MatplotLib 'saveFig()' Fullscreen

Question

I used MatplotLib with Cartopy to generate some data images. The problem is that when I set the frame size to fullscreen and use plt.show() the image is perfect and the resolution is fine.

However, when I save this figure using 'plt.savefig()' the image saved keeps with its original size (not fullscreen).

Showing outcome images:

My code is the following:

def plot_tec_cartopy(descfile): global matrixLon, matrixLat, matrixTec

ax = plt.axes(projection=cartopy.crs.PlateCarree())

v = np.linspace(0, 80, 46, endpoint=True)
cp = plt.contourf(matrixLon, matrixLat, matrixTec, v, cmap=plt.cm.rainbow)
plt.clim(0, 80)
plt.colorbar(cp)

ax.add_feature(cartopy.feature.COASTLINE)
ax.add_feature(cartopy.feature.BORDERS, linestyle=':')
ax.set_extent([-85, -30, -60, 15])

# Setting X and Y labels using LON/LAT format
ax.set_xticks([-85, -75, -65, -55, -45, -35])
ax.set_yticks([-60, -55, -50, -45, -40, -35, -30, -25, -20, -15, -10, -5, 0, 5, 10, 15])
lon_formatter = LongitudeFormatter(number_format='.0f',
                                   degree_symbol='',
                                   dateline_direction_label=True)
lat_formatter = LatitudeFormatter(number_format='.0f',
                                  degree_symbol='')
ax.xaxis.set_major_formatter(lon_formatter)
ax.yaxis.set_major_formatter(lat_formatter)

plt.title('Conteúdo Eletrônico Total', style='normal', fontsize='12')

# Acquiring Date
year, julianday = check_for_zero(descfile.split('.')[2]), descfile.split('.')[3]
hour, minute = descfile.split('.')[4], descfile.split('.')[5].replace('h','')
date = datetime.datetime(int(year), 1, 1, int(hour), int(minute)) + datetime.timedelta(int(julianday)-1)
month = date.month
day = date.day

# Set common labels
ax.text(1.22, 1.05, 'TEC', style='normal',
    verticalalignment='top', horizontalalignment='right',
    transform=ax.transAxes,
    color='black', fontsize=11)
ax.text(1, 0.005, 'EMBRACE/INPE', style='italic',
    verticalalignment='bottom', horizontalalignment='right',
    transform=ax.transAxes,
    color='black', fontsize=10)
ax.text(1, 0.995, str(date) + ' UT', style='italic',
    verticalalignment='top', horizontalalignment='right',
    transform=ax.transAxes,
    color='black', fontsize=10)
ax.text(0.5, -0.08, 'Copyright \N{COPYRIGHT SIGN} 2017 INPE - Instituto Nacional de',
    style='oblique', transform=ax.transAxes,
    verticalalignment='bottom', horizontalalignment='center',
    color='black', fontsize=8)
ax.text(0.5, -0.108, 'Pesquisas Espacias. Todos direitos reservados',
    style='oblique', transform=ax.transAxes,
    verticalalignment='bottom', horizontalalignment='center',
    color='black', fontsize=8)

manager = plt.get_current_fig_manager()
manager.resize(*manager.window.maxsize())

figName = 'tec.map' + '.' + str(year) + '.' + str(julianday) + '.' + str(hour) + '.' + str(minute) + 'h.png'
#plt.show()
plt.savefig(figName, dpi=500)
plt.clf()

Maybe I need to set some parameter into savefig() to say it that it needs to save my modified frame? Can someone help me with this issue?

Thanks in advance.

Answer 1

Coming from MATLAB, it is not intuitive that your displayed figure does not have to affect the saved one in terms of dimensions, etc. Each one is handled by a different backend, and you can modify the dpi and size_inches as you choose.

Increasing the DPI is definitely going to help you get a large figure, especially with a format like PNG, which does not know about the size in inches. However, it will not help you scale the text relative to the figure itself.

To do that, you will have to use the object oriented API, specifically, figure.set_size_inches, which I don't think has an equivalent in plt. Replace

plt.savefig(figName, dpi=500)

with

fig = plt.gcf()
fig.set_size_inches((8.5, 11), forward=False)
fig.savefig(figName, dpi=500)

The size 8.5, 11 is the width and height of the standard paper size in the US, respectively. You can set it to whatever you want. For example, you can use your screen size, but in that case be sure to get the DPI right as well.

Answer 2

Just to add some context on @Mad Physicist answer. If someone is trying to use it with a new version of matplotlib, you will get an AttributeError: 'Figure' object has no attribute 'save'. You also need to be careful about when you call plt.show() otherwise you will get a blank image. You need to update the code as follow:

# Create a plot
plt.barh(range(top), imp, align='center')
plt.yticks(range(top), names)

# Get the current figure like in MATLAB
fig = plt.gcf()
plt.show() # show it here (important, if done before you will get blank picture)
fig.set_size_inches((8.5, 11), forward=False)
fig.savefig(figName, dpi=500) # Change is over here

Hope it helps!