fix: update I/O branches (#126)

* added script to get I/O branches and updated list of branches to hit various thresholds
* moved some config options to a dedicated config file

Author: ekauffma

analyses/cms-open-data-ttbar/GetIOBranches.py

@@ -0,0 +1,47 @@
+import json
+import numpy as np
+
+### OPTIONS
+ratio_json_path = "nanoaod_branch_ratios.json"
+agc_original_branches = ["Jet_pt", "Jet_eta", "Jet_phi", "Jet_btagCSVV2", "Jet_mass", 
+                         "Muon_pt", "Electron_pt"]
+desired_percents = [15,25,50]
+
+
+def main():
+    
+    with open(ratio_json_path) as json_file:
+        branch_ratios = json.load(json_file)
+    
+    io_branch_dict = {}
+
+    # calculate percentage associated with original AGC branches
+    current_sum = 0
+    for key in branch_ratios.keys():
+        if key in agc_original_branches:
+            current_sum+=branch_ratios[key]
+    io_branch_dict[np.round(100*current_sum,1)] = agc_original_branches
+
+    sortind = np.flip(np.argsort(list(branch_ratios.values())))
+    keys = np.array(list(branch_ratios.keys()))[sortind]
+    values = np.array(list(branch_ratios.values()))[sortind]
+
+    for percent in desired_percents:
+        branch_names = []
+        current_sum = 0
+        for i, key in enumerate(keys):
+            if 100*values[i] > 1.02*(percent-100*current_sum):
+                continue
+            if 100*(current_sum+values[i])>=percent:
+                print(f"Expected Percentage = {percent}, Calculated Percentage = {100*np.round(current_sum,4)}, Number of Branches = {len(branch_names)}")
+                break
+            branch_names.append(key)
+            current_sum+=values[i]
+        io_branch_dict[percent] = branch_names
+
+    print(json.dumps(io_branch_dict, sort_keys=True, indent=4))
+
+
+if __name__ == "__main__":
+    main()
+

analyses/cms-open-data-ttbar/config.yaml

@@ -0,0 +1,82 @@
+global:
+
+  # ServiceX: ignore cache with repeated queries
+  SERVICEX_IGNORE_CACHE: false
+  
+  # analysis facility: set to "coffea_casa" for coffea-casa environments, "EAF" for FNAL, "local" for local setups
+  AF: coffea_casa
+  
+benchmarking:
+
+  # chunk size to use
+  CHUNKSIZE: 500000
+  
+  # metadata to propagate through to metrics
+  # "ssl-dev" allows for the switch to local data on /data
+  AF_NAME: coffea_casa
+  
+  # currently has no effect
+  SYSTEMATICS: all
+  
+  # does not do anything, only used for metric gathering (set to 2 for distributed coffea-casa)
+  CORES_PER_WORKER: 2
+  
+  # scaling for local setups with FuturesExecutor
+  NUM_CORES: 4
+  
+  # only I/O, all other processing disabled
+  DISABLE_PROCESSING: false
+  
+  # read additional branches (only with DISABLE_PROCESSING = True)
+  # acceptable values are 4.1, 15, 25, 50 (corresponding to % of file read), 4.1% corresponds to the standard branches used in the notebook
+  IO_FILE_PERCENT: '4.1'
+  
+  # nanoAOD branches that correspond to different values of IO_FILE_PERCENT
+  IO_BRANCHES:
+    '4.1':
+      - Jet_pt
+      - Jet_eta
+      - Jet_phi
+      - Jet_btagCSVV2
+      - Jet_mass
+      - Muon_pt
+      - Electron_pt
+    '15':
+      - LHEPdfWeight
+      - GenPart_pdgId
+      - CorrT1METJet_phi
+    '25':
+      - LHEPdfWeight
+      - GenPart_pt
+      - GenPart_eta
+      - GenPart_pdgId
+      - LHEScaleWeight
+    '50':
+      - LHEPdfWeight
+      - GenPart_pt
+      - GenPart_eta
+      - GenPart_phi
+      - GenPart_pdgId
+      - GenPart_genPartIdxMother
+      - GenPart_statusFlags
+      - GenPart_mass
+      - LHEScaleWeight
+      - GenJet_pt
+      - GenPart_status
+      - LHEPart_eta
+      - LHEPart_phi
+      - LHEPart_pt
+      - GenJet_eta
+      - GenJet_phi
+      - Jet_eta
+      - Jet_phi
+      - SoftActivityJet_pt
+      - SoftActivityJet_phi
+      - SoftActivityJet_eta
+      - GenJet_mass
+      - Jet_pt
+      - Jet_mass
+      - LHEPart_mass
+      - Jet_qgl
+      - Jet_muonSubtrFactor
+      - Jet_puIdDisc