merge

rkansal47 · Mar 5, 2024 · f58fde3 · f58fde3
1 parent d04b04d
commit f58fde3
Show file tree

Hide file tree

Showing 2 changed files with 50 additions and 58 deletions.
diff --git a/src/HHbbVV/postprocessing/PostProcessRes.ipynb b/src/HHbbVV/postprocessing/PostProcessRes.ipynb
@@ -12,7 +12,7 @@
     "import corrections\n",
     "\n",
     "from utils import CUT_MAX_VAL, ShapeVar\n",
-    "from hh_vars import (\n",
+    "from HHbbVV.hh_vars import (\n",
     "    years,\n",
     "    data_key,\n",
     "    qcd_key,\n",
@@ -92,6 +92,7 @@
    "source": [
     "MAIN_DIR = Path(\"../../../\")\n",
     "samples_dir = MAIN_DIR / \"../data/skimmer/24Feb25_update_skimmer\"\n",
+    "sig_samples_dir = MAIN_DIR / \"../data/skimmer/24Mar5_update_lp\"\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",
@@ -100,7 +101,7 @@
     "# res_signal_samples_dir = \"/eos/uscms/store/user/rkansal/bbVV/skimmer/Apr11/\"\n",
     "year = \"2017\"\n",
     "\n",
-    "date = \"24Mar2\"\n",
+    "date = \"24Mar5\"\n",
     "plot_dir = MAIN_DIR / f\"plots/PostProcessing/{date}/\"\n",
     "templates_dir = Path(f\"templates/{date}/\")\n",
     "\n",
@@ -133,22 +134,34 @@
     "    index=list(samples.keys()) + list(nonres_samples.keys()) + list(res_samples.keys())\n",
     ")\n",
     "\n",
-    "# hem cleaning in load_samples not implemented yet for res samples\n",
-    "hem_cleaning = True\n",
-    "\n",
-    "# utils.remove_empty_parquets(samples_dir, year)\n",
     "events_dict = postprocessing.load_samples(\n",
+    "    sig_samples_dir,\n",
+    "    {**nonres_samples, **res_samples},\n",
+    "    year,\n",
+    "    new_filters,\n",
+    ")\n",
+    "\n",
+    "events_dict |= postprocessing.load_samples(\n",
     "    samples_dir,\n",
-    "    {**nonres_samples, **res_samples, **samples},\n",
+    "    samples,\n",
     "    year,\n",
     "    new_filters,\n",
-    "    hem_cleaning=hem_cleaning,\n",
     ")\n",
     "\n",
     "utils.add_to_cutflow(events_dict, \"Preselection\", \"finalWeight\", cutflow)\n",
     "cutflow"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "events = pd.read_parquet(f\"{sig_samples_dir}/{year}/GluGluToHHTobbVV_node_cHHH1/parquet\")\n",
+    "events"
+   ]
+  },
   {
    "attachments": {},
    "cell_type": "markdown",
@@ -226,7 +239,7 @@
     "    bb_masks,\n",
     "    nonres_sig_keys + res_sig_keys,\n",
     "    control_plot_vars,\n",
-    "    f\"{plot_dir}/ControlPlots/{year}/\",\n",
+    "    plot_dir / f\"ControlPlots/{year}\",\n",
     "    year,\n",
     "    bg_keys=bg_keys,\n",
     "    sig_scale_dict={\"HHbbVV\": 1e5, \"VBFHHbbVV\": 2e6} | {key: 2e4 for key in res_sig_keys},\n",
@@ -243,6 +256,32 @@
     "Overall LP SF"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from postprocessing import Region, nonres_shape_vars\n",
+    "\n",
+    "# temp region to check systematics\n",
+    "selection_regions = {\n",
+    "    \"pass\": Region(\n",
+    "        cuts={\n",
+    "            \"bbFatJetParticleNetMD_Txbb\": [0.97, CUT_MAX_VAL],\n",
+    "            \"VVFatJetParTMD_THWWvsT\": [0.8, CUT_MAX_VAL],\n",
+    "        },\n",
+    "        label=\"Pass\",\n",
+    "    ),\n",
+    "    \"lpsf\": Region(\n",
+    "        cuts={\n",
+    "            \"VVFatJetParTMD_THWWvsT\": [0.8, CUT_MAX_VAL],\n",
+    "        },\n",
+    "        label=\"LP SF\",\n",
+    "    ),\n",
+    "}"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -305,26 +344,6 @@
     "# del selection_regions[\"fail\"], selection_regions[\"failBlinded\"]"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from postprocessing import Region, nonres_shape_vars\n",
-    "\n",
-    "# temp region to check systematics\n",
-    "selection_regions = {\n",
-    "    \"pass\": Region(\n",
-    "        cuts={\n",
-    "            \"bbFatJetParticleNetMD_Txbb\": [0.97, CUT_MAX_VAL],\n",
-    "            \"VVFatJetParTMD_THWWvsT\": [0.8, CUT_MAX_VAL],\n",
-    "        },\n",
-    "        label=\"Pass\",\n",
-    "    )\n",
-    "}"
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -395,33 +414,6 @@
     "    templates = {**templates, **ttemps}"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "ttemps"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "np.linalg.norm((ttemps[1].values() - ttemps[0])[:, 10]) / "
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "np.linalg.norm([1, 1])"
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": null,

diff --git a/src/HHbbVV/postprocessing/postprocessing.py b/src/HHbbVV/postprocessing/postprocessing.py
@@ -1399,7 +1399,7 @@ def control_plots(
     bb_masks: Dict[str, pd.DataFrame],
     sig_keys: List[str],
     control_plot_vars: List[ShapeVar],
-    plot_dir: str,
+    plot_dir: Path,
     year: str,
     weight_key: str = "finalWeight",
     hists: Dict = {},
@@ -1446,7 +1446,7 @@ def control_plots(
     if HEM2d and year == "2018":
         hists["HEM2d"] = hists_HEM2d(events_dict, bb_masks, weight_key, selection)
 
-    with open(f"{plot_dir}/hists.pkl", "wb") as f:
+    with (plot_dir / "hists.pkl").open("wb") as f:
         pickle.dump(hists, f)
 
     if sig_splits is None: