collective/wikiq
20
0
Fork 0
Code Issues Pull Requests 3 Activity

write regex captures to parquet arrays.

Browse Source
This commit is contained in:
Nathan TeBlunthuis
2022-03-29 17:52:26 -07:00
parent 32283aa4da
commit b124f9c7c8
9 changed files with 359 additions and 353 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
test/Wikiq_Unit_Test.py
View File

@@ -289,12 +289,13 @@ class Test_Redirect_Ikwiki(unittest.TestCase):
proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
proc.wait()
test_file = os.path.join(self.test_output_dir, "redirect_" + self.wikiq_out_name)
test_file = os.path.join("redirect_" + self.wikiq_out_name)
copyfile(self.call_output, test_file)
baseline_file = os.path.join(".", self.baseline_output_dir, test_file)
# as a test let's make sure that we get equal data frames
test = pd.read_table(test_file)
test = pd.read_table(os.path.join(self.test_output_dir, test_file))
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline, check_like=True)
@@ -326,12 +327,12 @@ class Test_Redirect_Wikia(unittest.TestCase):
proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
proc.wait()
test_file = os.path.join(self.test_output_dir, "redirect_" + self.wikiq_out_name)
test_file = os.path.join("redirect_" + self.wikiq_out_name)
copyfile(self.call_output, test_file)
baseline_file = os.path.join(".", self.baseline_output_dir, test_file)
# as a test let's make sure that we get equal data frames
test = pd.read_table(test_file)
test = pd.read_table(os.path.join(self.test_output_dir, test_file))
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline, check_like=True)
@@ -340,7 +341,7 @@ class Test_Redirect_Wikia(unittest.TestCase):
class Test_Stdout(unittest.TestCase):
def setUp(self):
self.wiki = 'itwiki'
self.wiki = 'sailormoon'
self.wikiq_out_name = self.wiki + ".tsv"
self.infile = "{0}.xml.7z".format(self.wiki)
@@ -359,7 +360,8 @@ class Test_Stdout(unittest.TestCase):
test_file = "noargs_" + self.wikiq_out_name
baseline_file = os.path.join(".", self.baseline_output_dir, test_file)
print(baseline_file)
test = pd.read_table(StringIO(outs))
s = StringIO(outs)
test = pd.read_table(s)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline, check_like=True)
@@ -415,9 +417,9 @@ class Test_Regex(unittest.TestCase):
def test_regex_inputs(self):
for input in self.bad_inputs_list:
for inpt in self.bad_inputs_list:
call = self.base_call.format(self.input_file)
call = call + " --stdout " + input
call = call + " --stdout " + inpt
print(call)
proc = subprocess.Popen(call,stdout=subprocess.PIPE,stderr=subprocess.PIPE,shell=True)
stdout,stderr = proc.communicate()
@@ -428,7 +430,7 @@ class Test_Regex(unittest.TestCase):
self.assertNotEqual(proc.returncode,0)
def test_basic_regex(self):
for i, input in enumerate(self.good_inputs_list):
for i, inpt in enumerate(self.good_inputs_list):
test_filename = "basic_{0}_{1}.tsv".format(self.wikiq_out_name[:-4], str(i))
#print(test_filename)
@@ -437,7 +439,7 @@ 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 + " " + inpt
print(call)
proc = subprocess.Popen(call,stdout=subprocess.PIPE,stderr=subprocess.PIPE,shell=True)
@@ -453,7 +455,7 @@ class Test_Regex(unittest.TestCase):
def test_capturegroup_regex(self):
for i, input in enumerate(self.cap_inputs_list):
for i, inpt in enumerate(self.cap_inputs_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)
@@ -461,7 +463,7 @@ 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 + " " + inpt
print(call)
proc = subprocess.Popen(call,stdout=subprocess.PIPE,stderr=subprocess.PIPE,shell=True)