updated data templates

rkansal47 · Mar 15, 2024 · 2a9fe3f · 2a9fe3f
1 parent 11a7575
commit 2a9fe3f
Show file tree

Hide file tree

Showing 86 changed files with 1,878 additions and 318 deletions.
diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
@@ -39,12 +39,12 @@ repos:
       - id: rst-directive-colons
       - id: rst-inline-touching-normal
 
-  - repo: https://github.com/pre-commit/mirrors-prettier
-    rev: "v4.0.0-alpha.8"
-    hooks:
-      - id: prettier
-        types_or: [yaml, markdown, html, css, scss, javascript, json]
-        args: [--prose-wrap=always]
+  # - repo: https://github.com/pre-commit/mirrors-prettier
+  #   rev: "v4.0.0-alpha.8"
+  #   hooks:
+  #     - id: prettier
+  #       types_or: [yaml, markdown, html, css, scss, javascript, json]
+  #       args: [--prose-wrap=always]
 
   - repo: https://github.com/astral-sh/ruff-pre-commit
     rev: "v0.3.2"

diff --git a/src/HHbbVV/postprocessing/BDTPreProcessing.ipynb b/src/HHbbVV/postprocessing/BDTPreProcessing.ipynb
diff --git a/src/HHbbVV/postprocessing/BDTPreProcessing.py b/src/HHbbVV/postprocessing/BDTPreProcessing.py
@@ -92,7 +92,6 @@ def save_bdt_data(
 
     bdt_events_dict = []
     for key in BDT_sample_order:
-        print(key)
         save_vars = BDT_data_vars + jec_jmsr_vars if key != "Data" else BDT_data_vars
         events = pd.DataFrame(
             {var: utils.get_feat(events_dict[key], var, bb_masks[key]) for var in save_vars}

diff --git a/src/HHbbVV/postprocessing/BDT_LPSFs.ipynb b/src/HHbbVV/postprocessing/BDT_LPSFs.ipynb
diff --git a/src/HHbbVV/postprocessing/PostProcessRes.ipynb b/src/HHbbVV/postprocessing/PostProcessRes.ipynb
@@ -91,14 +91,14 @@
    "outputs": [],
    "source": [
     "MAIN_DIR = Path(\"../../../\")\n",
-    "samples_dir = MAIN_DIR / \"../data/skimmer/24Mar5AllYears\"\n",
+    "samples_dir = MAIN_DIR / \"../data/skimmer/24Mar6AllYearsBDTVars\"\n",
     "# samples_dir = f\"{MAIN_DIR}/../data/skimmer/Feb24\"\n",
     "# nonres_signal_samples_dir = f\"{MAIN_DIR}/../data/skimmer/Jun10\"\n",
     "# res_signal_samples_dir = f\"{MAIN_DIR}/../data/skimmer/Apr11\"\n",
     "# samples_dir = \"/eos/uscms/store/user/rkansal/bbVV/skimmer/Feb24\"\n",
     "# nonres_signal_samples_dir = \"/eos/uscms/store/user/cmantill/bbVV/skimmer/Jun10/\"\n",
     "# res_signal_samples_dir = \"/eos/uscms/store/user/rkansal/bbVV/skimmer/Apr11/\"\n",
-    "year = \"2016\"\n",
+    "year = \"2016APV\"\n",
     "\n",
     "date = \"24Mar6\"\n",
     "plot_dir = MAIN_DIR / f\"plots/PostProcessing/{date}/\"\n",
@@ -135,7 +135,7 @@
     "\n",
     "events_dict = postprocessing.load_samples(\n",
     "    samples_dir,\n",
-    "    {**nonres_samples, **res_samples, **samples},\n",
+    "    {**nonres_samples},  # , **res_samples, **samples},\n",
     "    year,\n",
     "    load_filters,\n",
     ")\n",
@@ -144,6 +144,33 @@
     "cutflow"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "events = events_dict[\"HHbbVV\"]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "list(events.columns)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "events[\"DijetMass\"]"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,

diff --git a/src/HHbbVV/postprocessing/bash_scripts/BDTPreProcessing.sh b/src/HHbbVV/postprocessing/bash_scripts/BDTPreProcessing.sh
@@ -7,7 +7,7 @@
 ####################################################################################################
 
 MAIN_DIR="../../.."
-data_dir="$MAIN_DIR/../data/skimmer/24Mar6AllYearsBDTVars"
+data_dir="$MAIN_DIR/../data/skimmer/24Mar14UpdateData"
 
 for year in 2016APV 2016 2017 2018
 do

diff --git a/src/HHbbVV/postprocessing/bash_scripts/NonresTemplates.sh b/src/HHbbVV/postprocessing/bash_scripts/NonresTemplates.sh
@@ -7,7 +7,7 @@
 ####################################################################################################
 
 MAIN_DIR="../../.."
-data_dir="$MAIN_DIR/../data/skimmer/24Mar6AllYearsBDTVars"
+data_dir="$MAIN_DIR/../data/skimmer/24Mar14UpdateData"
 TAG=""
 
 
@@ -40,10 +40,11 @@ if [[ -z $TAG ]]; then
   exit 1
 fi
 
-for year in 2016APV 2016 2017 2018
+# for year in 2016APV 2016 2017 2018
+for year in 2017 2018 2016APV 2016
 do
     python -u postprocessing.py --year $year --data-dir "$data_dir" --templates \
-    --bdt-preds-dir "$MAIN_DIR/../data/skimmer/24Mar6AllYearsBDTVars/24_03_07_new_samples_max_depth_5/inferences" \
+    --bdt-preds-dir "$data_dir/24_03_07_new_samples_max_depth_5/inferences" \
     --plot-dir "${MAIN_DIR}/plots/PostProcessing/$TAG" \
     --template-dir "templates/$TAG" --plot-shifts
     # --control-plots --control-plot-vars "BDTScore" \

diff --git a/src/HHbbVV/postprocessing/postprocessing.py b/src/HHbbVV/postprocessing/postprocessing.py
@@ -1136,10 +1136,16 @@ def derive_variables(
     do_jshifts: bool = True,
 ):
     """Add Dijet variables"""
+    import time
+
+    start = time.time()
+
     for sample, events in events_dict.items():
         if not nonres_vars:
             continue
 
+        print(f"Deriving variables for {sample} {time.time() - start:.2f}s")
+
         bb_mask = bb_masks[sample]
         _add_nonres_columns(events, bb_mask, vbf_vars=vbf_vars)
 

diff --git a/src/HHbbVV/postprocessing/templates/24Mar15UpdateData/2016APV_templates.pkl b/src/HHbbVV/postprocessing/templates/24Mar15UpdateData/2016APV_templates.pkl
diff --git a/src/HHbbVV/postprocessing/templates/24Mar15UpdateData/2016_templates.pkl b/src/HHbbVV/postprocessing/templates/24Mar15UpdateData/2016_templates.pkl
diff --git a/src/HHbbVV/postprocessing/templates/24Mar15UpdateData/2017_templates.pkl b/src/HHbbVV/postprocessing/templates/24Mar15UpdateData/2017_templates.pkl
diff --git a/src/HHbbVV/postprocessing/templates/24Mar15UpdateData/2018_templates.pkl b/src/HHbbVV/postprocessing/templates/24Mar15UpdateData/2018_templates.pkl
diff --git a/src/HHbbVV/postprocessing/templates/24Mar15UpdateData/cutflows/2016/fail_cutflow.csv b/src/HHbbVV/postprocessing/templates/24Mar15UpdateData/cutflows/2016/fail_cutflow.csv
@@ -0,0 +1,25 @@
+,Pre-selection,QCD SF,LP SF,bbFatJetPt >= 300,VVFatJetPt >= 300,0.8 ≤ bbFatJetParticleNetMD_Txbb < 0.9735,0 ≤ nGoodElectronsHbb < 0.9,0 ≤ nGoodMuonsHbb < 0.9
+HHbbVV,0.9147081031687907,0.9147081031687907,0.9147081031687907,0.9088022319312213,0.9037950499185726,0.30175414428376585,0.30112232829530405,0.2990202748826515
+ggHH_kl_2p45_kt_1_HHbbVV,0.6688447704532685,0.6688447704532685,0.6688447704532685,0.6649531572120448,0.6617354385600798,0.2299036391879109,0.2295289421882006,0.2280924478485079
+ggHH_kl_5_kt_1_HHbbVV,0.6956895087373941,0.6956895087373941,0.6956895087373941,0.6920514447111688,0.6892047021562906,0.3063302486063932,0.30550315963569236,0.30391351000104144
+ggHH_kl_0_kt_1_HHbbVV,1.1944766869755938,1.1944766869755938,1.1944766869755938,1.1863493648635597,1.17930333767202,0.39586526060371247,0.39514188691284224,0.39256694301449424
+VBFHHbbVV,0.017115373482989717,0.017115373482989717,0.017115373482989717,0.017030078444247275,0.016924919012766975,0.00579761879633008,0.0057790354556334165,0.0057248101820241174
+qqHH_CV_1_C2V_0_kl_1_HHbbVV,9.060969753010587,9.060969753010587,9.060969753010587,9.041934826728792,9.02728294983494,2.4426924669829826,2.43781489891332,2.4243925435159057
+qqHH_CV_1p5_C2V_1_kl_1_HHbbVV,15.269096111629358,15.269096111629358,15.269096111629358,15.23647158421826,15.20829496277662,4.13073889572393,4.120987792038875,4.10015685148764
+qqHH_CV_1_C2V_1_kl_2_HHbbVV,0.0173959132140592,0.0173959132140592,0.0173959132140592,0.017299990201845,0.017214419089199705,0.00846784274434387,0.008440107570873445,0.008373264656125586
+qqHH_CV_1_C2V_2_kl_1_HHbbVV,8.138882127595444,8.138882127595444,8.138882127595444,8.123943352838433,8.11286122859791,2.1892512907726895,2.1837678445588895,2.1717354106435
+qqHH_CV_1_C2V_1_kl_0_HHbbVV,0.04617824244896124,0.04617824244896124,0.04617824244896124,0.0459311591300662,0.04565497610400018,0.014190872302543785,0.01416197212352274,0.014112105993637325
+QCD,717986.216814939,956785.8146709627,956785.8146709627,948406.4124347997,941617.267942029,754552.8113641972,752984.4192876149,750107.2737808328
+TT,63126.80979180276,63126.80979180276,63126.80979180276,62714.21430804499,62290.703474200425,49529.77449791308,44329.76646173106,39159.45860664165
+ST,5417.824558008516,5417.824558008516,5417.824558008516,5380.176293677084,5350.773696634727,4335.103148326725,3970.1436871856044,3563.758325231233
+W+Jets,9487.889178263631,9487.889178263631,9487.889178263631,9424.292575831972,9375.118673797197,7834.787103754956,7814.4993526839335,7789.446400547657
+Z+Jets,12055.822300436781,12055.822300436781,12055.822300436781,11968.016139487008,11904.514452147456,6790.915177024769,6773.707406230397,6734.8455822324595
+Diboson,544.751536187187,544.751536187187,544.751536187187,541.5192000426408,539.5255957808083,336.0954273567496,320.18549221259616,303.48286314956806
+ggFHbb,120.74204174793252,120.74204174793252,120.74204174793252,120.02707196890958,119.27175261578469,38.999571077551494,38.97813391667,38.788894151509346
+VBFHbb,32.24173303425971,32.24173303425971,32.24173303425971,32.089487243194796,31.921442997500897,9.845327900626657,9.826780907273822,9.728638615140152
+ZHbb,32.70020301080715,32.70020301080715,32.70020301080715,32.51666448179822,32.40029638484181,9.92855058278212,9.896574590344532,9.807368124790582
+WHbb,56.014320595270185,56.014320595270185,56.014320595270185,55.69337399997595,55.51504919284625,19.11120104088777,19.06683717160466,18.91991542799836
+ggZHbb,4.8805631912185055,4.8805631912185055,4.8805631912185055,4.852821845932874,4.827078996358681,1.9281741450543308,1.9243343671398156,1.9080303961975615
+ttHbb,156.94119216557863,156.94119216557863,156.94119216557863,155.9939482895008,154.97672213467678,97.04622661247781,85.884977667534,74.2696929751333
+HWW,80.56791059341468,80.56791059341468,80.56791059341468,80.1332628058381,79.59897023669294,63.60672027477934,52.77270242090221,43.58177362660231
+Data,1047903.0,1047903.0,1047903.0,1047903.0,1047903.0,863853.0,854944.0,844713.0