collective/mediawiki_dump_tools
16
0
Fork 0
Code Issues Pull Requests Activity

Reformat Wikiq_Unit_Test.py

Browse Source 
Separate out reformatting from editing.

Signed-off-by: Will Beason <willbeason@gmail.com>
This commit is contained in:
Will Beason 2025-05-26 15:07:39 -05:00
parent 9c5bf577e6
commit 09a84e7d11
1 changed files with 75 additions and 85 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
test/Wikiq_Unit_Test.py
View File

@ -9,6 +9,7 @@ import tracemalloc
tracemalloc.start() tracemalloc.start()
# with / without pwr DONE # with / without pwr DONE
# with / without url encode DONE # with / without url encode DONE
# with / without collapse user DONE # with / without collapse user DONE
@ -99,7 +100,6 @@ class Test_Wikipedia(unittest.TestCase):
assert_frame_equal(test, baseline, check_like=True) assert_frame_equal(test, baseline, check_like=True)
class Test_Basic(unittest.TestCase): class Test_Basic(unittest.TestCase):
def setUp(self): def setUp(self):
@ -136,7 +136,6 @@ class Test_Basic(unittest.TestCase):
baseline = pd.read_table(baseline_file) baseline = pd.read_table(baseline_file)
assert_frame_equal(test, baseline, check_like=True) assert_frame_equal(test, baseline, check_like=True)
def test_collapse_user(self): def test_collapse_user(self):
test_filename = "collapse-user_" + self.wikiq_out_name test_filename = "collapse-user_" + self.wikiq_out_name
test_file = os.path.join(self.test_output_dir, test_filename) test_file = os.path.join(self.test_output_dir, test_filename)
@ -167,7 +166,6 @@ class Test_Basic(unittest.TestCase):
proc = subprocess.Popen(call, stdout=subprocess.PIPE, shell=True) proc = subprocess.Popen(call, stdout=subprocess.PIPE, shell=True)
proc.wait() proc.wait()
copyfile(self.call_output, test_file) copyfile(self.call_output, test_file)
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename) baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
@ -187,7 +185,6 @@ class Test_Basic(unittest.TestCase):
proc = subprocess.Popen(call, stdout=subprocess.PIPE, shell=True) proc = subprocess.Popen(call, stdout=subprocess.PIPE, shell=True)
proc.wait() proc.wait()
copyfile(self.call_output, test_file) copyfile(self.call_output, test_file)
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename) baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
@ -207,7 +204,6 @@ class Test_Basic(unittest.TestCase):
proc = subprocess.Popen(call, stdout=subprocess.PIPE, shell=True) proc = subprocess.Popen(call, stdout=subprocess.PIPE, shell=True)
proc.wait() proc.wait()
copyfile(self.call_output, test_file) copyfile(self.call_output, test_file)
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename) baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
@ -218,7 +214,6 @@ class Test_Basic(unittest.TestCase):
test = test.reindex(columns=sorted(test.columns)) test = test.reindex(columns=sorted(test.columns))
assert_frame_equal(test, baseline, check_like=True) assert_frame_equal(test, baseline, check_like=True)
def test_url_encode(self): def test_url_encode(self):
test_filename = "url-encode_" + self.wikiq_out_name test_filename = "url-encode_" + self.wikiq_out_name
@ -255,9 +250,7 @@ class Test_Malformed(unittest.TestCase):
self.input_dir = "dumps" self.input_dir = "dumps"
self.input_file = os.path.join(".", self.input_dir, self.infile) self.input_file = os.path.join(".", self.input_dir, self.infile)
def test_malformed_noargs(self): def test_malformed_noargs(self):
call = self.base_call.format(self.input_file, self.test_output_dir) call = self.base_call.format(self.input_file, self.test_output_dir)
proc = subprocess.Popen(call, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True) proc = subprocess.Popen(call, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True)
proc.wait() proc.wait()
@ -265,6 +258,7 @@ class Test_Malformed(unittest.TestCase):
errlines = str(errs).split("\\n") errlines = str(errs).split("\\n")
self.assertEqual(errlines[-2], 'xml.etree.ElementTree.ParseError: no element found: line 1369, column 0') self.assertEqual(errlines[-2], 'xml.etree.ElementTree.ParseError: no element found: line 1369, column 0')
class Test_Stdout(unittest.TestCase): class Test_Stdout(unittest.TestCase):
def setUp(self): def setUp(self):
@ -278,7 +272,6 @@ class Test_Stdout(unittest.TestCase):
self.baseline_output_dir = "baseline_output" self.baseline_output_dir = "baseline_output"
def test_noargs(self): def test_noargs(self):
call = self.base_call.format(self.input_file) call = self.base_call.format(self.input_file)
print(call) print(call)
proc = subprocess.run(call, stdout=subprocess.PIPE, shell=True) proc = subprocess.run(call, stdout=subprocess.PIPE, shell=True)
@ -291,6 +284,7 @@ class Test_Stdout(unittest.TestCase):
baseline = pd.read_table(baseline_file) baseline = pd.read_table(baseline_file)
assert_frame_equal(test, baseline, check_like=True) assert_frame_equal(test, baseline, check_like=True)
class Test_Regex(unittest.TestCase): class Test_Regex(unittest.TestCase):
def setUp(self): def setUp(self):
@ -333,14 +327,11 @@ class Test_Regex(unittest.TestCase):
"-CP 'WP:EVADE' -CPl wp_evade" "-CP 'WP:EVADE' -CPl wp_evade"
] ]
self.cap_inputs_list = [ self.cap_inputs_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", "-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" "-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): def test_regex_inputs(self):
for input in self.bad_inputs_list: for input in self.bad_inputs_list:
call = self.base_call.format(self.input_file) call = self.base_call.format(self.input_file)
@ -378,7 +369,6 @@ class Test_Regex(unittest.TestCase):
assert_frame_equal(test, baseline, check_like=True) assert_frame_equal(test, baseline, check_like=True)
print(i) print(i)
def test_capturegroup_regex(self): def test_capturegroup_regex(self):
for i, input in enumerate(self.cap_inputs_list): for i, input in enumerate(self.cap_inputs_list):
test_filename = "capturegroup_{0}_{1}.tsv".format(self.wikiq_out_name[:-4], str(i)) test_filename = "capturegroup_{0}_{1}.tsv".format(self.wikiq_out_name[:-4], str(i))