Move main logic to main()

This avoids:
1) the main function running when sourcing the file
2) Creating many globally-scoped variables in the main logic

Also begin refactor of test output file logic

Signed-off-by: Will Beason <willbeason@gmail.com>

test/Wikiq_Unit_Test.py

@@ -22,9 +22,6 @@ tracemalloc.start()
 
 class Test_Wikipedia(unittest.TestCase):
     def setUp(self):
-        if not os.path.exists("test_output"):
-            os.mkdir("test_output")
-
         self.wiki = 'ikwiki-20180301-pages-meta-history'
         self.wikiq_out_name = self.wiki + ".tsv"
         self.test_output_dir = os.path.join(".", "test_output")
@@ -109,9 +106,6 @@ class Test_Wikipedia(unittest.TestCase):
 class Test_Basic(unittest.TestCase):
 
     def setUp(self):
-        if not os.path.exists("test_output"):
-            os.mkdir("test_output")
-
         self.wiki = 'sailormoon'
         self.wikiq_out_name = self.wiki + ".tsv"
         self.test_output_dir = os.path.join(".", "test_output")
@@ -318,14 +312,14 @@ class Test_Regex(unittest.TestCase):
 
         self.test_output_dir = os.path.join(".", "test_output")
         self.call_output = os.path.join(self.test_output_dir, self.wikiq_out_name)
-        # we have two base calls, one for checking inputs and the other for checking outputs
+        # we have two base calls, one for checking arguments and the other for checking outputs
         self.base_call = "../wikiq {0}"
         self.base_call_outs = "../wikiq {0} -o {1}"
 
         self.baseline_output_dir = "baseline_output"
 
-        # sample inputs for checking that bad inputs get terminated / test_regex_inputs
-        self.bad_inputs_list = [
+        # sample arguments for checking that bad arguments get terminated / test_regex_arguments
+        self.bad_arguments_list = [
             # label is missing
             "-RP '\\b\\d+\\b'",
             # number of reg and number of labels do not match
@@ -337,33 +331,33 @@ class Test_Regex(unittest.TestCase):
             "-RP '\\b\\w{3}\\b' -RPl threeletters -CP '\\b\\w{3}\\b'"
         ]
 
-        # sample inputs for checking the outcomes of good inputs / test_basic_regex
-        self.good_inputs_list = [
+        # sample arguments for checking the outcomes of good arguments / test_basic_regex
+        self.good_arguments_list = [
             "-RP '\\b\\d{3}\\b' -RPl threedigits",
             "-RP 'TestCase' -RP 'page' -RPl testcases -RPl page_word",
             "-CP 'Chevalier' -CPl chev_com -RP 'welcome to Wikipedia' -RPl wiki_welcome -CP 'Warning' -CPl warning",
             "-CP 'WP:EVADE' -CPl wp_evade"
         ]
 
-        self.cap_inputs_list = [
+        self.cap_arguments_list = [
             "-RP 'Li Chevalier' -RPl li_cheval -CP '(?P<letter>\\b[a-zA-Z]{3}\\b)|(?P<number>\\b\\d+\\b)|(?P<cat>\\bcat\\b)' -CPl three",
             "-CP '(?P<a>\\bTestCaseA\\b)|(?P<b>\\bTestCaseB\\b)|(?P<c>\\bTestCaseC\\b)|(?P<d>\\bTestCaseD\\b)' -CPl testcase -RP '(?P<npov>npov|NPOV)|(?P<neutral>neutral point of view)' -RPl npov"
         ]
 
-    def test_regex_inputs(self):
-        for input in self.bad_inputs_list:
+    def test_regex_arguments(self):
+        for arguments in self.bad_arguments_list:
             call = self.base_call.format(self.input_file)
-            call = call + " --stdout " + input
+            call = call + " --stdout " + arguments
             print(call)
             with subprocess.Popen(call, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True) as proc:
                 stdout, stderr = proc.communicate()
-                # we want to check that the bad inputs were caught and sys.exit is stopping the code
+                # we want to check that the bad arguments were caught and sys.exit is stopping the code
                 print(stderr.decode("utf-8"))
 
                 self.assertNotEqual(proc.returncode, 0)
 
     def test_basic_regex(self):
-        for i, input in enumerate(self.good_inputs_list):
+        for i, arguments in enumerate(self.good_arguments_list):
 
             test_filename = "basic_{0}_{1}.tsv".format(self.wikiq_out_name[:-4], str(i))
             # print(test_filename)
@@ -372,7 +366,7 @@ class Test_Regex(unittest.TestCase):
                 os.remove(test_file)
 
             call = self.base_call_outs.format(self.input_file, self.test_output_dir)
-            call = call + " " + input
+            call = call + " " + arguments
             print(call)
             with subprocess.Popen(call, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True) as proc:
                 proc.wait()
@@ -388,7 +382,7 @@ class Test_Regex(unittest.TestCase):
             print(i)
 
     def test_capturegroup_regex(self):
-        for i, input in enumerate(self.cap_inputs_list):
+        for i, arguments in enumerate(self.cap_arguments_list):
             test_filename = "capturegroup_{0}_{1}.tsv".format(self.wikiq_out_name[:-4], str(i))
             print(test_filename)
             test_file = os.path.join(self.test_output_dir, test_filename)
@@ -396,10 +390,9 @@ class Test_Regex(unittest.TestCase):
                 os.remove(test_file)
 
             call = self.base_call_outs.format(self.input_file, self.test_output_dir)
-            call = call + " " + input
+            call = call + " " + arguments
             print(call)
 
-            print(call)
             with subprocess.Popen(call, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True) as proc:
                 proc.wait()
                 assert (proc.returncode == 0)
@@ -414,4 +407,14 @@ class Test_Regex(unittest.TestCase):
 
 
 if __name__ == '__main__':
+    # Perform directory check and reset here as this is a one-time setup step as opposed to per-test setup.
+    if not os.path.exists("test_output"):
+        os.mkdir("test_output")
+    else:
+        # Avoid subsequent calls to tests interfering with each other.
+        # Otherwise, a test may erroneously pass if the program has no output
+        # but a previous run output what was expected.
+        for f in os.listdir("test_output"):
+            os.remove(os.path.join("test_output", f))
+
     unittest.main()