Add plot script for e/pi separation performance

• update the json output from epcut and ML so rejection power and uncertainties can be directly read from them
• add a json output for run_benchmark.py to collect all run information
• add a script reading the run info files and plot the e/pi separation power curve

benchmarks/imaging_shower_ML/scripts/draw_imaging_event3d.py

+"""
+    A script to draw imaging events in 3D
+    Assumed the input reconstruction file has only single-particle events
+
+    Chao Peng (ANL)
+    2022/12/06
+"""
+import ROOT
+import os
+import gc
+import sys
+import json
+import numpy as np
+import pandas as pd
+import argparse
+from utils import flatten_collection, cartesian_to_polar, imcal_info, center_of_gravity
+
+
+pd.set_option('display.max_rows', 500)
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(description='Visualize the cluster (layer-wise) from analysis')
+    parser.add_argument(
+            'file', type=str,
+            help='path to the input file.'
+            )
+    parser.add_argument(
+            '-o', type=str,
+            default='.',
+            dest='outdir',
+            help='output directory.'
+            )
+    parser.add_argument(
+            '-e', type=int,
+            default=0,
+            dest='event',
+            help='event number to draw.'
+            )
+    parser.add_argument(
+            '--nhits', type=int,
+            default=50,
+            dest='nhits',
+            help='number of hits per layer.'
+            )
+    parser.add_argument(
+            '-b', '--branch', type=str,
+            dest='branch',
+            default='EcalBarrelImagingHitsReco',
+            help='name of data branch (edm4eic::CalorimeterHitCollection).'
+            )
+    parser.add_argument(
+            '--eta-ngrid', type=int,
+            dest='eta_ngrid',
+            default=40,
+            help='number of eta grids for drawing.')
+    parser.add_argument(
+            '--eta-grid', type=float,
+            dest='eta_grid',
+            default=0.01,
+            help='eta grid size for drawing.')
+    parser.add_argument(
+            '--phi-ngrid', type=int,
+            dest='phi_ngrid',
+            default=40,
+            help='number of phi grids for drawing.')
+    parser.add_argument(
+            '--phi-grid', type=float,
+            dest='phi_grid',
+            default=0.01,
+            help='phi grid size for drawing (rad).')
+    parser.add_argument('--topo-range', type=float, default=50.0, dest='topo_range',
+                        help='half range for projection plot (mrad)')
+    args = parser.parse_args()
+
+    # read data and MCParticles
+    rdf = ROOT.RDataFrame("events", args.file)
+    # event range
+    ntotal = rdf.Count().GetValue()
+    iev = int(np.clip(args.event, 0, ntotal))
+
+    # data
+    data = flatten_collection(rdf, args.branch, (iev, iev + 1), cols=[
+            'layer',
+            'energy',
+            'position.x', 'position.y', 'position.z',
+            ])
+    etot = data['energy'].sum()
+    xc, yc, zc = center_of_gravity(*data[['position.x', 'position.y', 'position.z']].values.T,
+                                   np.log(data['energy'].values/etot), 4.6)
+
+    rc, thetac, phic, r0c, etac = cartesian_to_polar(xc, yc, zc)
+    r, theta, phi, rc, eta = cartesian_to_polar(*data[['position.x', 'position.y', 'position.z']].values.T)
+
+    dfg = pd.DataFrame(data=np.vstack((*data[['layer', 'energy']].values.T, eta, phi)).T,
+                       columns=['layer', 'energy', 'eta', 'phi'])
+    eta_bins = np.arange(args.eta_ngrid + 1)*args.eta_grid
+    eta_bins = eta_bins - eta_bins[-1]/2. + etac
+    phi_bins = np.arange(args.phi_ngrid + 1)*args.phi_grid
+    phi_bins = phi_bins - phi_bins[-1]/2. + phic
+    dfg.loc[:, 'eta_bin'] = np.digitize(dfg['eta'], eta_bins)
+    dfg.loc[:, 'phi_bin'] = np.digitize(dfg['phi'], phi_bins)
+    print(dfg)
+    # eta = eta - etac
+    # phi = phi - phic
+
+    # determine eta and phi windows according to event (cluster) center
+