quickplot/quickplot/quickplot.py at master · baxen/quickplot

This repository was archived by the owner on Jun 8, 2020. It is now read-only.
805 lines (672 loc) · 28.9 KB
#!/usr/bin/env python
An extension to rootpy for convenience and speed in plotting.
Matplotlib is the backend, because it is more powerful than ROOT.
But, for convenience and familiarity, ROOT classes are used to
hold and analyze data, through rootpy.
This module provides quick setup functions for matplotlib plots
and an interface for good color choices.
import glob
import json
import types
import shlex
import argparse
import matplotlib
matplotlib.use('PDF')
import retrieve
from helpers import *
# Right now this uses an adaptable system:
#   Full latex support
#   Full vector rendering
# I would like to upgrade to pgf based output if they work out bugs
custom_preamble = {
    'font.size'          : 20, # default is too small
    'figure.autolayout'  : True,
    'text.usetex'        : True,
    'axes.edgecolor'     : '#737373',
    'grid.color'         : '#B3B3B3',
    'grid.linestyle'     : '--',
    'grid.linewidth'     : 2,
    'ytick.major.size'   : 8,
    'ytick.minor.size'   : 4,
    'xtick.major.size'   : 8,
    'xtick.minor.size'   : 4,
#    'lines.linewidth'    : 2.0,
#    'lines.markersize'   : 16.0,
    'legend.numpoints'   : 1
matplotlib.rcParams.update(custom_preamble)
fontdir = os.path.join(os.path.dirname(os.path.realpath(__file__)),"fonts/")
pgf_update = {
    'font.family'       : 'serif', # actually uses latex main font from below
    'pgf.texsystem'     : 'xelatex',
    'pgf.rcfonts'       : False,   # don't setup fonts from rc parameters
    'pgf.preamble'      : [
        r'\usepackage{mathspec}',
        '\setallmainfonts(Digits,Latin,Greek)[Path = {0}, BoldFont={{HelveticaBold}}, ItalicFont={{HelveticaOblique}}, BoldItalicFont={{HelveticaBoldOblique}}]{{Helvetica}}'.format(fontdir)
import matplotlib.pyplot as plt
import matplotlib.gridspec as gridspec
import rootpy.plotting.root2matplotlib as rplt
import matplotlib.ticker as ticker
import matplotlib.offsetbox as offsetbox
import matplotlib.patches as mpatches
import rootpy.ROOT as ROOT
from rootpy.plotting.style import set_style
from rootpy.plotting.utils import get_limits, draw
from mpl_toolkits.axes_grid.anchored_artists import AnchoredText
from matplotlib.colors import LogNorm
from rootpy.plotting.style import get_style, set_style
from rootpy.io import root_open
from rootpy.tree import Tree, TreeChain
from rootpy.plotting import Hist
from itertools import izip, chain, product
set_style('ATLAS',mpl=True)
# Update for fonts
matplotlib.rcParams['text.usetex'] = True
matplotlib.rcParams['text.latex.preamble'] =  [r'\usepackage{tgheros}',    # helvetica font
                                               r'\usepackage{sansmath}',   # math-font matching  helvetica
                                               r'\sansmath'                # actually tell tex to use it!
                                               r'\usepackage{siunitx}',    # micro symbols
                                               r'\sisetup{detect-all}',    # force siunitx to use the fonts
# ----------------------------------------
# Utility Functions and Variables
# ----------------------------------------
canvas_json = 'canvas.json'
SIM_INTERNAL = r'$\textbf{\textit{ATLAS} Simulation Internal}$'
INTERNAL = r'$\textbf{\textit{ATLAS}}$'
PRELIM = r'$\textbf{\textit{ATLAS} Preliminary}$'
#PRELIM = r'$\textbf{\textit{ATLAS}}$'
def is_listy(x):
    return any(isinstance(x, t) for t in [types.TupleType,types.ListType])
def make_iterable(x):
    if not is_listy(x):
        return [x]
    return x
def set_font(font="helvetica"):
    fonts = {"helvetica": '\setallmainfonts(Digits,Latin,Greek)[Path = {0}, BoldFont={{HelveticaBold}}, ItalicFont={{HelveticaOblique}}, BoldItalicFont={{HelveticaBoldOblique}}]{{Helvetica}}'.format(fontdir),
             "lato": '\setallmainfonts(Digits,Latin,Greek)[Path = {0}, BoldFont={{Lato-Bold}}, ItalicFont={{Lato-LightItalic}}, BoldItalicFont={{Lato-BoldItalic}}]{{Lato-Light}}'.format(fontdir)}
    if font.lower() not in fonts:
        print "Attempting to use unsupported font", font, "; not enabled."
        return
    matplotlib.rcParams.update({'pgf.preamble':[r'\usepackage{mathspec}', fonts[font.lower()]]}) 
# ----------------------------------------
# Utility Classes
# ----------------------------------------
class PlotError(Exception):
# ----------------------------------------
# Plot Layouts
# ----------------------------------------
def build_axes(ratio=False, double=False, **kwargs):
    Builds a figure and axes with either one or two major axes
    and an optional ratio axes.
    kwargs are ignored
    # Two axes and ratio
    if ratio and double:
        fig = plt.figure(figsize=(8.0, 8.0))
        gs = gridspec.GridSpec(3,1,height_ratios=[2,2,1], hspace=0.0, wspace=0.0)
        ax = plt.subplot(gs[0])
        axes = [ax, plt.subplot(gs[1], sharex=ax), plt.subplot(gs[2], sharex=ax)]
    # Single axes and ratio
    elif ratio:
        fig = plt.figure(figsize=(9.0, 6.0))
        gs = gridspec.GridSpec(2,1,height_ratios=[3,1], hspace=0.0, wspace=0.0)
        ax = plt.subplot(gs[0])
        axes = [ax, plt.subplot(gs[1], sharex=ax)]
    # Single axis
        if 'customization' in kwargs and kwargs['customization'] == 'jes_cor':
            fig = plt.figure(figsize=(8.0,7.0))
        else:
            fig = plt.figure()
        axes = [fig.add_subplot(111)]
    for ax in axes:
        ax.ratio = False
        ax.primary = False
        ax.secondary = False
        ax.extra_handles = []
    axes[0].primary = True
    if double: axes[-2].secondary = True
    if ratio: axes[-1].ratio = True
    return fig, axes
def process_hists(hists, ratio=False, double=False, rindex=0, **kwargs):
    Process and group histograms to match the setup created by build_axes.
    kwargs are ignored
    # Processing of hists prior to ratio
    for j,hist in enumerate(hists):
        if not_empty(kwargs, 'xscale'):
            hists[j] = xscale(hist, kwargs['xscale'])
        if not_empty(kwargs, 'xefficiency') or not_empty(kwargs,'xinefficiency'):
            index = int(kwargs['xefficiency']) if not_empty(kwargs, 'xefficiency') else int(kwargs['xinefficiency'])
            hists[j] = running_integral(hist, neg=index<0)
        if not_empty(kwargs, 'overflow') and kwargs['overflow']:
            # Add overflow to last bin for plotting
            hist[hist.bins_range()[-1]] += hist[hist.bins_range(overflow=True)[-1]]
    if not_empty(kwargs, 'errorbands') and kwargs['errorbands']:
        for i in xrange(1,len(hists)):
            htop = hists[0] + hists[i]
            htop.decorate(hists[i])
            htop.title = hists[i].title
            hbot = hists[0] - hists[i]
            hbot.decorate(hists[i])
            hbot.title = "" # make sure only one gets a legend entry
            hists[i] = htop
            hists.append(hbot)
    # If two axes, half go on each in order
    if double:
        hgroups = [hists[:len(hists)/2], hists[len(hists)/2:]]
        hgroups = [hists]
    # If ratio, create a ratio from each hist by dividing by the histogram at rindex for each group
    if ratio=='paired':
        hgroups.append(sum(([histify(group[i])/histify(group[i-1]) for i in xrange(1,len(group),2)] for group in hgroups), []))
    elif ratio=='sb':
        hgroups.append(sum(([histify(h)/sqrt_hist(histify(group[rindex])) for i,h in enumerate(group) if i != rindex] for group in hgroups), []))
    elif ratio:
        hgroups.append(sum(([histify(h)/histify(group[rindex]) for i,h in enumerate(group) if i != rindex] for group in hgroups), []))
    # Processing of hists after ratio
    for i, hgroup in enumerate(hgroups):
        if ratio and i == len(hgroups) - 1:
            continue
        for j,hist in enumerate(hgroup):
            if not_empty(kwargs, 'xefficiency') or not_empty(kwargs,'xinefficiency'):
                hgroups[i][j] = efficiency_divide(hist,flat_hist(hist,index))
            if not_empty(kwargs, 'xinefficiency'):
                hgroups[i][j] = hgroups[i][j]*-1 + 1
    return hgroups
# ----------------------------------------
# Plot Setup
# ----------------------------------------
def set_scales(axes, logx=False, logy=False, **kwargs):
    Set log scales on all axes.
    Does not set logy on an axis marked ratio
    kwargs are ignored for convenience in passing
    for ax in axes:
        if logx:
            ax.set_xscale('log')
        if logy and (not hasattr(ax,'ratio') or not ax.ratio):
            ax.set_yscale('log')
def normalize(hists, rebin='', norm='', first_norm='', bin_norm='', index_norm='', range_norm='', overflow=False, **kwargs):
    for hist in hists:
        if rebin:
            hist.rebin(int(rebin))
        if norm: 
            hist.Scale(float(norm)/hist.integral(overflow=overflow))
        if first_norm:
            hist.Scale(1.0/hist[1])
        if index_norm: 
            hist.Scale(hist[index_norm].Integral()//hist.integral(overflow=overflow))
        if range_norm:
            low, high = [float(s) for s in range_norm.split(',')]
            hist.Scale(hists[0].integral(hist.find_bin(low), hist.find_bin(high))/hist.integral(hist.find_bin(low), hist.find_bin(high)))
        if bin_norm:
            for i in hist.bins_range():
                hist.SetBinContent(i, hist.GetBinContent(i)/hist.GetBinWidth(i))
                hist.SetBinError(i, hist.GetBinError(i)/hist.GetBinWidth(i)) 
def setup_axes(*axes, **kwargs):
    """ Setup an axis with standard options. 
    Just a shortcut for calling a bunch of axes member functions.
    Keyword Arguments:
    title -- Text to be placed on top center of axes.
    xmin -- Set minimum of x-axis.
    xmax -- Set maximum of x-axis.
    ymin -- Set minimum of y-axis.
    ymax -- Set maximum of y-axis.
    xlabel -- Label for x-axis.
    ylabel -- Label for y-axis.
    xticks -- String holding space-separated tick locations
    xticklabels -- String hold space-separated tick labels
    xtickrot -- Rotation for tick labels
    # Handle some shortcut syntaxes
    if not_empty(kwargs, 'xticks') and not not_empty(kwargs, 'xticklabels'):
        kwargs['xticklabels'] = kwargs['xticks']
    if not_empty(kwargs, 'yticks') and not not_empty(kwargs, 'yticklabels'):
        kwargs['yticklabels'] = kwargs['yticks']
    if not_empty(kwargs, 'xticks'):
        try:
            kwargs['xticks'] = [float(x) for x in kwargs['xticks'].split()]
        except (AttributeError):
            pass
    if not_empty(kwargs, 'xticklabels'):
        try:
            kwargs['xticklabels'] = kwargs['xticklabels'].split()
        except (AttributeError, KeyError):
            pass
    if not_empty(kwargs, 'yticks'):
        try:
            kwargs['yticks'] = [float(x) for x in kwargs['yticks'].split()]
        except (AttributeError, KeyError):
            pass
    if not_empty(kwargs, 'yticklabels'):
        try:
            kwargs['yticklabels'] = kwargs['yticklabels'].split()
        except (AttributeError, KeyError):
            pass
    if not_empty(kwargs, 'rticks'):
        try:
            kwargs['rticks'] = [float(x) for x in kwargs['rticks'].split()]
        except (AttributeError, KeyError):
            pass
    if not_empty(kwargs, 'rticklabels'):
        try:
            kwargs['rticklabels'] = kwargs['rticklabels'].split()
        except (AttributeError, KeyError):
            pass
    # Setup each axis
    for ax in axes:
        if not_empty(kwargs, 'title') and ax.primary:
            # Only put title on the top set of axes
            ax.set_title(kwargs['title'])
        if ax.ratio:
            # Use ratio y-axis values
            if not_empty(kwargs, 'rmin'):
                ax.set_ylim(bottom=kwargs['rmin'])
            if not_empty(kwargs, 'rmax'):
                ax.set_ylim(top=kwargs['rmax'])
            if not_empty(kwargs, 'rlabel'):
                ax.set_ylabel(kwargs['rlabel'], y=1, ha='right')
        else:
            # Use normal y-axis values
            if not_empty(kwargs, 'ymin'):
                ax.set_ylim(bottom=float(kwargs['ymin']))
            if not_empty(kwargs, 'ymax'):
                ax.set_ylim(top=float(kwargs['ymax']))
            if not_empty(kwargs, 'ylabel'):
                ax.set_ylabel(kwargs['ylabel'], y=1, ha='right')
            if not_empty(kwargs, 'yticks'):
                ax.set_yticks(kwargs['yticks'])
            if not_empty(kwargs, 'yticklabels'):
                if 'ytickrot' in kwargs:
                    ax.set_yticklabels(kwargs['yticklabels'], rotation=kwargs['ytickrot'])
                else:
                    ax.set_yticklabels(kwargs['yticklabels'])
        if not_empty(kwargs, 'xmin'):
            ax.set_xlim(left=float(kwargs['xmin']))
        if not_empty(kwargs, 'xmax'):
            ax.set_xlim(right=float(kwargs['xmax']))
        if not_empty(kwargs, 'xlabel'):
            ax.set_xlabel(kwargs['xlabel'], x=1, ha='right')        
        if not_empty(kwargs, 'xticks'):
            ax.set_xticks(kwargs['xticks'])
        if not_empty(kwargs, 'xticklabels'):
            if 'xtickrot' in kwargs:
                ax.set_xticklabels(kwargs['xticklabels'], rotation=kwargs['xtickrot'])
            else:
                ax.set_xticklabels(kwargs['xticklabels'])
        if ax.get_xscale() != 'log':
            ax.xaxis.set_minor_locator(ticker.AutoMinorLocator())
        if ax.get_yscale() != 'log':
            if 'yticks' not in kwargs or not kwargs['yticks']:
                ax.yaxis.set_major_locator(ticker.MaxNLocator(5, prune='upper' if (not ax.ratio and not ax.primary) else None))
            ax.yaxis.set_minor_locator(ticker.AutoMinorLocator())
        ax.yaxis.set_label_coords(-0.15,1)
        for spine in ax.spines.values(): 
            spine.set_zorder(100)
    # Clean up for ratio plots
    if axes[-1].ratio:
        for ax in axes[:-1]:
            for tick in ax.get_xticklabels():
                tick.set_visible(False)
            ax.set_xlabel("")
        axes[-1].yaxis.set_major_locator(ticker.MaxNLocator(4, prune='upper'))
        # Show a line at 1.0 if it's in range
        if axes[-1].get_ylim()[0] < 1.0 < axes[-1].get_ylim()[1]:
            axes[-1].plot(axes[-1].get_xlim(), [1.0,1.0], color='#B3B3B3', linewidth=2.0,linestyle='--')
        axes[-1].yaxis.grid(True) # Show lines on tick marks
        if not_empty(kwargs, 'ratio_sb') and kwargs['ratio_sb']:
            axes[-1].yaxis.get_major_formatter().set_powerlimits((-1,1))
        if not_empty(kwargs, 'rticks'):
            axes[-1].set_yticks(kwargs['rticks'])
        if not_empty(kwargs, 'rticklabels'):
            axes[-1].set_yticklabels(kwargs['rticklabels'])
def text(ax, secondary=False, text=None, text_location="upper left", text_secondary=None, text_location_secondary="upper left", **kwargs):
    if secondary:
        text = text_secondary
        text_location = text_location_secondary
    if text:
        locations = {'best': 0,
                     'upper right':1,
                     'upper left':2,
                     'lower left':3,
                     'lower right':4,
                     'right':5,
                     'center left':6,
                     'center right':7,
                     'lower center':8,
                     'upper center':9,
                     'center':10}
        bbox = {}
        if text_location in locations:
            text_location = locations[text_location]
        elif text_location in locations.values():
            pass
        elif '(' in text_location:
            # Specify x,y coordinates for upper left corner
            bbox['bbox_to_anchor'] = tuple(float(x) for x in text_location[1:-1].split(','))
            bbox['bbox_transform'] = ax.transAxes
            text_location = 2
        text = text.replace("SIM_INTERNAL", SIM_INTERNAL)
        text = text.replace("INTERNAL", INTERNAL)
        text = text.replace("PRELIM", PRELIM)
        text = text.replace(";", "\n")
        anchored = AnchoredText(text, prop=dict(size=18), loc=text_location, frameon=False, **bbox)
        ax.add_artist(anchored)
def legend(ax, hists, legend_loc='best', legend_text=None, **kwargs):
    # Don't create a legend if there are not labels
    if not any(h.title for h in hists):
        return
    # Don't make a legend on the ratio plot
    if ax.ratio:
        return
    bbox = {}
    # Specify x,y coordinates for upper left corner
    if '(' in legend_loc:
        bbox['bbox_to_anchor'] = tuple(float(x) for x in text_location[1:-1].split(','))
        bbox['bbox_transform'] = ax.transAxes
        legend_loc = 2
    # Shortcut for common labels
    if legend_text == "INTERNAL":
        legend_text = INTERNAL
    if legend_text == "PRELIM":
        legend_text = PRELIM
    handles, labels = ax.get_legend_handles_labels()
    handles = handles[::-1]
    labels = labels[::-1]
    for extra in ax.extra_handles:
        handles.append(extra)
        labels.append(extra.get_label())
    legend = ax.legend(handles, labels, frameon=False, loc=legend_loc if legend_loc else 'best', prop={'size':18}, labelspacing=0.25, ncol=1 if len(hists) < 6 else 2, **bbox)
    # Add extra text above legend.
    if legend_text and legend_text is not "none":
        for sub in legend_text.split(";")[::-1]:
            txt=offsetbox.TextArea(sub, {'size':18}) 
            box = legend._legend_box 
            box.get_children().insert(0,txt) 
            box.set_figure(box.figure) 
    return legend
def customize(kwargs, hists):
        # Handle customizations
    if  kwargs.get('customization') == 'jes_cor':
        kwargs['xticks'] = '1 8 16 20 27 35'
        kwargs['yticks'] = '1 8 16 20 27 35'
        kwargs['xticklabels'] = ['30', '300\n' +r'$|\eta| < 0.6$', '2500', '30', '300\n' +r'$0.6 < |\eta| < 1.1$', '2500']
        kwargs['yticklabels'] = ['30', r'\begin{center}$|\eta| < 0.6$\\300\end{center}', '2500', '30', r'\begin{center}$0.6 < |\eta| < 1.1$\\300\end{center}', '2500']
        kwargs['ytickrot'] = 'vertical'
    if kwargs.get('customization') == 'fig16':
        hists[0][-4] = 0.0
    if kwargs.get('customization') == 'systematics':
        # hists[-2] is the central value, hists[-1] is an upper systematic
        for bin1,bin2 in zip(hists[-2], hists[-1]): 
            bin2.error = abs(bin2.value - bin1.value)
            bin2.value = bin1.value
    if kwargs.get('customization') == 'halfshift':
        for i,h in enumerate(hists):
            htmp = Hist([x+0.5 for x in h.xedges()])
            htmp.decorate(h)
            htmp.drawstyle = h.drawstyle
            htmp.title = h.title
            for bin in h.bins_range():
                htmp[bin] = h[bin]
            hists[i] = htmp
    if kwargs.get('customization') == 'ztrunc':
        for i,h in enumerate(hists):
            htmp = Hist([0] + list(h.xedges())[1:])
            htmp.decorate(h)
            htmp.drawstyle = h.drawstyle
            htmp.title = h.title
            for bin in h.bins_range():
                htmp[bin] = h[bin]
            hists[i] = htmp
def plot(draw, name, hists, **kwargs):
    Performs basic canvas setup and draws hists with draw, saving the plot to name.
    kwargs are passed around to all subfunctions
    customize(kwargs, hists)
    fig, axes = build_axes(**kwargs)
    # Set log scales first thing, because some functions check for log scales
    set_scales(axes, **kwargs)
    # Normalize input hists if specified
    normalize(hists, **kwargs)
    # Split histograms into group for each axes
    hgroups = process_hists(hists, **kwargs)
    # It might be necessary for some plotting functions to set approx ylims first?
    # Come back to this...
    if ymin is None or ymin == "":
        l,h = min(chain(*[list(h.y()) for h in hists])), max(chain(*[list(h.y()) for h in hists]))
        if logy: ymin = 0.8*l
        else: ymin = l - .1*(h-l)
    ax.set_ylim(bottom=ymin)
    if ymax is None or ymax == "":
        l,h = min(chain(*[list(h.y()) for h in hists])), max(chain(*[list(h.y()) for h in hists]))        
        if logy: ymax = h/0.8
        else: ymax = h + .1*(h-l)
    ax.set_ylim(top=ymax)
    for ax, group in izip(axes, hgroups):
        draw(ax, group, **kwargs)
        if ax.primary: text(ax, **kwargs)
        if ax.secondary: text(ax, secondary=True, **kwargs)
        if draw != draw2d:
            legend(ax, group, **kwargs)
    # Setup axes last so that nothing gets overwritten
    setup_axes(*axes, **kwargs)
    if any(x in name for x in ["png","jpg"]):
        fig.savefig(name, dpi=500)
        fig.savefig(name)
    plt.close(fig)
# ----------------------------------------
# Drawing Functions
# ----------------------------------------
def errorbar(ax, hists, **kwargs):
    Plot histograms or graphs as points with errors bars.
    kwargs are passed to errorbar
    # Override some defaults
    defaults = {"snap":False,
                "emptybins":False,
                "capsize":0}
    if "logy" in defaults:
        defaults.pop("logy")
    rplt.errorbar(hists, axes=ax, **defaults)
def band(ax, hists, **kwargs):
    Fill the region between bottom and top, for each pair.
    kwargs are passed to fill_between
    # Override some defaults
    #rplt.hist(hists, axes=ax, **kwargs)
    defaults = {"linewidth":0.0, "alpha":0.8}
    for h in make_iterable(hists):
        b = h.Clone()
        t = h.Clone()
        b.title = ""
        t.title = ""
        for i in t.bins_range():
            b.SetBinContent(i,b.GetBinContent(i) - b.GetBinError(i))
            t.SetBinContent(i,t.GetBinContent(i) + t.GetBinError(i))
        rplt._set_defaults(h, defaults, ['common', 'line']) 
        if defaults.get('color') is None:
            defaults['color'] = h.GetLineColor('mpl')
        rplt.fill_between(b, t, axes=ax, **defaults)
        # Add patch for the legend
        if h.title:
            ax.extra_handles.append(mpatches.Patch(color=h.GetLineColor('mpl'), alpha=defaults["alpha"], label=h.title))
def _step(ax, h, **kwargs):
    # Make horizontal line segments
    x, y = list(h.xedges()), list(h.y())
    x_pairs = [x[i:i+2] for i in xrange(len(x)-1)]
    y_pairs = [[y[i],y[i]] for i in xrange(len(y))]
    segments = [[x,y] for x,y in izip(x_pairs, y_pairs)]
    segments = sum(segments, [])
    # Borrowed style setting from rootpy
    rplt._set_defaults(h, kwargs, ['common', 'line']) 
    if kwargs.get('color') is None:
        kwargs['color'] = h.GetLineColor('mpl')
    # Only have a label for the first one
    ax.plot(*segments[:2], **kwargs)
    kwargs['label'] = None
    ax.plot(*segments[2:], **kwargs)
def step(ax, hists, **kwargs):
    Plot the histogram as horizontal lines at the bin content
    kwargs are passed to Axes.plot
    hiter = make_iterable(hists)
    for h in hiter:
        if h == hiter[-1] and 'customization' in kwargs and kwargs['customization'] == 'jes_unc':
            print 'Using updated zorder on ', h.title
            _step(ax, h, zorder=50)
        else:
            _step(ax, h)
def steperr(ax, hists, **kwargs):
    step(ax, hists, **kwargs)
    for h in make_iterable(hists):
        eb = rplt.errorbar(hists, axes=ax, fmt='none', capsize=0, xerr=None, label='', elinewidth=h.linewidth)
        eb[-1][0].set_linestyle(h.linestyle) 
def hist(ax, hists, **kwargs):
    defaults = {'fill':None,
                'capsize':0,
                'stacked':False}
    for hist in make_iterable(hists):
        # Need the histify syntax to handle graphs
        rplt.hist(histify(hist), axes=ax, **defaults)
def herr(ax, hists, **kwargs):
    defaults = {'marker':None,
                'capsize':0,
                'stacked':False}
    hist(ax, hists, **defaults)
    for h in make_iterable(hists):
        eb = rplt.errorbar(h, axes=ax, fmt='none', xerr=None, label='', capsize=0, elinewidth=h.linewidth)
        eb[-1][0].set_linestyle(h.linestyle) 
def stack(ax, hists, **kwargs):
    defaults = {'fill':'solid',
                'capsize':0,
                'yerr':None}
    rplt.hist(hists[::-1], axes=ax, **defaults)
def hist2d(ax, hist, logz=False, numbers=False, cmap='Blues', **kwargs):
    defaults = {"colorbar":True}
    if logz:
        defaults["norm"] = LogNorm()
    defaults['cmap'] = cmap
    rplt.hist2d(hist, axes=ax, **defaults)
    if numbers:
        for (x,y),z in izip(product(hist.x(), hist.y()), chain(*hist.z())):
            ax.text(x, y, "{:.3g}".format(z), va='center', ha='center')
_plot_functions = {"errorbar":errorbar,
                   "hist":hist,
                   "herr":herr,
                   "stack":stack,
                   "band":band,
                   "step":step,
                   "steperr":steperr}
def draw1d(ax, hists, **kwargs):
    stacked = [h for h in hists if h.drawstyle == 'stack']
    unstacked = [h for h in hists if h.drawstyle != 'stack']
    if stacked: stack(ax, stacked)
    for h in unstacked:
        if h.drawstyle == '': h.drawstyle = 'errorbar'
        if h.drawstyle not in _plot_functions:
            raise PlotError("Do not recognize drawstyle: " + h.drawstyle)
        pargs = dict(customization=kwargs.pop('customization',None))
        _plot_functions[h.drawstyle](ax, h, **pargs)
def draw2d(ax, hists, logz=False, **kwargs):
    hist2d(ax, hists[0], logz=logz, cmap=hists[0].cmap, **kwargs)
# ----------------------------------------
# ROOT Helpers
# ----------------------------------------
def fit(hists, functions):
    graphs, pars, errors = [], [], []
    if isinstance(functions, types.StringTypes):
        functions = [functions for h in hists]
    for hist,function in zip(hists,functions):
        hist.Fit(function, "RB+")
        func = hist.GetFunction(function)
        graphs.append(ROOT.TH1D(func.GetHistogram()))
        params = func.GetParameters()
        param_errors = func.GetParErrors()
        pars.append(params)
        errors.append(param_errors)
    return graphs, pars, errors
# ----------------------------------------
# Command Line Plotting
# ----------------------------------------
def main(args):
    parser = argparse.ArgumentParser('Create a plot from ntuples, using variable and sample lookup information from spreadsheets. Any additional arguments are taken as overrides for canvas options. (e.g. --title "Title")')
    parser.add_argument('output', nargs="?", help='Name of the plot to save.')
    parser.add_argument('--variables', nargs='*', help='The names of variables to add to the plot. Each variable gets a separate graphic.')
    parser.add_argument('--canvas', help="Canvas to use for the plot, if unspecified will start with a blank canvas.")
    parser.add_argument('--selection', help='Additional global selection to apply for all loaded data.')
    parser.add_argument('--batch', help='Specify a file which contains a set of arguments to run on each line.')
    parser.add_argument('--dump', action='store_true', help='Save all plotted histograms to a root file.')
    parser.add_argument('--internal', action='store_true', help='Use ATLAS Internal instead of ATLAS label.')
    args, extras = parser.parse_known_args(args)
    extras = dict(zip([key.replace('--','') for key in extras[:-1:2]], extras[1::2]))
    if (not args.output or not len(args.variables)) and not args.batch:
        raise PlotError("Must provide an output name and variable for the plot or a batch file to run on.")
    global INTERNAL
    global SIM_INTERNAL
    if args.internal:
        INTERNAL = r'$\textbf{\textit{ATLAS} Internal}$'
        SIM_INTERNAL = r'$\textbf{\textit{ATLAS} Simulation Internal}$'
        INTERNAL = r'$\textbf{\textit{ATLAS}}$'
        SIM_INTERNAL = r'$\textbf{\textit{ATLAS} Simulation}$'
    if args.batch:
        with open(args.batch) as f:
            lines = f.readlines()
        lines = [l.split("#")[0] for l in lines]
        # Maybe ship this out in parallel?
        for next_args in (l for l in lines if l.strip() != ""):
            try:
                main(shlex.split(next_args))
            except BaseException as e:
                print "Plotting failed with ", e
                print "  and arguments", next_args
        return
    hists = retrieve.retrieve_all(args.variables, args.selection)
    if args.dump:
        f = root_open(os.path.splitext(args.output)[0] + ".root", "recreate")
        for h in hists:
            htmp = histify(h)
            htmp.name = htmp.title.replace("/","over").replace(" ","_").replace("$","").replace(">","gt").replace(",","").lower()
            htmp.write()
        f.close()
    if args.canvas:
        with open(canvas_json) as f:    
            canvas = json.load(f)[args.canvas]
        canvas.update(extras)
        plot_args = canvas
        plot_args = extras
    # ----------------------------------------
    # Select Draw Function and Plot
    # ----------------------------------------
    if '2' in hists[0].__class__.__name__:
        draw_function = draw2d
        draw_function = draw1d
    # Create figure
    plot(draw_function, args.output, hists, **plot_args)
if __name__ == "__main__":
    main(sys.argv[1:])
Provide feedback

Saved searches

Use saved searches to filter your results more quickly

FilesExpand file tree

quickplot.py

Latest commit

History

quickplot.py

File metadata and controls
