Python for Astronomers

matplotlib (1)

matplotlib

library for working with 2D plots project was started to emulate matlab Three main parts:

the pylab interface the matplotlib API (matplotlib frontend) the backends (GTK, GTKAgg, PostScript, PDF, png, ...)

ipython -pylab

matplotlib can be used interactively and non-interactively start an interactive session for matplotlib with ipython -pylab

simple plotting
>>> plot([1,2,4]) >>> ylabel(just numbers) >>> >>> x = [1,2,3,4] >>> y = [1,4,9,16] >>> plot(x,y)

matplotlib
home, forward, back pan & zoom zoom to rect configure subplot save to file

simple plotting
>>> clf() # clear current figure >>> plot(x, y, 'ro-')
Colors: blue, green, red, cyan, magenta, yellow, black, white Marker symbols: o - circle, ^ - triangle up, < - triangle left, s - square, + - plus symbol, . - points, , - pixels, rest see docs Linestyle: - solid, -- dashed, -. dash-dotted, : dotted

simple plotting
>>> t = linspace(0,2*pi,10) >>> s = linspace(-pi, pi, 10) >>> >>> plot(t, sin(t), s, cos(s)) >>> plot(t, 0.2+cos(t), 'g^-', ... s, 0.2+sin(s), 'rs:')

line properties
>>> t = linspace(0,2*pi,1000) >>> st = sin(t) >>> >>> plot(t, st, color='green') >>> plot(t, st**2, color='#C000C0') >>> plot(t, st+1, color=(0.5,0,0,1)) >>> # RGB as hex #C000C0 >>> # RGB(A) as tuple (0.5,0,0,0.5) >>> # alpha: 0.0 transparent, 1 opaque

line properties
>>> x = arange(10) >>> y = x**2 >>> clf() >>> plot(x,y, color='red', linewidth=2, ... >>> >>> plot(x,y+1, c='green', ... lw=3, ls='--') linestyle=':')

marker
>>> plot(x,y, color='r', ... ... >>> >>> # !!! to get unfilled markers: >>> plot(x,y+1, color='red', ... ... marker='v', markerfacecolor='None') marker='s', markerfacecolor='m', markeredgecolor='g')

lines
>>> clf() >>> lines = plot(x,y, x,y+10) >>> print lines >>> lines[0].set_color('red') # mpl API >>> draw() >>> l0 = lines[0] >>> l0.set<Tab> >>> >>> setp(lines, linestyle='--') >>> setp(lines, marker='D')

Hipparcos CMD
>>> import numpy >>> filename = '/home/mmetz/hipcmd.dat' >>> cmd = numpy.loadtxt(filename) >>> >>> # recall: 2nd column BT-VT >>> # >>> clf() >>> plot(cmd[:,1], cmd[:,2], 'k,') 3rd column VT

Hipparcos CMD
>>> xlim(-1,3) >>> ylim(26,8) >>> >>> xlabel('BT-VT') >>> ylabel('VT')

Hipparcos CMD
>>> # numpy arrays >>> color = cmd[:,1] >>> red = (color>0.8) >>> print red.dtype >>> blue = ~red >>> >>> plot(cmd[red,1], cmd[red,2], 'r,') >>> plot(cmd[blue,1], cmd[blue,2],'b,')

numpy bool arrays

>>> # testing numpy boolean arrays ! >>> lst = [] >>> if lst: print 'ok' >>> >>> ba = numpy.array([True, False]) >>> if ba.any(): print 'any' >>> if ba.all(): print 'all'

plotting
>>> clf() >>> >>> t = linspace(0,2*pi,1000) >>> s = linspace(-pi,pi,1000) >>> >>> plot(t, sin(t)) >>> plot(s, cos(s))

Figure & axes

Figure

Axes

Axis

axes
>>> # pylab interface: >>> xlim(-pi, 2*pi) >>> xlabel('T') >>> >>> # matplolib API: >>> ax = gca() # get current axes >>> ax.set_xlim(-pi, pi)# lims of xaxis >>> ax.set_xlabel('t') >>> draw()

Figure
>>> fig = gcf() >>> >>> print fig.get_figwidth() >>> >>> >>> >>> fig.clf() >>> draw() # figure width in inch # resize the figure and try again # get current Figure

Figures
>>> fig1 = figure() >>> fig2 = figure( figsize=(8,8) ) >>> draw() >>> nax = fig2.gca() >>> draw() >>> >>> ax = gca() >>> print nax is ax # pylab

subplots
>>> fig2.clf() >>> draw() >>> >>> ax1 = subplot(211) # pylab >>> # equiv: subplot(2,1,1) >>> # (nrows, ncols, nplot) >>> >>> ax2 = fig2.add_subplot(212) >>> draw()

subplots
>>> # pylab >>> plot(t, sin(t)) # plot to current >>> >>> >>> >>> # matplotlib API >>> ax1.plot(s, sin(s)) >>> draw() # axes in current # figure, ie. ax2

subplots
>>> # what if I forgot to assign the >>> # return value of subplot to a var? >>> fig = gcf() >>> print fig.axes >>> ax1 = fig.axes[0] >>> >>> # for figures: >>> f1 = figure(1) >>> f2 = figure(num=2)

Exercise subplots

create 2x2 subplots and plot x, x**2, x**3 and sqrt(x) for x=linspace(0,1,100) in the 4 subplots experiment with setting x/ylabels for the subplots and axis limits

working with text

>>> clf() >>> plot(t, sin(t)) >>> >>> xlabel('t') >>> xlabel('$t$') >>> ylabel('$\sin(t)$') >>> title('plot of $\sin(t)$') >>> suptitle('The figure title')

Exercise

open the configure subplot dialog and change the values left and top reset view to home

working with text

>>> # add text to the plot >>> text(4, 0.8, 'text in plot') >>> >>> v = 3*pi/4 >>> text(v, sin(v), '$s = \sin(t)$')

working with text

>>> text(pi,0, 'More Text', ... ... >>> >>> # also short forms: ha & va; can be >>> # ha: left, center, right >>> # va: top, center, bottom, baseline >>> # defaults are: left, bottom horizontalalignment='right') verticalalignment='top') >>> text(pi,0, 'Last Text',

working with text

>>> # text(x,y, 'text') places the text >>> # at x,y in data coords >>> ax = gca() >>> text(0.1, 0.9, 'in axes coords', ... >>> >>> # fig/axes coords x,y: >>> # >>> # 0,0 left, bottom 1,1 right, top transform=ax.transAxes) >>> figtext(0.5, 0.5, 'in fig coords')

working with text

>>> text(1, -0.5, 'Some Text', ... ... ... >>> >>> # matplotlib API >>> txt = text(0.5, -0.5, 'More Text') >>> txt.set_color('red') >>> draw() fontsize=14, color='red', fontweight='bold', style='italic')

working with text

>>> text(1.5, 0.8, 'Cats & Dogs', ... >>> >>> text(1.5, 0.2, 'Dragons', ... >>> >>> # available font families: >>> # >>> # serif, sans-serif, cursive, fantasy, monospace family='serif') name='Verdana')

working with text

>>> text(0.5, 0.5, 'Inclinde text', ... >>> >>> text(0.5, 0.5, 'Hurt\nmy eyes', ... ... color = 'red', backgroundcolor='green') rotation=30)

working with text

>>> boxprops = {'facecolor':'red', ... ... >>> >>> text(2,0.5, 'Text in a box', ... bbox=boxprops) 'alpha':0.8, 'pad':5}

Legends
>>> clf() >>> plot(t, sin(t)) >>> plot(t, cos(t)) >>> legend( ('sin', 'cos') )

Legends
>>> clf() >>> plot(t, sin(t), label='sin') >>> plot(t, cos(t), label='$\cos$') >>> legend() >>> >>> # matplotlib API >>> lines = plot(t, sin(t)) >>> lines[0].set_label('$\sin(t)$')

Legends
>>> # location code: >>> legend(loc=2) >>> >>> # location string: >>> legend(loc='lower left') >>> >>> # axes coords (not plot coords !): >>> legend(loc=(0.5, 0.5))

Exercise Text

Add some text in a plot, once using data coordinates and once using axes coords (... ,transform=ax.transAxes) Change the x/y limits or use the Pan/Zoom Tools to see the different behaviour add multiline text ('Hallo\nUniverse') to the plot, and check the behaviour of the parameter multialignment='right' ('center', 'left')

log-log plots
>>> clf() >>> semilogy(t, t**2) >>> >>> clf() >>> semilogx(t, sqrt(t)) >>> >>> clf() >>> loglog(t, sqrt(t**3))

scatter plots
>>> import numpy >>> r1 = numpy.random.random(10) >>> r2 = numpy.random.random(10) >>> >>> # scatter plot using plot >>> plot(r1, r2, ls='None', marker='o', ... markersize=10)

scatter plots
>>> scatter(r2, r1) >>> scatter(r2, r1, s=80, c='red', ... >>> >>> # plot: markersize linear scale >>> # >>> # default=6 ie. diam = sqrt(area) >>> # scatter: s area scale, default=20 marker='s')

scatter plots
>>> cols = ['red', 'blue', 'green'] >>> scatter(r1+0.1, r2, c=cols) >>> >>> sizes = linspace(20,80,10) >>> scatter(r1+0.2, r2, s=sizes)

histograms
>>> clf() >>> m, s = 40., 3. >>> X = m + s * numpy.random.randn(888) >>> hist(X) >>> >>> n, b, p = hist(X) >>> # numbers, binedges, patches >>> print n, sum(n) >>> print b

histograms
>>> clf() >>> binedges = linspace(30,50,21) >>> print binedges >>> >>> n, b, p = hist(X, bins=binedges) >>> print b >>> print len(n) >>> print len(b)

histograms
>>> clf() >>> >>> n, b, p = \ ... hist(X, bins=20, range=(30,50)) >>> print b >>> print len(b)

histograms
>>> clf() >>> hist(X, histtype='stepfilled') >>> >>> hist(X, histtype='step', ... >>> >>> clf() >>> hist(X, normed=True) cumulative=True)

Final note

If you want to use pylab in a script !

>>> import pylab >>> import numpy >>> x = numpy.linspace(0,1,100) >>> pylab.plot(x, x**2) >>> pylab.show()

Exercise

The files HD152200.uvspec and HD152234.uvspec in /home/mmetz contain FUSE spectra (thanks to Ole !) of the star HD* as ASCII data (wavelength, count rate). i) Read the data, normalise the count rates, and plot both spectra into one subplot. ii) Create a subplot below the first one and plot binned spectra (use numpy to bin the data)