Python mpldatacursor.datacursor函数代码示例

 def plot_histogram(self, bf, row, col, plot_sigma=True,
                    hold=False):
     self.bf = bf
     if self.bf is None:
         raise Exception("Need to pass in a BinFile instance to plot object")
     if not hasattr(self.bf, 'data_out'):
         raise Exception("BinFile instance does not have any data.")
     if not hold:
         self.clear()
         self.lines = []
     try:
         h = self.bf.data_out[row, col, :, :].real.flatten()
     except:
         print "Error in extracting histogram"
     if hasattr(self.bf, 'pixel_label'):
         label="Row%d, Col%d (Pix: %s)" % (self.bf.row_start+row, self.bf.col_start+col, self.bf.pixel_label.get((row, col), 'NC'))
     else:
         label="Row%d, Col%d" % (self.bf.row_start+row, self.bf.col_start+col)
     self.hist(h, bins=20, label=label)
     if MATPLOTLIBV1_0:
         datacursor()
     self.set_subplot_title("%s" % self.bf.basename)
     if plot_sigma:
         y1, y2 = self.get_ylims()
         for x in (-8, 8):
             self.plot([x, x], [y1, y2], 'r--', linewidth=2, 
                       label="_nolegend_")
     self.set_xlim(-64, 64)
     self.set_legend(loc='best')

    def plot(self):
        specs = pl.array( list(set([ l['spec'] for l in self.lines ])) )
        specs.sort()
        self.specs = specs

        pl.figure()
        pl.hold('on')
        pl.grid('on')
        pl.jet()

        lines = [] 
        lines_spec = list(pl.zeros(len(specs)))
        for i in range(0,len(self.lines)):
            ispc = pl.find( specs == self.lines[i]['spec'] )
            self.colr = pl.cm.get_cmap()( float(ispc)/len(specs) )
            wl = self.lines[i]['wave']
            ri = float(self.lines[i]['rel_int'])
            lines.append( pl.plot( [wl, wl], [0., ri if not isnan(ri) else 0.], '.-', color=self.colr )[0] )
            lines_spec[ispc] = lines[-1]
        datacursor(lines,formatter='x={x:8.3f}\ny={y:8.3f}'.format)

        pl.rc('text',usetex=True)
        pl.xlabel('$\lambda ~ [\AA]$')
        pl.ylabel('relative intensity [arb]')
        pl.title('Spectrum for '+self.spec+' from NIST ASD')

        if len(specs) > 1:
            pl.legend( lines_spec,specs )

        pl.show()

    def measure_iv(self):
        self.get_vd_values()
        self.textBrowser.append('Measuring I-V from ' + str(self.vd_start) +
                                'V to ' + str(self.vd_stop) + ' V')
        vd_values = np.linspace(self.vd_start, self.vd_stop, num=self.vd_steps,
                                endpoint=True)
        id_values = np.zeros_like(vd_values)
#        self.measure_current()
        for i, vd in np.ndenumerate(vd_values):
            self.set_voltage_vd(vd)
            id_values[i] = self.measure_current()
            if self.stop_engaged:
                print('stop')
                self.stop_engaged = False
                break

            self.update_progress((i[0] + 1.0)/self.vd_steps)
        self.set_voltage_vd(0)
        self.textBrowser.append('Measurement completed')
        data = pd.DataFrame({'Voltage': vd_values, 'Current': id_values})
        data.set_index('Voltage', inplace=True)
        self.data_iv = data
        data.to_csv('measurement_iv.csv')
        data.to_msgpack('measurement_iv.msgpack')
        ax = data.plot()
        datacursor(ax)
#        datacursor(display='single', draggable=True)
        winsound.Beep(750, 1000)
        fig = ax.get_figure()
        fig.savefig('measurement_iv.png')
        plt.show()

def plot_k_walls(k_walls, plot_range=None,
                 plot_data_points=False,):
    """
    Plot K-walls for debugging purpose.
    """
    pyplot.figure()
    pyplot.axes().set_aspect('equal')

    for k_wall in k_walls:
        xs = k_wall.get_xs() 
        ys = k_wall.get_ys() 
        pyplot.plot(xs, ys, '-', label=k_wall.identifier)

        if(plot_data_points == True):
            pyplot.plot(xs, ys, 'o', color='k', markersize=4)

    if plot_range is None:
        pyplot.autoscale(enable=True, axis='both', tight=None)
    else:
        [[x_min, x_max], [y_min, y_max]] = plot_range
        pyplot.xlim(x_min, x_max)
        pyplot.ylim(y_min, y_max)

    mpldatacursor.datacursor(
        formatter='{label}'.format,
        hover=True,
    )

    pyplot.show()

def exploredata1d(data, slider='chans', stack='ants'):
    """ Set up interactive 1d (line) plotting for vis data of dimension (ints, ants, chans). """
    
    axdict = {'ints': 0, 'ants': 1, 'chans': 2}
    assert slider in axdict.keys() and stack in axdict.keys(), 'slider or stack param not allowed'

    slax = axdict[slider]
    # need to account for axis shift after first 'take'
    stax = axdict[stack] if axdict[stack] <= slax else axdict[stack] - 1
    slmax = data.shape[axdict[slider]]
    stmax = data.shape[axdict[stack]]

    xaxis = [name for name in axdict.keys() if name != slider and name != stack][0]
    
    fcndict = {'Real': np.real, 'Imag': np.imag, 'Amp': np.abs, 'Phase': np.angle}

    @interact(sl=(0, slmax, 1), f=['Real', 'Imag', 'Amp', 'Phase'])
    def plotautos(sl, f):
        pl.figure(figsize=(15,8))
        pl.clf()
        pl.xlabel(xaxis)
        pl.ylabel(f)
        
        fcn =  fcndict[f]
        
        for st in range(stmax):
            pl.plot(fcn(data.take(sl, axis=slax).take(st, axis=stax)), label='{0} {1}'.format(stack.rstrip('s'), st))
            
        print('Plotting {0} vs. {1}.'.format(f, xaxis))
        print('Slider for {0}. A line per {1}.'.format(slider, stack.rstrip('s')))
        print('Click on a line to see {0} number'.format(stack.rstrip('s')))
    datacursor(formatter='{label}'.format)

def Graph_Instant_boot():
    print ('inside Graph_Instant_boot graph')
    Plot1 = np.loadtxt('Output_VHFS_Instant_boot.txt')
    Plot2 = np.loadtxt('Output_VHFS_Instant_boot_2.txt')
    fig = plt.figure()
    #pl.subplot(311)
    
    pl.title('Instant_Boot: VHFS1 vs VHFS2')    
    pl.xlabel('Incrementals')
    pl.ylabel('Time Taken in secs')
    pl.xlim(-1,11)
    
    x = Plot1[:,0]
    y = Plot1[:,1]
    y2 = Plot2[:,1]
    #yerr = Plot1[:,4]
    #xticks = Plot1[:,1]
    #plt.xticks(xticks)
    #plt.xscale('log')
    #plt.yscale('log')
    #plt.xticks(xticks)
    #pl.errorbar(x,y,yerr,ecolor='b')
    pl.plot(x,y, 'b', label = 'VHFS1', marker = 'o')
    pl.plot(x,y2, 'y', label = 'VHFS2', marker = 'o')
    pl.legend(loc = 'upper right', numpoints = 1)
    datacursor(display='multiple', draggable=True)
    #pl.show()
 
    pl.savefig('Instant_boot_VHFS_vs_VHD.ps')
    pl.savefig('Instant_boot_VHFS_vs_VHD.png')
    pl.savefig('Instant_boot_VHFS_vs_VHD.pdf')
    print ('endddd')
    
    plt.close(fig)

def main(icorr_mean_list):
    suffix = rate2suffix(icorr_mean_list)
    # load data
    datapath = os.path.join(os.path.abspath('./'), 'data')
    filename = 'popdata_'+suffix+'.npz'
    datafile = os.path.join(datapath,filename)
    if os.path.isfile(datafile) is False:
        print 'no data available, execute Life_Cycle_Optimization.py with \
                icorr_mean_list={} fist'.format(icorr_mean_list)
        sys.exit(1)
    else:
        popdata = np.load(datafile)
        allpop = popdata['allpop']
        allfits = popdata['allfits']
        front = popdata['front']
        frontfits = popdata['frontfits']
        pop = popdata['pop']
        popfits = popdata['popfits']

    plt.ion()
    plt.close('all')

    ##plt.semilogx(np.array(frontfits)[:,0], np.array(frontfits)[:,1], 'bo', markeredgecolor='b')
    #plt.semilogx(np.array(popfits)[:,0], np.array(popfits)[:,1], 'bo', markeredgecolor='b')
    #for popfit in popfits:
    for ind, popfit in zip(pop, popfits):
        ## journal version
        #plt.semilogx(popfit[0], popfit[1], 'bo',markeredgecolor='b',
                #label=u'flexure: {:d}, shear: {:d}, deck: {:d}'.format(ind[0], ind[1], ind[2]))
        # conference version
        plt.semilogx(popfit[0], popfit[1], 'bo',markeredgecolor='b',
                label=u'({:d},{:d},{:d})'.format(ind[0], ind[1], ind[2]))

    plt.ylim((-1,np.max(popfits)*1.1))
    ax = plt.gca()
    ax.ticklabel_format(axis='y', style='sci', scilimits=(-3,3))

    plt.xlabel(u'Failure probability')
    plt.ylabel(u'Strengthening cost (mm\\textsuperscript{3})')

    ## journal version
    #annotate_text = u'P={x:.2e}, C={y:.2e}\n {{ {label} }}'
    #datacursor(formatter=annotate_text.format,display='multiple', draggable=True,
            #bbox=None, fontsize=9,
            #arrowprops=dict(arrowstyle='-|>', connectionstyle='arc3', facecolor='k'))
    # conference version
    annotate_text = u'{label}'
    datacursor(formatter=annotate_text.format,display='multiple', draggable=True,
            bbox=None, fontsize=9,
            arrowprops=dict(arrowstyle='-|>', connectionstyle='arc3', facecolor='k'))

    pause = raw_input('press any key after annotation...')

    plt.semilogx(np.array(allfits)[:,0], np.array(allfits)[:,1], 'o', markerfacecolor='lightgrey',
            markeredgecolor='lightgrey', alpha=0.8)
    plt.semilogx(np.array(popfits)[:,0], np.array(popfits)[:,1], 'bo', markeredgecolor='b')

    def implot_data(self, bf, data_type='amp',
                    vmin=None, vmax=None,
                    hold=False, title=None,
                    colorbar=True, sti=False, **kwargs):
        self.bf = bf
        if self.bf is None:
            raise Exception("Need to pass in a BinFile instance to plot object")
        self.check_alive() 
        if not hold:
            self.clear()
        if not hasattr(self.bf, 'cross_corr') and not hasattr(self.bf, 'sti_cc'):
            raise Exception("BinFile does not have cross correlation data. get_cross_corr_data() first on binfile")
        if sti:
            if not hasattr(self.bf, 'sti_cc'):
                raise Exception("BinFile does not have sti cross corr data. Run sti_cross_correlate first")
            self.bf.cc = self.bf.sti_cc.mean(axis=3).mean(axis=2)
        else:
            self.bf.cc = self.bf.cross_corr.mean(axis=2)
        if MATPLOTLIBV1_0:
            interpolation = 'none'
        else:
            interpolation = 'nearest'
        if data_type == 'amp':
            self.image = self.imshow(10*numpy.log10(numpy.abs(self.bf.cc)),
                                     cmap=cm.spectral, interpolation=interpolation,
                                     vmin=vmin, vmax=vmax,
                                     **kwargs)
        else:
            self.image = self.imshow(numpy.angle(self.bf.cc),
                                     cmap=cm.spectral, interpolation=interpolation,
                                     vmin=vmin, vmax=vmax,
                                     **kwargs)            
        ax, kw = self.plotobj._get_current_axes()
        if MATPLOTLIBV1_0:
            datacursor(self.image, display='single',bbox=dict(fc='white'),
                       arrowprops=dict(arrowstyle='simple', fc='white', alpha=0.5),
                       formatter="x: {x:.0f}\ny: {y:.0f}\nz: {z:.2f}".format)
        def format_coord(x, y):
            if data_type == 'amp':
                z = (10*numpy.log10(numpy.abs(self.bf.cc)))[x, y]
            else:
                z = (numpy.angle(self.bf.cc))[x, y]
            return 'x=%.1f, y=%.1f, z=%.2f' % (x, y, z)


        ax, kw = self.plotobj._get_current_axes()
        ax.format_coord = format_coord
        # self.set_subplot_title(title)
        if title is None:
            title = "%s" % self.bf.basename        
        self.set_subplot_title(title)
        if colorbar:
            self.colorbar()

def omni_view(reds,vis,pol,int=10,chan=500,norm=False,cursor=True,save=None,colors=None,symbols=None, ex_ants=[]):
    if not colors:
        colors = ["#006BA4", "#FF7F0E", "#2CA02C", "#D61D28", "#9467BD", "#8C564B", "#E377C2", "#7F7F7F", "#BCBD22", "#17BECF"]
    if not symbols: 
        symbols = ["o", "v", "^", "<", ">", "*"]
    points = []
    sym = []
    col = []
    bl = []
    ngps = len(reds)
    if save:
        plt.clf()
        plt.cla()
    for i,gp in enumerate(reds):
        c = colors[i%len(colors)]
        s = symbols[i/len(colors)]
        for r in gp:
            if np.any([ant in r for ant in ex_ants]): continue
            try:
                points.append(vis[r][pol][int,chan])
                bl.append(r)
            except(KeyError):
                points.append(np.conj(vis[r[::-1]][pol][int,chan]))
                bl.append(r[::-1])
            sym.append(s)
            col.append(c)
    points = np.array(points)
    max_x=0
    max_y=0
    ax = plt.subplots(111)
    for i,pt in enumerate(points):
        if norm:
            ax.scatter(pt.real/np.abs(pt), pt.imag/np.abs(pt), c=col[i], marker=sym[i], s=50, label='{}'.format(bl[i]))
        else:        
            ax.scatter(pt.real, pt.imag, c=col[i], marker=sym[i], s=50, label='{}'.format(bl[i]))
            if np.abs(pt.real) > max_x: max_x = np.abs(pt.real)
            if np.abs(pt.imag) > max_y: max_y = np.abs(pt.imag)
            
    if norm:         
        plt.xlim(-1,1)
        plt.ylim(-1,1)
    else: 
        plt.xlim(-max_x-.1*max_x,max_x+.1*max_x)
        plt.ylim(-max_y-.1*max_y,max_y+.1*max_y)
    plt.ylabel('imag(V)')
    plt.xlabel('real(V)')
    if cursor:
        from mpldatacursor import datacursor
        datacursor(formatter='{label}'.format)
    if save:
        plt.savefig(save)
    return None

def lifetimefitting(icorr_mean_list, str_yr_list=[0., 0., 0.]):
    from scipy import stats
    from scipy.optimize import curve_fit

    # load data
    suffix = rate2suffix(icorr_mean_list)
    filename = 'pfhistory_str_'
    for ti in str_yr_list:
        filename = filename + str(int(ti)) + '_'
    datapath = os.path.join(os.path.abspath('./'), 'data')
    filename = filename+suffix+'.npz'
    datafile = os.path.join(datapath,filename)

    if os.path.isfile(datafile) is False:
        print 'no data available, execute Life_Cycle_History.py with \
                icorr_mean_list={} and str_yr_list={} fist'.format(icorr_mean_list,
                        str_yr_list)
        sys.exit(1)
    else:
        pfhistory = np.load(datafile)
        time_array = pfhistory['time']
        pf_sys = pfhistory['system']
        # lifetime fitting
        # candidate cdfs
        def fitexpon(xdata, *params):
            lbd = params[0]
            return stats.expon.cdf(xdata, scale=1./lbd)
        def fitweibull(xdata, *params):
            k = params[0]    # shape param k in Weibull wiki
            lbd = params[1]    # scale param lbd in Weibull wiki
            return stats.weibull_min.cdf(xdata, k, scale=lbd)
        def fitgamma(xdata, *params):
            k=params[0]
            theta = params[1]
            return stats.gamma.cdf(xdata, k, scale=theta)
        poptExpon, pcovExpon = curve_fit(fitexpon, time_array, pf_sys, p0=[1.], bounds=(0.,np.inf))
        poptWbl, pcovWbl = curve_fit(fitweibull, time_array, pf_sys, p0=[1.,5.], bounds=([0.,0.],[np.inf, np.inf]))
        poptGamma, pcovGamma = curve_fit(fitgamma, time_array, pf_sys, p0=[1.,1.], bounds=([0.,0.],[np.inf,np.inf]))

    plt.ion()
    plt.figure()
    plt.semilogy(time_array, pf_sys, 'o', label='$T_f$ data')
    plt.semilogy(time_array, fitexpon(time_array, poptExpon[0]), ls='--', label='Exponential')
    plt.semilogy(time_array, fitweibull(time_array,poptWbl[0],poptWbl[1]), ls='-', label='Weibull')
    plt.semilogy(time_array, fitgamma(time_array,poptGamma[0],poptGamma[1]), ls=':', label='Gamma')
    plt.xlabel('Time (year)')
    plt.ylabel('Failure probability')
    #plt.legend(loc='lower right', fontsize=9)

    datacursor(formatter='{label}'.format,display='multiple', draggable=True,
            bbox=None, fontsize=9,
            arrowprops=dict(arrowstyle='-|>', connectionstyle='arc3', facecolor='k'))

def Graph_GFS_write():
    print ('inside Graph_GFS graph')
    Plot3 = np.loadtxt('Merge_write.txt')
    Plot4 = np.loadtxt('Merge_read.txt')
    fig = plt.figure()
    pl.subplot(311)
    
    pl.title('Write :: Chunk_and_Block Size vs Speed')    
    pl.xlabel('Chunk_and_Block Size in MB')
    pl.ylabel('Speed in MBps')
    #pl.xlim(0,10)
    
    x = Plot3[:,1]
    y = Plot3[:,3]
    yerr = Plot3[:,4]
    xticks = Plot3[:,1]
    #plt.xticks(xticks)
    plt.xscale('log')
    #plt.yscale('log')
    #plt.xticks(xticks)
    pl.errorbar(x,y,yerr,ecolor='b')
    #pl.plot(x,y, 'b-', marker ='o',)
    datacursor(display='multiple', draggable=True)
    #cursor = Dcursor.FollowDotCursor(ax, x, y)
    #plt.show()
    #pl.show()

    
    pl.subplot(313)    
    pl.title('Read :: Chunk_and_Block Size vs Speed')    
    pl.xlabel('Chunk_and_Block Size in MB')
    pl.ylabel('Speed in MBps')
    #pl.xlim(0,10)
    
    x = Plot4[:,1]
    y = Plot4[:,3]
    yerr = Plot4[:,4]
    xticks = Plot4[:,1]
    pl.xticks(xticks)
    plt.xscale('log')
    pl.errorbar(x,y,yerr,ecolor='b')
    datacursor(display='multiple', draggable=True)
    #pl.plot(x,y, 'b-', marker = 'o')
    #pl.show()
    #Dcursor2.DataCursor([write,read])
    pl.show()
    pl.savefig('GFS_Reading.ps')
    pl.savefig('GFS_Reading.png')
    pl.savefig('GFS_Reading.pdf')
    print ('endddd')
    
    plt.close(fig)

def plot_avg_time_locked_data(timeLockedData, timeAxis, subplot=None, timeToPlot=None, remove_channels=None, picker=None, labels=False, figure_id=0, figure=None):
    if timeToPlot==None:
        if np.size(np.shape(timeLockedData)) > 1:
            samplesToPlot = [0, np.shape(timeLockedData)[1]]
        else:
            samplesToPlot = [0, np.shape(timeLockedData)[0]]
    else:
        Freq = len(timeAxis)/(timeAxis[-1]-timeAxis[0])
        startingTimeDif = timeToPlot[0] - timeAxis[0]
        endingTimeDif = timeToPlot[1] - timeAxis[0]
        if startingTimeDif < 0:
            raise ArithmeticError("The starting time to plot must be after the starting time of the trial")
        if endingTimeDif < 0:
            raise ArithmeticError("The end time to plot must be after the starting time of the trial")
        samplesToPlot = [startingTimeDif*Freq, endingTimeDif*Freq]

    if figure is None:
        fig = plt.figure(figure_id)
    else:
        fig = figure

    if subplot is not None:
        ax = fig.add_subplot(subplot)
    else:
        ax = fig.add_subplot(111)

    if picker:
        def on_pick(event):
            event.artist.set_visible(not event.artist.get_visible())
            print(ax.lines.index(event.artist))
            fig.canvas.draw()
        fig.canvas.callbacks.connect('pick_event', on_pick)


    if remove_channels is not None:
        timeLockedData[remove_channels, :] = float('nan')
    if np.size(np.shape(timeLockedData)) > 1:
        lines = ax.plot(timeAxis[samplesToPlot[0]:samplesToPlot[1]], np.transpose(timeLockedData[:, samplesToPlot[0]:samplesToPlot[1]]), picker=picker)
    else:
        lines = ax.plot(timeAxis[samplesToPlot[0]:samplesToPlot[1]], timeLockedData[samplesToPlot[0]:samplesToPlot[1]], picker=picker)

    if labels:
        datacursor(hover=True)
        for i in np.arange(0,len(lines)):
            lines[i].set_label(str(i))

    fig.add_subplot(ax)

    plt.show()
    return ax

def plot_iv(data,range,res,symbol):
    fig=plt.figure()
    ax1 = fig.add_subplot(111)
    ax2 = ax1.twiny()
    act,=ax1.plot(data.DELTA,data.VOLATILITY,'r*',label='Actual Smile')
    fit,=ax1.plot(range,res,'g',label='Fitted Smile')
    ax1.set_xlabel('DELTA')
    ax1.set_ylabel('VOLATILITY')
    #ax2.plot(data.STRIKE_PR,data.VOLATILITY,'b*',label='Actual Smile')
    ax2.set_xticks(data.DELTA)
    ax2.set_xticklabels(data.STRIKE_PR,rotation=90,fontsize=12)
    
    ax2.set_xlabel('STRIKE')
    #plt.show()
    datacursor()
    plt.savefig("%s_VOL.jpg"%symbol)

    def datacursor(self, widget, unit=None, labname='Label'):
        ''' Plot datacursors (useful when the number of lines gets confusing)

        Parameters
        ----------

        unit and labname: 
            to customize the info box'''

        # Clean previous cursors  (prevent overlaps with remaining cursors)
        while len(self.cursors) > 0:
            dc = self.cursors.pop()
            dc.hide().disable()

        if self.actionDatacursor.isChecked() and CURSOR_AVAIL:

            def formatter(x=None, y=None, z=None, s=None, label=None, **kwargs):

                ax = kwargs['event'].mouseevent.inaxes

                output = []
                output.append(u't: {0:0.3e} s'.format(x))
                output.append(u'y: {0:0.3e} {1}'.format(y, unit))

                for key, val in zip(['z', 's'], [z, s]):
                    if val is not None:
                        try:
                            output.append(
                                u'{key}: {val:0.3e}'.format(key=key, val=val))
                        except ValueError:
                            # X & Y will be strings at this point.
                            # For masked arrays, etc, "z" and s values may be a
                            # string
                            output.append(
                                u'{key}: {val}'.format(key=key, val=val))

                # label may be None or an empty string (for an un-labeled AxesImage)...
                # Un-labeled Line2D's will have labels that start with an
                # underscore
                if label and not label.startswith('_'):
                    output.append(u'{0}: {1}'.format(labname, label))

                if kwargs.get(u'point_label', None) is not None:
                    output.append(
                        u'Point: ' + u', '.join(kwargs['point_label']))

                return u'\n'.join(output)

            for ax in widget.axes:
                if not ax.cursorlines is None:
                    self.cursors.append(mpldatacursor.datacursor(
                        ax.cursorlines, hover=True, size=10, color='k',
                        bbox=dict(fc='white', alpha=0.9),
                        formatter=formatter))
        return None

def front_deprecated(icorr_mean_list):
    suffix = rate2suffix(icorr_mean_list)
    # load data
    datapath = os.path.join(os.path.abspath('./'), 'data')
    filename = 'popdata_'+suffix+'.npz'
    datafile = os.path.join(datapath,filename)
    if os.path.isfile(datafile) is False:
        print 'no data available, execute Life_Cycle_Optimization.py with \
                icorr_mean_list={} fist'.format(icorr_mean_list)
        sys.exit(1)
    else:
        popdata = np.load(datafile)
        # all pop is the same as pop
        allpop = popdata['allpop']
        allfits = popdata['allfits']
        front = popdata['front']
        frontfits = popdata['frontfits']
        pop = popdata['pop']
        popfits = popdata['popfits']

    plt.ion()
    plt.figure()

    ##plt.semilogx(np.array(frontfits)[:,0], np.array(frontfits)[:,1], 'bo', markeredgecolor='b')
    for ind, popfit in zip(front, frontfits):
        plt.semilogx(popfit[0], popfit[1], 'bo',
                label=u'flexure: {:d}, shear: {:d}, deck: {:d}'.format(ind[0], ind[1], ind[2]))

    plt.ylim((-1,np.max(popfits)*1.01))
    ax = plt.gca()
    ax.ticklabel_format(axis='y', style='sci', scilimits=(-3,3))

    plt.xlabel(u'Failure probability (log scale)', fontsize=12)
    plt.ylabel(u'Strengthening cost (mm\\textsuperscript{3})', fontsize=12)

    annotate_text = u'P={x:.2e}, C={y:.2e}\n {{ {label} }}'
    datacursor(formatter=annotate_text.format,display='multiple', draggable=True,
            bbox=None, fontsize=12,
            arrowprops=dict(arrowstyle='-|>', connectionstyle='arc3', facecolor='k'))

    pause = raw_input('press any key after annotation...')