fix sporadic depth error (#77)

* fix sporadic depth error * update max_depth calc per meeting discussion * redo max_depth calc per gh discussion * update tests for new max_depth procedure * Apply suggestions from code review Co-authored-by: Ezio Melotti <[email protected]> * Apply suggestions from code review 2 Co-authored-by: Ezio Melotti <[email protected]> * Specify default type of `depth_base`. Co-authored-by: Ezio Melotti <[email protected]>
kstaats · Jul 6, 2022 · db2c47a · db2c47a
1 parent b5d1ada
commit db2c47a
Show file tree

Hide file tree

Showing 11 changed files with 101 additions and 107 deletions.
diff --git a/karoo-gp.py b/karoo-gp.py
@@ -332,9 +332,9 @@
                          '(c)lassification, or (m)atching')
     ap.add_argument('-typ', action='store', dest='type', default='r',
                     help='[f,g,r] Tree type: (f)ull, (g)row, or (r)amped half/half')
-    ap.add_argument('-bas', action='store', dest='depth_base', default=4,
+    ap.add_argument('-bas', action='store', dest='depth_base', type=int, default=4,
                     help='[3...10] maximum Tree depth for the initial population')
-    ap.add_argument('-max', action='store', dest='depth_max', default=4,
+    ap.add_argument('-max', action='store', dest='depth_max', type=int, default=None,
                     help='[3...10] maximum Tree depth for the entire run')
     ap.add_argument('-min', action='store', dest='depth_min', default=3,
                     help='minimum nodes, from 3 to 2^(base_depth +1) - 1')
@@ -368,8 +368,8 @@
     # pass the argparse defaults and/or user inputs to the required variables
     kernel = str(args.kernel)
     tree_type = str(args.type)
-    tree_depth_base = int(args.depth_base)
-    tree_depth_max = int(args.depth_max)
+    tree_depth_base = args.depth_base
+    tree_depth_max = args.depth_max
     tree_depth_min = int(args.depth_min)
     tree_pop_max = int(args.pop_max)
     gen_max = int(args.gen_max)

diff --git a/karoo_gp/base_class.py b/karoo_gp/base_class.py
@@ -105,7 +105,7 @@ class BaseGP(object):
     """
 
     def __init__(
-        self, tree_type='r', tree_depth_base=3, tree_depth_max=3,
+        self, tree_type='r', tree_depth_base=3, tree_depth_max=None,
         tree_depth_min=1, tree_pop_max=100, gen_max=10, tourn_size=7,
         filename='', output_dir='', evolve_repro=0.1, evolve_point=0.1,
         evolve_branch=0.2, evolve_cross=0.6, display='s', precision=None,
@@ -159,7 +159,15 @@ def __init__(
         self.X_hash = None                   # hash of last-used fit data
         self.gen_max = gen_max               # number of generations to evolve
         self.swim = swim                     # culling method
-        self.tree_depth_max = tree_depth_max # max allowed depth
+
+        if tree_depth_max is None:
+            self.tree_depth_max = tree_depth_base
+        elif tree_depth_max >= tree_depth_base:
+            self.tree_depth_max = tree_depth_max
+        else:
+            raise ValueError(f'Max depth ({tree_depth_max}) must be greater '
+                             f'or equal to base depth ({tree_depth_base})')
+
         self.tree_depth_min = tree_depth_min # min allowed number of nodes
         self.tourn_size = tourn_size         # number of Trees per tournament
         self.evolve_repro = evolve_repro     # ratio of next_gen reproduced

diff --git a/karoo_gp/test/data/results[c-f].json b/karoo_gp/test/data/results[c-f].json
@@ -1,13 +1,13 @@
 {
     "package": "Karoo GP",
-    "launched": "2022-06-12_06-15-25-662209",
+    "launched": "2022-07-06_20-39-53-790684",
     "dataset": "karoo_gp/files/data_CLASSIFY.csv",
     "config": {
         "kernel": "c",
         "precision": 6,
         "tree_type": "f",
         "tree_depth_base": 3,
-        "tree_depth_max": 4,
+        "tree_depth_max": 3,
         "min_node_count": 3,
         "genetic_operators": {
             "reproduction": 0.1,
@@ -21,41 +21,41 @@
     },
     "outcome": "SUCCESS",
     "fittest_tree": {
-        "id": 5,
-        "expression": "pl**2/sw - sl - pw/(pl**2*sw)"
+        "id": 10,
+        "expression": "pl*sl/sw - pl/pw - sl"
     },
     "score": {
-        "fitness": 26.0,
+        "fitness": 23.0,
         "classification_report": {
             "0.0": {
-                "precision": 1.0,
+                "precision": 0.6,
                 "recall": 1.0,
-                "f1-score": 1.0,
+                "f1-score": 0.7499999999999999,
                 "support": 9
             },
             "1.0": {
                 "precision": 1.0,
-                "recall": 0.6666666666666666,
-                "f1-score": 0.8,
+                "recall": 0.4166666666666667,
+                "f1-score": 0.5882352941176471,
                 "support": 12
             },
             "2.0": {
-                "precision": 0.6923076923076923,
+                "precision": 0.9,
                 "recall": 1.0,
-                "f1-score": 0.8181818181818181,
+                "f1-score": 0.9473684210526316,
                 "support": 9
             },
-            "accuracy": 0.8666666666666667,
+            "accuracy": 0.7666666666666667,
             "macro avg": {
-                "precision": 0.8974358974358975,
-                "recall": 0.8888888888888888,
-                "f1-score": 0.8727272727272727,
+                "precision": 0.8333333333333334,
+                "recall": 0.8055555555555557,
+                "f1-score": 0.7618679050567595,
                 "support": 30
             },
             "weighted avg": {
-                "precision": 0.9076923076923077,
-                "recall": 0.8666666666666667,
-                "f1-score": 0.8654545454545455,
+                "precision": 0.85,
+                "recall": 0.7666666666666667,
+                "f1-score": 0.7445046439628483,
                 "support": 30
             }
         },
@@ -66,9 +66,9 @@
                 0
             ],
             [
-                0,
-                8,
-                4
+                6,
+                5,
+                1
             ],
             [
                 0,

diff --git a/karoo_gp/test/data/results[c-g].json b/karoo_gp/test/data/results[c-g].json
@@ -1,13 +1,13 @@
 {
     "package": "Karoo GP",
-    "launched": "2022-06-12_06-16-20-639880",
+    "launched": "2022-07-06_20-39-14-375734",
     "dataset": "karoo_gp/files/data_CLASSIFY.csv",
     "config": {
         "kernel": "c",
         "precision": 6,
         "tree_type": "g",
         "tree_depth_base": 3,
-        "tree_depth_max": 4,
+        "tree_depth_max": 3,
         "min_node_count": 3,
         "genetic_operators": {
             "reproduction": 0.1,
@@ -21,41 +21,41 @@
     },
     "outcome": "SUCCESS",
     "fittest_tree": {
-        "id": 9,
-        "expression": "pl*pw**2 + pw*sw - sl"
+        "id": 10,
+        "expression": "2*pl - sw**2 + 1"
     },
     "score": {
-        "fitness": 19.0,
+        "fitness": 20.0,
         "classification_report": {
             "0.0": {
-                "precision": 1.0,
+                "precision": 0.9,
                 "recall": 1.0,
-                "f1-score": 1.0,
+                "f1-score": 0.9473684210526316,
                 "support": 9
             },
             "1.0": {
                 "precision": 1.0,
-                "recall": 0.08333333333333333,
-                "f1-score": 0.15384615384615385,
+                "recall": 0.16666666666666666,
+                "f1-score": 0.2857142857142857,
                 "support": 12
             },
             "2.0": {
-                "precision": 0.45,
+                "precision": 0.5,
                 "recall": 1.0,
-                "f1-score": 0.6206896551724138,
+                "f1-score": 0.6666666666666666,
                 "support": 9
             },
-            "accuracy": 0.6333333333333333,
+            "accuracy": 0.6666666666666666,
             "macro avg": {
-                "precision": 0.8166666666666668,
-                "recall": 0.6944444444444443,
-                "f1-score": 0.5915119363395225,
+                "precision": 0.7999999999999999,
+                "recall": 0.7222222222222223,
+                "f1-score": 0.633249791144528,
                 "support": 30
             },
             "weighted avg": {
-                "precision": 0.8350000000000001,
-                "recall": 0.6333333333333333,
-                "f1-score": 0.5477453580901857,
+                "precision": 0.8200000000000001,
+                "recall": 0.6666666666666666,
+                "f1-score": 0.5984962406015039,
                 "support": 30
             }
         },
@@ -66,9 +66,9 @@
                 0
             ],
             [
-                0,
                 1,
-                11
+                2,
+                9
             ],
             [
                 0,

diff --git a/karoo_gp/test/data/results[c-r].json b/karoo_gp/test/data/results[c-r].json
@@ -1,13 +1,13 @@
 {
     "package": "Karoo GP",
-    "launched": "2022-06-12_06-16-39-210454",
+    "launched": "2022-07-06_20-38-32-741216",
     "dataset": "karoo_gp/files/data_CLASSIFY.csv",
     "config": {
         "kernel": "c",
         "precision": 6,
         "tree_type": "r",
         "tree_depth_base": 3,
-        "tree_depth_max": 4,
+        "tree_depth_max": 3,
         "min_node_count": 3,
         "genetic_operators": {
             "reproduction": 0.1,
@@ -21,41 +21,41 @@
     },
     "outcome": "SUCCESS",
     "fittest_tree": {
-        "id": 3,
-        "expression": "pl + pw - sl"
+        "id": 6,
+        "expression": "pw - 1"
     },
     "score": {
-        "fitness": 18.0,
+        "fitness": 27.0,
         "classification_report": {
             "0.0": {
-                "precision": 0.5625,
+                "precision": 1.0,
                 "recall": 1.0,
-                "f1-score": 0.72,
+                "f1-score": 1.0,
                 "support": 9
             },
             "1.0": {
-                "precision": 0.5,
-                "recall": 0.25,
-                "f1-score": 0.3333333333333333,
+                "precision": 0.9090909090909091,
+                "recall": 0.8333333333333334,
+                "f1-score": 0.8695652173913043,
                 "support": 12
             },
             "2.0": {
-                "precision": 0.75,
-                "recall": 0.6666666666666666,
-                "f1-score": 0.7058823529411765,
+                "precision": 0.8,
+                "recall": 0.8888888888888888,
+                "f1-score": 0.8421052631578948,
                 "support": 9
             },
-            "accuracy": 0.6,
+            "accuracy": 0.9,
             "macro avg": {
-                "precision": 0.6041666666666666,
-                "recall": 0.6388888888888888,
-                "f1-score": 0.5864052287581699,
+                "precision": 0.903030303030303,
+                "recall": 0.9074074074074074,
+                "f1-score": 0.9038901601830663,
                 "support": 30
             },
             "weighted avg": {
-                "precision": 0.59375,
-                "recall": 0.6,
-                "f1-score": 0.5610980392156864,
+                "precision": 0.9036363636363636,
+                "recall": 0.9,
+                "f1-score": 0.9004576659038901,
                 "support": 30
             }
         },
@@ -66,14 +66,14 @@
                 0
             ],
             [
-                7,
-                3,
+                0,
+                10,
                 2
             ],
             [
                 0,
-                3,
-                6
+                1,
+                8
             ]
         ]
     }

diff --git a/karoo_gp/test/data/results[m-f].json b/karoo_gp/test/data/results[m-f].json
@@ -1,13 +1,13 @@
 {
     "package": "Karoo GP",
-    "launched": "2022-06-12_06-17-46-867022",
+    "launched": "2022-07-06_20-37-24-150346",
     "dataset": "karoo_gp/files/data_MATCH.csv",
     "config": {
         "kernel": "m",
         "precision": 6,
         "tree_type": "f",
         "tree_depth_base": 3,
-        "tree_depth_max": 4,
+        "tree_depth_max": 3,
         "min_node_count": 3,
         "genetic_operators": {
             "reproduction": 0.1,
@@ -19,12 +19,5 @@
         "population": 50,
         "number_of_generations": 10
     },
-    "outcome": "SUCCESS",
-    "fittest_tree": {
-        "id": 45,
-        "expression": "-a + 3*b + c - 2"
-    },
-    "score": {
-        "fitness": 10.0
-    }
+    "outcome": "FAILURE"
 }
diff --git a/karoo_gp/test/data/results[m-g].json b/karoo_gp/test/data/results[m-g].json
@@ -1,13 +1,13 @@
 {
     "package": "Karoo GP",
-    "launched": "2022-06-12_06-18-20-189034",
+    "launched": "2022-07-06_20-36-33-411609",
     "dataset": "karoo_gp/files/data_MATCH.csv",
     "config": {
         "kernel": "m",
         "precision": 6,
         "tree_type": "g",
         "tree_depth_base": 3,
-        "tree_depth_max": 4,
+        "tree_depth_max": 3,
         "min_node_count": 3,
         "genetic_operators": {
             "reproduction": 0.1,
@@ -21,7 +21,7 @@
     },
     "outcome": "SUCCESS",
     "fittest_tree": {
-        "id": 25,
+        "id": 27,
         "expression": "a + b + c"
     },
     "score": {