Fix normalize error

analysis_mysql.py

@@ -20,8 +20,8 @@ models_ensemble = ["odte", "adaBoost", "bagging", "TBRaF", "TBRoF", "TBRRoF"]
 description = ["samp", "var", "cls"]
 complexity = ["nodes", "leaves", "depth"]
 title = "Best model results"
-lengths_acc = [30, 4, 3, 3, 3, 3, 3, 12, 12, 12, 12, 12, 12, 12]
-lengths_time = [30, 4, 3, 3, 3, 3, 3, 17, 17, 17, 17, 17, 17, 17]
+lengths_acc = [30, 4, 3, 3, 6, 6, 5, 12, 12, 12, 12, 12, 12, 12]
+lengths_time = [30, 4, 3, 3, 6, 6, 5, 17, 17, 17, 17, 17, 17, 17]
 
 
 def parse_arguments() -> Tuple[str, str, str, bool, bool]:
@@ -59,9 +59,10 @@ def parse_arguments() -> Tuple[str, str, str, bool, bool]:
     ap.add_argument(
         "-o",
         "--compare",
-        type=bool,
+        type=int,
         required=False,
-        default=False,
+        default=1,
+        help="1=stree optimized, 2=stree_default, 3=both",
     )
     ap.add_argument(
         "-i",
@@ -176,8 +177,10 @@ def report_header(title, experiment, model_type):
 
 def report_line(line):
     output = f"{line['dataset']:{lengths[0] + 5}s} "
-    for key, item in enumerate(description + complexity):
+    for key, item in enumerate(description):
         output += f"{line[item]:{lengths[key + 1]}d} "
+    for key, item in enumerate(complexity):
+        output += f"{line[item]:{lengths[key + len(description) + 1]}.2f} "
     data = models.copy()
     for key, model in enumerate(data):
         output += f"{line[model]:{lengths[key + 7]}s} "
@@ -215,17 +218,22 @@ fields = (
     "Samp",
     "Var",
     "Cls",
-    "Nod",
-    "Lea",
-    "Dep",
+    "Nodes",
+    "Leave",
+    "Depth",
 )
 lengths = lengths_time if time_info else lengths_acc
+reference_model = "stree"
 if tex_output:
-    # We need the stree_std column for the tex output
-    compare = True
-if not compare:
+    # We need the stree & stree_std column for the tex output
+    compare = 3
+if compare != 3:
     # remove stree_default from fields list and lengths
-    models_tree.pop(1)
+    if compare == 1:
+        models_tree.pop(1)
+    else:
+        models_tree.pop(0)
+        reference_model = "stree_default"
     lengths.pop(7)
 models = models_tree if model_type == "tree" else models_ensemble
 for item in models:
@@ -256,16 +264,16 @@ for number, dataset in enumerate(dt):
     max_accuracy = (
         0.0 if record is None else record[9] if time_info else record[5]
     )
-    line["nodes"] = 0
-    line["leaves"] = 0
-    line["depth"] = 0
+    line["nodes"] = 0.0
+    line["leaves"] = 0.0
+    line["depth"] = 0.0
     line_tex = line.copy()
     for column, model in enumerate(models):
         record = dbh.find_best(dataset[0], model, experiment)
         if record is None:
             line[model] = color + "-" * 12
         else:
-            if model == "stree":
+            if model == reference_model:
                 line["nodes"] = record[12]
                 line["leaves"] = record[13]
                 line["depth"] = record[14]