Final performance tests

Author: TKasekamp

core/tests/test_perf.py

@@ -35,12 +35,12 @@ def calculate_helper_hyperopt(self, job):
 
     def test_class_randomForest(self):
         job = self.get_job()
-        job['label'] = LabelContainer(NEXT_ACTIVITY)
         add_default_config(job)
         self.calculate_helper(job)
 
-    def test_ne_randomForest(self):
+    def ne_randomForest(self):
         job = self.get_job()
+        job['label'] = LabelContainer(NEXT_ACTIVITY)
         add_default_config(job)
         self.calculate_helper(job)
 
@@ -56,7 +56,7 @@ def test_class_decision(self):
         add_default_config(job)
         self.calculate_helper(job)
 
-    def test_reg_hyperopt(self):
+    def test_class_hyperopt(self):
         job = self.get_job()
         job['label'] = LabelContainer(NEXT_ACTIVITY)
         job['hyperopt'] = {'use_hyperopt': True, 'max_evals': 10, 'performance_metric': 'f1score'}
@@ -72,7 +72,7 @@ def get_job(self):
         json["split"] = split_single()
         json["split"]["original_log_path"] = "log_cache/BPI Challenge 2017.xes.gz"
         json["method"] = "randomForest"
-        json["encoding"] = "boolean"
+        json["encoding"] = EncodingContainer(BOOLEAN, prefix_length=20)
         json["prefix_length"] = 20
         json["type"] = "regression"
         json["padding"] = 'no_padding'
@@ -89,18 +89,18 @@ def calculate_helper_hyperopt(self, job):
         calculate_hyperopt(job)
         print("Total for %s %s seconds" % (job['method'], time.time() - start_time))
 
-    def test_class_randomForest(self):
+    def test_reg_randomForest(self):
         job = self.get_job()
         add_default_config(job)
         self.calculate_helper(job)
 
-    def test_class_linear(self):
+    def test_reg_linear(self):
         job = self.get_job()
         job['method'] = 'linear'
         add_default_config(job)
         self.calculate_helper(job)
 
-    def test_class_lasso(self):
+    def test_reg_lasso(self):
         job = self.get_job()
         job['method'] = 'lasso'
         add_default_config(job)

encoders/tests/test_performance.py

@@ -61,22 +61,30 @@ def test_performance(self):
 class TestTraceLengthTime(TestCase):
     def setUp(self):
         self.label = LabelContainer(NO_LABEL)
+        start_time = time.time()
+        self.log1 = get_logs("log_cache/Sepsis Cases - Event Log.xes.gz")[0]
+        print("Total for %s %s seconds" % ("sepsis", time.time() - start_time))
+        start_time = time.time()
+        self.log2 = get_logs("log_cache/financial_log.xes.gz")[0]
+        print("Total for %s %s seconds" % ("financial", time.time() - start_time))
+        start_time = time.time()
+        self.log3 = get_logs("log_cache/BPI Challenge 2017.xes.gz")[0]
+        print("Total for %s %s seconds" % ("2017", time.time() - start_time))
 
-    def do_test(self, encoding, log_path):
+    def do_test(self, encoding, log):
         start_time = time.time()
-        log = get_logs(log_path)[0]
+        # log = get_logs(log_path)[0]
         add_col = get_global_event_attributes(log)
         event_names = unique_events(log)
         encoding = EncodingContainer(encoding, prefix_length=20, padding=ZERO_PADDING)
         log = encode_label_log(log, encoding, REGRESSION, self.label, event_names=event_names,
                                additional_columns=add_col)
         print(log.shape)
-        print("Total %s for %s %s seconds" % (log_path, encoding.method, time.time() - start_time))
+        print("Total for %s %s seconds" % (encoding.method, time.time() - start_time))
 
     def test_performance(self):
         encs = [SIMPLE_INDEX, BOOLEAN, FREQUENCY, COMPLEX, LAST_PAYLOAD]
-        logs = ["log_cache/Sepsis Cases - Event Log.xes.gz", "log_cache/financial_log.xes.gz",
-                "log_cache/BPI Challenge 2017.xes.gz"]
+        logs = [self.log3]
         for l in logs:
             for e in encs:
                 self.do_test(e, l)