plotIt

plotIt is a utility to plot ROOT histogram stacks. The input files corresponding to the different contributions and the histograms to fit are configured through a yaml file.

First time setup instructions

git clone -o upstream https://github.com/cp3-llbb/plotIt.git # inside $CMSSW_BASE/src/cp3_llbb
cd plotIt/

# Initialize the git remotes
source firstsetup.sh 
# Within-CMSSW and on ingrid specific install
cms_env # specific to ingrid, aka 'module purge; module load grid/grid_environment_sl6; module load crab/crab3; module load cms/cmssw;'
cmsenv
source setup_for_cms_env.sh
# For a non-CMSSW and non-ingrid install (beware there is no cmsenv at all in this case):
# source setup_sl6_env.sh

# Build externals
cd external
./build-external.sh
# Build the executable itself
cd ..
make -j 4

Outside CMSSW and with a recent version of ROOT, plotIt can als be built with CMake (and an install prefix passed with -DCMAKE_INSTALL_PREFIX, or the exectuble copied from the build directory to install), e.g.

git clone -o upstream https://github.com/cp3-llbb/plotIt.git
mkdir plotit-build
cd plotit-build
cmake ../plotIt
make -j2
cd -

Test run (command line)

# Load the proper environment (if not already done)
source setup_sl6_env.sh
# Create some dumb root files to play with
cd test
root -l -b -q generate_files.C
# Now plot stuff
./../plotIt -o plots/ example.yml
# Go to the plots directory to observe the beautiful plots

Optional command-line arguments

The command above is the minimal way to run plotIt: with a configuration file and output directory. There are also a number of optional arguments that change the behaviour:

• -h: print help for the command-line interface and exit
• -v: verbose output, will print more progress messages, a summary, and the LaTeX yields table (if -y is also specified) to stdout
• -b: print systematics details for each MC process in the summary
• -o: output directory
• -i: input directory (file names are taken relative to this directory, default is the current directory)
• -y: produce a yields table (written to <output>/yields.tex)
• -p: do not produce the plots
• --ignore-scales: ignore (global and per-file) scale parameters for the normalisation
• -e: pass one era to make plots for (otherwise combined plots, summing over all eras specified in the configuration, are made)
• -u: unblind, i.e. ignore any blinded-range in the configuration

Configuration file reference

• Points and Ranges are lists of two numbers: [x,y]
• Positions are lists of four numbers: [x1, y1, x2, y2]
• Some strings are formatted using boost::format. Arguments passed to the formatting are described below.
• Strings used for labels, axis titles, legends, ... can use TLatex
• Line styles and widths, fill styles and marker styles are numbers, see the corresponding TAtt... class references or the lists below for some common values
• Colors are either basic ROOT colors (ie 42) or hexadecimal (alpha)RGB (ie #FFFFCC)

plotIt general ('configuration')

Field	Type	Action	Default
width	number	Width (in pixels) of the plots	800
height	number	Height (in pixels) of the plots	800
experiment	string	Name of the experiment, appearing in bold above the plot	CMS
extra-label	string	Additional text above the plot
root	string	Path to the folder where the files are located	./
scale	number	Global scaling of all non-data entries	1
eras	list of strings	List of data-taking periods to combine	[""]
luminosity	number or map	Integrated luminosity (in pb) used to normalise MC to data. Used by the luminosity label. When using eras: map of { era : luminosity }
luminosity-error	number	Fractional uncertainty on the integrated luminosity. Used for the yields table (uncertainty on MC and data/MC ratio) and the plot's MC & ratio syst. error bands.	0
luminosity-label	string	Luminosity label appearing above the plot. In the string, %lumi% gets replaced by the integrated luminosity in fb.
error-fill-color	color		42
error-fill-style	fill style		3154
fit-line-color	color		46
fit-line-style	line style		1
fit-line-width	line width		1
fit-error-fill-color	color		42
fit-error-fill-style	fill style		1001
fit-n-points	number		1000
ratio-fit-line-color	color		46
ratio-fit-line-style	line style		1
ratio-fit-line-width	line width		1
ratio-fit-error-fill-color	color		42
ratio-fit-error-fill-style	fill style		1001
ratio-fit-n-points	number		1000
blinded-range-fill-color	color		42
blinded-range-fill-style	fill style		1001
y-axis-format	formatted string	Arg. 1: plot's y-axis title; Arg. 2: plot's first bin's width	%1% / %2$.2f
ratio-y-axis	string	Y axis title of the ratio plot	Data / MC
ratio-style	string	Draw option for the ratio histogram	P0
mode	string	tree (fill TH1s from a tree on-the-fly) or hist (retrieve TH1s from files)	hist
tree-name	string	Name of the TTree used to fill the histograms in tree mode
labels	Label	Extra labels' configurations (see below).
show-overflow	bool	Show under- and overflow	false
errors-type	string	normal (ie. sqrt(N)), poisson (68% Poisson interval) or poisson2	poisson
yields-table-stretch	number	Stretch of LaTeX table produced	1.15
yields-table-align	string	Alignment of the LaTeX table. For now only h (processes horizontal, categories vertical) is supported.	h
yields-table-text-align	string	l/r/c: alignment of text in the LaTeX table's entries.	c
yields-table-numerical-precision-yields	number	Number of decimals used for the MC yields in the LaTeX table.	1
yields-table-numerical-precision-ratio	number	Number of decimals used for the data/MC ratio in the LaTeX table.	2
line-color	color	Default color of lines	1
line-type	line style	Default style (type) of lines	1
line-width	number	Default width of lines	1
book-keeping-file	string	If set, store all the produced plots into this file in addition to create graphic files	none
x-axis-label-size	number	The size of the labels on the x-axis, in pixel	18
y-axis-label-size	number	The size of the labels on the y-axis, in pixel	18
x-axis-top-ticks	bool	Show ticks at the top of the frame	True
y-axis-right-ticks	bool	Show ticks at the right of the frame	True
transparent-background	bool	Transparent background for canvas and legend	False

Plots configuration ('plots')

Field	Type	Action	Default
exclude	string	Regexp allowing to exlude histograms present in the files whose name matches the plot's name
x-axis	string	x-axis title
y-axis	string	y-axis title	Events
y-axis-format	formatted string	Overrides plotIt option.
normalized	bool	Normalize data/each signal/total MC to 1.	false
no-data	bool	Do not plot data.	false
override	bool	If any plot has this field set to true, only plots which do will be produced.	false
log-y	bool	Log-scale on y-axis. Special value: Both.	false
log-x	bool	Log-scale on x-axis. Special value: Both.	false
save-extensions	list of strings	Write plot to file in each of this formats (supported by TCanvas::Print())	[pdf]
show-ratio	bool	Show ratio plot under main plot.	false
fit	bool	Do a fit	false
fit-function	string	Function used for the fit.	gaus
fit-legend	formatted string	Fit results. Arg. i: fitted parameter i.	#scale[1.6]{#splitline{#mu = %2$.3f}{#sigma = %3$.3f}}
fit-legend-position	Point	Position of the fit results	[0.22, 0.87]
fit-range	Range	Restrict fit to this range
fit-ratio	bool	Fit the ratio plot	false
ratio-fit-function	string	Function used for the ratio fit.	pol1
ratio-fit-legend	formatted string	Ratio fit results. Arg. i: fitted parameter i.
ratio-fit-legend-position	Point	Position of the ratio fit results	[0.20, 0.38]
ratio-fit-range	Range	Restrict ratio fit to this range
show-errors	bool	Show errors on data points.	true
x-axis-range	Range	Range of x-axis.	TH1's range
log-x-axis-range	Range	Range of x-axis to be applied only if the x axis is in log scale.	Same as x-axis-range
y-axis-range	Range	Range of y-axis.	Computed from each TH1's range
log-y-axis-range	Range	Range of y-axis to be applied only if the y axis is in log scale.	Same as y-axis-range
ratio-y-axis-range	Range	Range of y-axis for ratio	[0.5, 1.5]
blinded-range	Range	Blind a range of data in the plot. Note that the bins that include the values used to specify the range will be blinded, and that the left edge is included in a bin (so you may have to write 4.9999 to blind up to 5, itself excluded).
y-axis-show-zero	bool	Force y-axis range to go down to zero (might not be the case for flat distributions).	false
inherits-from	string		TH1
rebin	number	Rebin: merge x bins into one	1
labels	Label	Extra labels' configurations (see below).
extra-label	string
legend-position	Position	Overrides plotIt option.
legend-columns	number	Overrides plotIt option.
show-overflow	bool	Overrides plotIt option.
errors-type	string	Overrides plotIt option.
binning-x	number
binning-y	number
draw-string	string	In tree mode, formula used to fill the histogram
selection-string	string	In tree mode, formula used to select events when filling histogram
for-yields	bool	Plot defines an entry as a category in the LaTeX yields table.	false
yields-title	string	Name of the category in the table. Must be valid LaTeX text string!	plot's name
yields-table-order	number	Order of appearance of the category in the yields table (small=up).	0
vertical-lines	Line	Draw vertical lines on the plot	Empty
horizontal-lines	Line	Draw horizontal lines on the plot	Empty
lines	Line	Draw arbitrary lines on the plot	Empty
rename	List of renaming operations	Apply a set of renaming operations to the output filename
sort-by-yields	bool	If true, histograms inside a given stack are sorted by yields, the smallest one at the bottom of the stack. This is the default behavior. Set to false to disable.	true
x-axis-label-size	number	The size of the labels on the x-axis, in pixel	Value of x-axis-label-size of the main configuration
y-axis-label-size	number	The size of the labels on the y-axis, in pixel	Value of y-axis-label-size of the main configuration
x-axis-hide-ticks	bool	Hide ticks on the x axis	False
y-axis-hide-ticks	bool	Hide ticks on the y axis	False
ratio-y-axis	string	Y axis title of the ratio plot	Value of ratio-y-axis of the main configuration

Files configuration ('files')

Each MC contribution is scaled by plotIt.scale*<luminosity>*scale*cross-section*branching-ratio/generated-events, where <luminosity> is the luminosity for the era of the file, or the total luminosity if no era is specified.

Field	Type	Action	Default
type	string	data, mc (all MC are stacked based on their stack-index) or signal (not stacked)	mc
scale	number	Scale this contribution	1
cross-section	number	Cross-section of the process	1
branching-ratio	number	Branching ratio of the process	1
generated-events	number	Careful! Sum of generated event weights! Can be different from the number of generated events.	1
era	string	era (used to select the corresponding integrated luminosity value for normalisation of MC, when using eras)
order	number	Order in which the histograms are stacked (concerns MC only). Small=low.
legend	string	Name of the file in the legend, unless group is defined .
group	Group	Refers to a legend group (see below)
yields-group	string	Used to group contributions in the LaTeX yields table (does't use a "group" node such as the legend groups).	In this order, defaults to group, legend or the file's name.
systematics	Systematics	Set of systematics nodes, as defined below.
drawing-options	string	Possibility to override TH1::Draw() option.	hist for signals and MC, P for data.
fill-color	color	For this and the following: -1 means that the quantity is not set.	MC: 1, else: -1
fill-type	fill style		MC: 1001 , signal: 0, else: -1
line-color	color		MC: -1, else: 1
line-type	line style		signal: 2, else: -1
line-width	line width		MC: 0, else: 1
marker-color	color		MC & signal: -1, else: 1
marker-type	marker style		MC & signal: -1, else: 20
marker-size	number		MC & signal: -1, else: 1
stack-index	number	Only MC files with the same stack-index are stacked together. Useful to have multiple stacks in the same plot	0
rename	List of renaming operations	Apply a set of renaming operations to all the histograms loaded from this file. It can be used for example to plot two different histograms from the same input file

Additional fields

Groups ('groups')

Legend ('legend')

Systematics configuration ('systematics')

Different types of systematic uncertainties are supported: shape, log-normal and constant.

Each shape uncertainty requires an up and down variated histogram. They are retrieved using its name, specified in the yml file as:

systematics:
  - <systematic>

Or, equivalently:

systematics:
  - <systematic>:
     type: shape

Based on this, plotIt will either look for two histograms called <nominal>__<systematic>[up|down], where <nominal> is the nominal histogram, in the file containing all the nominal histograms, or look for histograms called <nominal> in two files called <nominalFile>__<systematic>[up|down].root (where <nominalFile> is the file containing the nominal histograms for that process).

A constant systematic is just a rate uncertainty, which can specified either as:

systematics:
  - <systematic>: <value>

or as:

systematics:
  - <systematic>:
     type: const
     value: <value>

where <value> is the value with which the nominal histogram should be scaled up and down (by <value> and 2-<value>): e.g. with 1.025, the histograms are scaled by 1.025 and 0.975, i.e. by +/- 2.5%.

Any kind of systematics can be restricted to a subset of files by means of a regular expression:

systematics:
  - <systematic>:
     type: shape
     on: 'ttbar'

will only apply the <systematic> on files containing ttbar.

When plotting, all uncertainties will be added in quadrature, assuming zero correlations between all of them.

Additional labels ('labels')

Each plot can be assigned a set of labels. A label is a custom string which can be positioned anywhere on the canvas. You can defined labels as follow:

labels:
  - {<label1>}
  - {<label2>}

Each label has the following fields:

Field	Type	Action	Default
text	string	Text of the label. Support TLatex syntax (#splitline{}{} is especially useful for multi-line labels)
position	Point	Position of the label on the canvas
size	number	Font size of the label (in pixel)	18

Example

plots:
  'Plot':
    x-axis: "X axis"
    y-axis: "Y axis"
    labels:
      - {text: 'Label', position: [0.7, 0.7]}
      - {text: 'Label 2', position: [0.3, 0.7]}

Lines

You can draw lines over a plot using the vertical-lines, horizontal-lines and lines plot options. Each option accepts an array of line. A line is defined:

• By a number for vertical-lines and horizontal-lines. In this case, the number represents the x or y coordinate of the line (in plot axis units), and the line will span across all the plot.
• By an array of 2 Points for lines, where the Points represent respectively the start point and the end point of the line.

Note

in lines definition, you can use the special value .nan for a point coordinate. This value will automatically be replaced by the maximum or minimum of the x or y axis, depending on which place it's used.

In addition, styling can be applied to lines either globally in the configuration part, or individually for each line.

• In global configuration, you can set the line-color, line-width and line-type options. This will affect all lines.
• If you want to change the style individually, then you must convert the line definition into a map. The value key will hold the previous line definition (ie, either a number or an array of 2 Points), and you can set the line-color, line-width and line-type as you would do in the global configuration.

If you want a specific line to be drawn on the ratio instead of the main plot, you can set the pad option to bottom (it defaults to top).

Example

plots:
  'histo1':
    # Draw two vertical lines at x = 3 and x = 7. The first line is stylized individually and drawn on the ratio
    vertical-lines:
      - {
          line-color: 10,
          line-type: 4,
          pad: bottom,
          value: 3
        },
      - 7
    # Draw one horizontal line at y = 80
    horizontal-lines:
      - 80
    # Draw:
    # 1 line from (3, 20) -> (8, 90)
    # 1 vertical line at x = 5 (.nan is replaced automatically by the minimum and maximum of the y-axis)
    lines:
      - [[3, 20], [8, 90]]
      - [[5, .nan], [5, .nan]]

Renaming operations

A renaming operation is defined in YAML as a dictionary with the keys from and to. from is a regular expression following the extended POSIX syntax. Any string matching this regular expression will be replaced by to. Matching groups are supported and can be accessed using the \N syntax, where N is the index of the matching group.

The rename option expect a list of operation. Renaming operations are applied in sequence, starting from the first occurrence.

Example

files:
    'file.root':
        rename:
            - from: '([[:alpha:]+)_cut$'
              to: '\1_nocut'

Parameters cheatsheets

Line styles

Line widths

Fill styles

Marker styles

Colors

ROOT basic colors