='minor', labelsize=8)
if contour_flag:
logger.debug('Contouring the data')
cs = ax.contourf(xi,
yi,
zi,
levels=np.linspace(zmin, zmax, nlevels),
cmap=self.cm_jetplus,
extend='both')
# this will show the points where the contouring occurs
#ax.scatter(x,y,marker='.',s=2,c='k',lw=0)
else:
logger.debug('Plotting the data')
cs = ax.scatter(x,
y,
c=z,
s=20,
marker='.',
vmin=zmin,
vmax=zmax,
lw=0,
alpha=1.0,
cmap=self.cm_jetplus)
# limit the number of ticks
max_yticks = 5
yloc = plt.MaxNLocator(max_yticks)
ax.yaxis.set_major_locator(yloc)
except Exception as e:
logger.debug('Error - trying to plot the data')
cs = ax.scatter(x,
y,
c=z,
s=20,
marker='.',
vmin=zmin,
vmax=zmax,
lw=0,
alpha=1.0,
cmap=self.cm_jetplus)
return cs, zi
data[0][0] + (data[0][1] - data[0][0]) * i for i in range(len(data[0]))
]
#magic number comes from experiments
width = int(len(data[0]) / 29)
length = 4
if width * 100 > pow(2, 16):
width = int(pow(2, 16))
print('Reached max width for plot: ' + str(width) + ' pixels')
fig, ax = plt.subplots(figsize=(width, length))
ax.grid(True)
plt.MaxNLocator(5)
for i in range(len(data) - 1):
ax.plot(data[0], data[i + 1], label=labels[i + 1])
start, end = ax.get_xlim()
# formatting x axis marking
hours = mdates.HourLocator(interval=1)
h_fmt = mdates.DateFormatter('%H:%M:%S')
# formatting density of x axis markings
ax.xaxis.set_major_locator(hours)
ax.xaxis.set_major_formatter(h_fmt)
ax.xaxis.set_ticks(np.arange(start, end, 0.00035))
ax.tick_params(axis='x', rotation=45)
fig, ax1 = plt.subplots()
color = "blue"
ax1.set_xlabel('date')
for tick in ax1.get_xticklabels():
tick.set_rotation(45)
ax1.set_ylabel('Number of Postive Cases', color=color)
ax1.plot(date, num_pos, color=color)
#ax1.plot(date, ave_num, color=color)
ax1.tick_params(axis='y', labelcolor=color)
ax2 = ax1.twinx()
color = 'red'
ax2.set_ylabel('Postive Rate', color=color)
ax2.plot(date, rate_pos, color=color)
#ax2.plot(date, ave_rate, color=color)
ax2.tick_params(axis='y', labelcolor=color)
ax1.xaxis.set_major_locator(plt.MaxNLocator(25))
fig.tight_layout()
#plt.show()
#7-Day Rolling Average
fig,ax3 = plt.subplots()
color = "blue"
ax3.set_xlabel('date')
for tick in ax3.get_xticklabels():
tick.set_rotation(45)
ax3.set_ylabel('Number of Postive Cases', color=color)
ax3.plot(date, ave_num, color=color)
ax3.tick_params(axis='y', labelcolor=color)
ax4 = ax3.twinx()
color = 'red'
ax4.set_ylabel('Postive Rate', color=color)
def model_comparisson(halo, mask=False):
fig, axes = plt.subplots(ncols=3, nrows=1, sharey=True)
bmin = halo.bmin
bmaj = halo.bmaj
scale = 1.
model4 = halo.result4.model
model6 = halo.result6.model
model8 = halo.result8.model
ra = halo.ra.value
dec = halo.dec.value
for axi in axes.flat:
axi.xaxis.set_major_locator(plt.MaxNLocator(5))
axi.xaxis.set_major_formatter(ScalarFormatter(useOffset=False))
axi.yaxis.set_major_formatter(ScalarFormatter(useOffset=False))
fig.set_size_inches(3.2*5,5.1)
vmin=-2*(halo.rmsnoise/halo.pix_area).to(uJyarcsec2).value
vmax=4*(halo.result4.params_units[0])
data = (np.copy(halo.result4.data)/halo.pix_area).to(uJyarcsec2).value
noise = (halo.rmsnoise/halo.pix_area).to(uJyarcsec2).value
masked_data = data.copy()
#if mask:
masked_data[halo.result4.image_mask==1]= -10000.
LEVEL = np.arange(1,7)*(halo.rmsnoise/halo.pix_area).to(uJyarcsec2).value
#NORM = mplc.LogNorm(vmin=0.4*(halo.rmsnoise/halo.pix_area).to(uJyarcsec2).value,
# vmax=20*(halo.rmsnoise/halo.pix_area).to(uJyarcsec2).value)
#NORM = SymLogNorm(2.*halo.result4.params_units[0] , linscale=1.0, vmin=vmin, vmax=vmax)
#NORMres = mplc.Normalize(vmin=-2.*(halo.rmsnoise/halo.pix_area).to(uJyarcsec2).value,
# vmax=1.*(data/halo.pix_area).to(uJyarcsec2).value.max())
NORMres = mplc.Normalize(vmin=-2.*noise, vmax=2.*masked_data.max())
im1 = axes[0].imshow(masked_data,
cmap='inferno', origin='lower',
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres)
try:
cont1 = axes[0].contour((model4/halo.pix_area).to(uJyarcsec2).value,
colors='white', levels=LEVEL, alpha=0.6,
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres,linewidths=1.)
cont2 = axes[0].contour(masked_data,
colors='lightgreen', levels=np.array([-999.8]), alpha=0.9, linestyles='-',
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres,linewidths=1.5)
except:
print('PROCESSING: Failed making contours')
pass
axes[0].set_title('Circular\n $S_{\\mathrm{1.5 GHz}}=%.1f\\pm%.1f$ mJy' % (halo.result4.flux_val.value, halo.result4.flux_err.value), fontsize=15)
axes[0].set_xlabel('RA [deg]', fontsize=labelsize)
axes[0].set_ylabel('DEC [deg]', fontsize=labelsize)
axes[0].grid(color='white', linestyle='-', alpha=0.25)
draw_sizebar(halo,axes[0], scale)
draw_ellipse(halo,axes[0], bmin, bmaj)
plt.tight_layout()
im2 = axes[1].imshow(masked_data,
cmap='inferno', origin='lower',
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres)
try:
cont3 = axes[1].contour((model6/halo.pix_area).to(uJyarcsec2).value,
colors='white', levels=LEVEL, alpha=0.6,
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres,linewidths=1.)
cont4 = axes[1].contour(masked_data,
colors='lightgreen', levels=np.array([-999.8]), alpha=0.9, linestyles='-',
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres,linewidths=1.5)
except:
print('PROCESSING: Failed making contours')
pass
axes[1].set_title('Elliptical\n $S_{\\mathrm{1.5 GHz}}=%.1f\\pm%.1f$ mJy' % (halo.result6.flux_val.value, halo.result8.flux_err.value), fontsize=15)
axes[1].set_xlabel('RA [deg]', fontsize=labelsize)
axes[1].set_ylabel('DEC [deg]', fontsize=labelsize)
axes[1].grid(color='white', linestyle='-', alpha=0.25)
draw_sizebar(halo,axes[0], scale)
draw_ellipse(halo,axes[0], bmin, bmaj)
plt.tight_layout()
im3 = axes[2].imshow(masked_data,
cmap='inferno', origin='lower',
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres)
try:
cont5 = axes[2].contour((model8/halo.pix_area).to(uJyarcsec2).value,
colors='white', levels=LEVEL, alpha=0.6,
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres,linewidths=1.)
cont6 = axes[2].contour(masked_data,
colors='lightgreen', levels=np.array([-999.8]), alpha=0.9, linestyles='-',
extent=(ra.max(),ra.min(),dec.min(),dec.max()), norm = NORMres,linewidths=1.5)
except:
print('PROCESSING: Failed making contours')
pass
axes[2].set_title('Skewed \n $S_{\\mathrm{1.5 GHz}}=%.1f\\pm%.1f$ mJy' % (halo.result8.flux_val.value, halo.result8.flux_err.value), fontsize=15)
axes[2].set_xlabel('RA [deg]', fontsize=labelsize)
axes[2].set_ylabel('DEC [deg]', fontsize=labelsize)
axes[2].grid(color='white', linestyle='-', alpha=0.25)
draw_sizebar(halo,axes[0], scale)
draw_ellipse(halo,axes[0], bmin, bmaj)
plt.tight_layout()
import matplotlib.ticker as ticker
cbar = fig.colorbar(im3)
cbar.ax.set_ylabel('$\\mu$Jy arcsec$^{-2}$',fontsize=labelsize)
#cbar.formatter = ScalarFormatter(useMathText=False)
#cbar.formatter = ticker.LogFormatter(base=10.,labelOnlyBase=True)
cbar.formatter = ticker.StrMethodFormatter('%.2f')
plt.savefig(halo.plotPath +halo.file.replace('.fits','')+'_mcmc_model_ALL.pdf')
#plt.show()
plt.clf()
plt.close(fig)
