Merge branch 'dev' into plot_postfit_shapes

hbw/analysis/create_analysis.py

@@ -7,7 +7,6 @@
 from __future__ import annotations
 
 import os
-import importlib
 
 import law
 import order as od
@@ -24,6 +23,8 @@
     ml_inputs_producer,
 )
 
+from hbw.tasks.campaigns import BuildCampaignSummary
+
 
 @timeit_multiple
 def create_hbw_analysis(
@@ -74,6 +75,8 @@ def create_hbw_analysis(
     analysis_inst.x.default_weight_producer = "default"
     analysis_inst.x.ml_inputs_producer = ml_inputs_producer(analysis_inst)
     analysis_inst.x.default_ml_model = default_ml_model
+    analysis_inst.x.default_variables = ["jet0_pt", "mll", "n_jet", "ptll", "lepton0_pt", "lepton1_pt"]
+    analysis_inst.x.default_categories = ["incl", "sr", "ttcr", "dycr"]
 
     #
     # define configs
@@ -82,7 +85,6 @@ def create_hbw_analysis(
     from hbw.config.config_run2 import add_config
 
     def add_lazy_config(
-        campaigns: dict[str, str],
         config_name: str,
         config_id: int,
         **kwargs,
@@ -101,22 +103,22 @@ def create_factory(
         ):
             @timeit_multiple
             def analysis_factory(configs: od.UniqueObjectIndex):
-                hbw_campaign_inst = None
-
-                for mod, campaign in campaigns.items():
-                    # import the campaign
-                    mod = importlib.import_module(mod)
-                    if not hbw_campaign_inst:
-                        # copy the main campaign
-                        hbw_campaign_inst = getattr(mod, campaign).copy()
-                    else:
-                        # add datasets to the main campaign
-                        campaign_inst = getattr(mod, campaign).copy()
-                        for dataset in list(campaign_inst.datasets):
-                            dataset.x.campaign = campaign
-                            if not hbw_campaign_inst.has_dataset(dataset.name):
-                                hbw_campaign_inst.add_dataset(dataset)
-
+                cpn_task = BuildCampaignSummary(
+                    config=config_name,
+                )
+                if cpn_task.complete():
+                    logger.warning(
+                        f"Using pickled campaign for config {config_name}; to re-initialize, run:\n"
+                        f"law run {cpn_task.task_family} --config {config_name} --remove-output 0,a,y",
+                    )
+                else:
+                    raise ValueError(
+                        f"Campaign used for {config_name} is not yet initialized; to initialize, run: \n"
+                        f"law run {cpn_task.task_family} --config {config_name} --remove-output 0,a,y",
+                    )
+                    # cpn_task.run()
+
+                hbw_campaign_inst = cpn_task.output()["hbw_campaign_inst"].load(formatter="pickle")
                 return add_config(
                     analysis_inst,
                     hbw_campaign_inst,
@@ -139,41 +141,39 @@ def analysis_factory(configs: od.UniqueObjectIndex):
 
     # 2017
     add_lazy_config(
-        {
-            "cmsdb.campaigns.run2_2017_nano_v9": "campaign_run2_2017_nano_v9",
-        },
         "c17",
         1700,
     )
 
     # 2022 preEE
     add_lazy_config(
-        {
-            "cmsdb.campaigns.run3_2022_preEE_nano_v12": "campaign_run3_2022_preEE_nano_v12",
-            "cmsdb.campaigns.run3_2022_preEE_nano_v13": "campaign_run3_2022_preEE_nano_v13",
-        },
         "c22pre",
         2200,
     )
 
     # 2022 postEE
     add_lazy_config(
-        {
-            "cmsdb.campaigns.run3_2022_postEE_nano_v12": "campaign_run3_2022_postEE_nano_v12",
-            "cmsdb.campaigns.run3_2022_postEE_nano_v13": "campaign_run3_2022_postEE_nano_v13",
-        },
         "c22post",
         2210,
     )
 
+    add_lazy_config(
+        "c22pre_das",
+        2201,
+    )
+    add_lazy_config(
+        "c22post_das",
+        2211,
+    )
+
     #
     # modify store_parts
     #
 
     software_tasks = ("cf.BundleBashSandbox", "cf.BundleCMSSWSandbox", "cf.BundleSoftware")
     shareable_analysis_tasks = ("cf.CalibrateEvents", "cf.GetDatasetLFNs")
     limited_config_shared_tasks = ("cf.CalibrateEvents", "cf.GetDatasetLFNs", "cf.SelectEvents", "cf.ReduceEvents")
-    skip_new_version_schema = ("cf.CalibrateEvents", "cf.GetDatasetLFNs")
+    skip_new_version_schema = ()
     known_parts = (
         # from cf
         "analysis", "task_family", "config", "configs", "dataset", "shift", "version",
@@ -234,11 +234,13 @@ def reorganize_parts(task, store_parts):
             "analysis",
             "calibrator", "calibrators", "calib",
             "selector", "sel",
-            "producer", "producers", "prod",
+            "config", "configs",
+            "producers", "prod",
             "ml_data", "ml_model", "ml_models",
             "weightprod", "inf_model",
             "task_family",
-            "config", "dataset", "shift",
+            "calibrator", "producer",
+            "shift", "dataset",
         ]
         parts_order_end = ["version"]
 

hbw/analysis/processes.py

@@ -0,0 +1,83 @@
+# coding: utf-8
+
+"""
+Creation and modification of processes in the HH -> bbWW analysis.
+NOTE: it is crucial to modify processes before the campaign is created. Otherwise,
+the changes will not be reflected in the campaign and there will be inconsistencies.
+"""
+
+# import order as od
+
+
+from hbw.config.processes import create_parent_process
+from hbw.config.styling import color_palette
+from cmsdb.util import add_decay_process
+
+
+def modify_cmsdb_processes():
+    from cmsdb.processes import (
+        qcd_mu, qcd_em, qcd_bctoe,
+        tt, ttv, st, w_lnu, vv, h,
+        dy, dy_m4to10, dy_m10to50, dy_m50toinf, dy_m50toinf_0j, dy_m50toinf_1j, dy_m50toinf_2j,
+    )
+
+    qcd_mu.label = "QCD Muon enriched"
+    qcd_ele = create_parent_process(
+        [qcd_em, qcd_bctoe],
+        name="qcd_ele",
+        id=31199,
+        label="QCD Electron enriched",
+    )
+
+    v_lep = create_parent_process(
+        [w_lnu, dy],
+        name="v_lep",
+        id=64575573,  # random number
+        label="W and DY",
+    )
+
+    t_bkg = create_parent_process(
+        [st, tt, ttv],
+        name="t_bkg",
+        id=97842611,  # random number
+        label="tt + st",
+    )
+
+    background = create_parent_process(  # noqa: F841
+        [t_bkg, v_lep, vv, w_lnu, h, qcd_ele, qcd_mu],
+        name="background",
+        id=99999,
+        label="background",
+        color=color_palette["blue"],
+    )
+
+    decay_map = {
+        "lf": {
+            "name": "lf",
+            "id": 50,
+            "label": "(lf)",
+            "br": -1,
+        },
+        "hf": {
+            "name": "hf",
+            "id": 60,
+            "label": "(hf)",
+            "br": -1,
+        },
+    }
+
+    for dy_proc_inst in (
+        dy, dy_m4to10, dy_m10to50, dy_m50toinf, dy_m50toinf_0j, dy_m50toinf_1j, dy_m50toinf_2j,
+    ):
+        add_production_mode_parent = dy_proc_inst.name != "dy"
+        for flavour in ("hf", "lf"):
+            # the 'add_decay_process' function helps us to create all parent-daughter relationships
+            add_decay_process(
+                dy_proc_inst,
+                decay_map[flavour],
+                add_production_mode_parent=add_production_mode_parent,
+                name_func=lambda parent_name, decay_name: f"{parent_name}_{decay_name}",
+                label_func=lambda parent_label, decay_label: f"{parent_label} {decay_label}",
+                xsecs=None,
+                aux={"flavour": flavour},
+            )