Dump full database in data result dir in analysis (#658)

* Dump full database in data result dir in analysis

* Create MC result dir in analyzer if not existing

Co-authored-by: Benedikt Volkel <[email protected]>

Author: benedikt-voelkel

machine_learning_hep/analysis/analyzer.py

@@ -12,14 +12,28 @@
 ##   along with this program. if not, see <https://www.gnu.org/licenses/>. ##
 #############################################################################
 
+from os.path import exists, join
+from os import makedirs
+
 # HF specific imports
 from machine_learning_hep.workflow.workflow_base import WorkflowBase
+from machine_learning_hep.io import dump_yaml_from_dict
 
 
 class Analyzer(WorkflowBase):
     def __init__(self, datap, case, typean, period):
         super().__init__(datap, case, typean, period)
 
+        # The only thing here is to dump the database in the data analysis directory
+        for mcordata in ("mc", "data"):
+            results_dir = datap["analysis"][typean][mcordata]["results"][period] \
+                    if period is not None else datap["analysis"][typean][mcordata]["resultsallp"]
+            if not exists(results_dir):
+                # create otput directories in case they do not exist
+                makedirs(results_dir)
+            if mcordata == "data":
+                dump_yaml_from_dict({case: datap},
+                                    join(results_dir, f"database_{case}_{typean}.yml"))
 
 
 class AnalyzerAfterBurner(WorkflowBase):

machine_learning_hep/analysis/analyzer_jet.py

@@ -29,7 +29,7 @@
 from ROOT import gROOT, gStyle
 from ROOT import RooUnfoldBayes
 # HF specific imports
-from machine_learning_hep.utilities import folding, equal_binning_lists, make_message_notfound, checkmakedir
+from machine_learning_hep.utilities import folding, equal_binning_lists, make_message_notfound
 from machine_learning_hep.analysis.analyzer import Analyzer
 from machine_learning_hep.utilities import setup_histogram, setup_canvas, get_colour, get_marker, get_y_window_gr, get_y_window_his, get_plot_range
 from machine_learning_hep.utilities import setup_legend, setup_tgraph, draw_latex, tg_sys
@@ -215,10 +215,6 @@ def __init__(self, datap, case, typean, period):
                 if period is not None else datap["analysis"][typean]["mc"]["resultsallp"]
         self.d_resultsallpdata = datap["analysis"][typean]["data"]["results"][period] \
                 if period is not None else datap["analysis"][typean]["data"]["resultsallp"]
-        # create otput directories in case they do not exist
-        for dir_out in [self.d_resultsallpmc, self.d_resultsallpdata]:
-            if not os.path.exists(dir_out):
-                checkmakedir(dir_out)
 
         # input directories (processor output)
         self.d_resultsallpmc_proc = self.d_resultsallpmc
@@ -245,10 +241,6 @@ def __init__(self, datap, case, typean, period):
         self.efficiency_filename = "efficiencies"
         self.sideband_subtracted_filename = "sideband_subtracted"
 
-        # Save the database in the results directory.
-        path_db_out = os.path.join(self.d_resultsallpdata, "database_%s_%s.yml" % (self.case, self.typean))
-        with open(path_db_out, "w") as file_db_out:
-            yaml.safe_dump(datap, file_db_out, default_flow_style=False)
 
     def fit(self):
         self.loadstyle()