cmt.base_tasks.plotting

Plotting tasks.

Class BasePlotTask

class cmt.base_tasks.plotting.BasePlotTask(*args, **kwargs)

Bases: ConfigTaskWithCategory

Task that wraps parameters used in all plotting tasks. Can’t be run.

Parameters:

• feature_names (csv string) – names of features to plot. Uses all features when empty.

• feature_tags (csv string) – list of tags for filtering features selected via feature names.

• skip_feature_names (csv string) – names or name pattern of features to skip

• skip_feature_tags (csv string) – list of tags of features to skip

• apply_weights (bool) – whether to apply weights and scaling to all histograms.

• n_bins (int) – Custom number of bins for plotting, defaults to the value configured by the feature when empty.

• systematics (csv list) – NOT YET IMPLEMENTED. List of custom systematics to be considered.

• store_systematics (bool) – whether to store systematic templates inside the output root files.

• shape_region (str from choice list) – shape region used for QCD computation.

• remove_horns (bool) – NOT YET IMPLEMENTED. Whether to remove the eta horns present in 2017

• optimization_method (str) – Optimization method to be used. Only bayesian_blocks available.

Class PrePlot

class cmt.base_tasks.plotting.PrePlot(*args, **kwargs)

Bases: DatasetTaskWithCategory, BasePlotTask, LocalWorkflow, HTCondorWorkflow, SGEWorkflow, RDFModuleTask

Performs the filling of histograms for all features considered. If systematics are considered, it also produces the same histograms after applying those.

Parameters:

• skip_processing (bool) – whether to skip the preprocessing and categorization steps.

• skip_merging (bool) – whether to skip the MergeCategorization task.

• preplot_modules_file (str) – filename inside cmt/config/ or ../config/ (w/o extension) with the RDF modules to run.

get_syst_list()

Returns a list of systematic names that affect present or past selections, so dedicated input ntuples are needed as requirements

create_branch_map()

Returns:

number of files after merging (usually 1) unless skip_processing == True

Return type:

int

workflow_requires()

requires()

Each branch requires one input file for the central value and two per systematic considered

output()

Returns:

One file per input file with all histograms to be plotted for each feature

Return type:

.root

get_weight(category, syst_name, syst_direction, **kwargs)

Obtains the product of all weights depending on the category/channel applied. Returns “1” if it’s a data sample or the apply_weights parameter is set to False.

Returns:

Product of all weights to be applied

Return type:

str

run()

Creates one RDataFrame per input file, runs the desired RDFModules and produces a set of plots per each feature, one for the nominal value and others (if available) for all systematics.

Class FeaturePlot

class cmt.base_tasks.plotting.FeaturePlot(*args, **kwargs)

Bases: BasePlotTask, DatasetWrapperTask

Performs the actual histogram plotting: loads the histograms obtained in the PrePlot tasks, rescales them if needed and plots and saves them.

Example command:

law run FeaturePlot --version test --category-name etau --config-name ul_2018 --process-group-name etau --feature-names Htt_svfit_mass,lep1_pt,bjet1_pt,lep1_eta,bjet1_eta --workers 20 --PrePlot-workflow local --stack --hide-data False --do-qcd --region-name etau_os_iso--dataset-names tt_dl,tt_sl,dy_high,wjets,data_etau_a,data_etau_b,data_etau_c,data_etau_d --MergeCategorizationStats-version test_old

Parameters:

• stack (bool) – whether to show all backgrounds stacked (True) or normalized to 1 (False)

• do_qcd (bool) – whether to estimate QCD using the ABCD method

• qcd_wp (str from choice list) – working point to use for QCD estimation

• qcd_signal_region_wp (str) – region to use as signal region for QCD estimation

• shape_region (str from choice list) – region to use as shape region for QCD estimation

• qcd_sym_shape (bool) – whether to symmetrise the shape coming from both possible shape regions

• qcd_category_name (str) – category name used for the same sign regions in QCD estimation

• hide_data (bool) – whether to show (False) or hide (True) the data histograms

• normalize_signals (bool) – whether to normalize signals to the total background yield (True) or not (False)

• avoid_normalization (bool) – whether to avoid normalizing by cross section and initial number of events

• blinded (bool) – whether to blind data in specified regions. The blinding ranges are specified using the blinded_range parameter in the Feature definition. This parameter can include a list ([initial, final]) or a list of lists ([[init_1, fin_1], [init_2, fin_2], …])

• save_png (bool) – whether to save plots in png

• save_pdf (bool) – whether to save plots in pdf

• save_root (bool) – whether to write plots in a root file

• save_yields (bool) – whether to save histogram yields in a json file

• process_group_name (str) – name of the process grouping name

• bins_in_x_axis (bool) – (NOT YET IMPLEMENTED) whether to plot histograms with the bin numbers in the x axis instead of the actual values

• plot_systematics (bool) – (NOT YET FULLY IMPLEMENTED) whether to plot histograms with their uncertainties

• fixed_colors (bool) – whether to plot histograms with their defined colors (False) or fixed colors (True) starting from ROOT color 2.

• log_y (bool) – whether to set y axis to log scale

• include_fit (str) – YAML file inside config folder (w/o extension) including input parameters for the fit

• propagate_syst_qcd (bool) – whether to propagate systematics to qcd background

requires()

All requirements needed:

• Histograms coming from the PrePlot task.

• Number of total events coming from the MergeCategorizationStats task (to normalize MC histograms).

• If estimating QCD, FeaturePlot for the three additional QCD regions needed.

get_output_postfix(key='pdf')

Returns:

string to be included in the output filenames

Return type:

str

output()

Output files to be filled: pdf, png, root or json

complete()

Task is completed when all output are present

setup_signal_hist(hist, color)

Method to apply signal format to an histogram

setup_background_hist(hist, color)

Method to apply background format to an histogram

setup_data_hist(hist, color)

Method to apply data format to an histogram

plot(feature, ifeat=0)

Performs the actual plotting.

run()

Splits processes into data, signal and background. Creates histograms from each process loading them from the input files. Scales the histograms and applies the correct format to them.