Changed flatten subsample behaviour not working on batch

docs/shapes/python/samples.py

@@ -90,10 +90,12 @@
 >>>     + "*( !(Sum(PhotonGen_isPrompt==1 && PhotonGen_pt>15 && abs(PhotonGen_eta)<2.6) > 0)) * ewknloW",
 >>>     "FilesPerJob": 5,
 >>>     "subsamples": dys,
->>>     "flatten_samples_map": lambda sname, sub: "%s" % (sub) 
->>>     # in this way flatten sampled are simply "DY_hardJets", "DY_PUJets", 
->>>     # and "DY_inclusive_rwgt"
->>>     # default flatten_samples_map is lambda sname, sub: '%s_%s' % (sname, sub)
+>>>     "flatten_samples_map": 1 
+>>>     # By default (flatten_samples_map=0), subsamples will carry the sample
+>>>     # name as prefix (e.g. DY_DY_hardJets, DY_DY_PUJets, ...). By specifying 
+>>>     # flatten_samples_map=1 the sample prefix will be dropped and subsamples
+>>>     # names will be the one appearing in the subsamples dictionary keys 
+>>>     # (for this example subsample shape names will be DY_hardJets, DY_PUJets)
 >>> }
 
 

mkShapesRDF/shapeAnalysis/latinos/LatinosUtils.py

@@ -17,10 +17,12 @@ def flatten_samples(samples):
         sample = samples[sname]
         if "subsamples" not in sample:
             continue
-        flatten_samples_map = samples[sname].get(
-            "flatten_samples_map", lambda sname, sub: "%s_%s" % (sname, sub)
+        active_fsm = samples[sname].get(
+            "flatten_samples_map", 0
         )
-
+        if active_fsm == 0:  flatten_samples_map = lambda sname, sub: "%s_%s" % (sname, sub)
+        else: flatten_samples_map = lambda sname, sub: "%s" % (sub)
+
         subsamplesmap.append((sname, []))
         for sub in sample["subsamples"]:
             new_subsample_name = flatten_samples_map(sname, sub)

mkShapesRDF/shapeAnalysis/latinos/mkDatacards.py

@@ -289,12 +289,6 @@ def makeDatacards(
 
                     card.write("-" * 100 + "\n")
                     card.write("bin         %s" % tagNameToAppearInDatacard + "\n")
-                    if len(data) == 0:
-                        self._logger.warning("no data, no fun! ")
-                        # raise RuntimeError('No Data found!')
-                        yieldsData["data"] = 0
-
-                    card.write("observation %.0f\n" % yieldsData["data"])
 
                     card.write(
                         "shapes  *           * "
@@ -305,14 +299,21 @@ def makeDatacards(
                         + "\n"
                     )
 
-                    card.write(
-                        "shapes  data_obs           * "
-                        + "shapes/histos_"
-                        + tagNameToAppearInDatacard
-                        + ".root"
-                        + "     histo_Data"
-                        + "\n"
-                    )
+                    if len(data) == 0:
+                        print("no data, no fun! ")
+                        yieldsData["data"] = 0
+
+                    else:
+                        card.write("observation %.0f\n" % yieldsData["data"])
+
+                        card.write(
+                            "shapes  data_obs           * "
+                            + "shapes/histos_"
+                            + tagNameToAppearInDatacard
+                            + ".root"
+                            + "     histo_Data"
+                            + "\n"
+                        )
 
                     #   shapes  *           * shapes/hww-19.36fb.mH125.of_vh2j_shape_mll.root     histo_$PROCESS histo_$PROCESS_$SYSTEMATIC
                     #   shapes  data_obs    * shapes/hww-19.36fb.mH125.of_vh2j_shape_mll.root     histo_Data

mkShapesRDF/shapeAnalysis/runner.py

@@ -854,9 +854,13 @@ def splitSubsamples(self):
             _sample = list(filter(lambda k: k[0] == sampleName, self.samples))[0]
             for subsample in list(_sample[6].keys()):
                 # _sample[5] is the original dict, i.e. samples[sampleName]
-                flatten_samples_map = _sample[5].get(
-                    "flatten_samples_map", lambda sname, sub: "%s_%s" % (sname, sub)
+
+                active_fsm = _sample[5].get(
+                    "flatten_samples_map", 0
                 )
+                if active_fsm == 0:  flatten_samples_map = lambda sname, sub: "%s_%s" % (sname, sub)
+                else: flatten_samples_map = lambda sname, sub: "%s" % (sub)
+
                 new_subsample_name = flatten_samples_map(sampleName, subsample)
                 self.dfs[new_subsample_name] = {}
                 for index in self.dfs[sampleName].keys():