fix bugs and unit tests

This commit is contained in:
2021-10-18 13:33:05 -07:00
parent 9a0c157ebb
commit b1bea09ad6
18 changed files with 9551 additions and 9541 deletions

View File

@@ -3,7 +3,7 @@ import os
import subprocess
from shutil import copyfile
import pandas as pd
from pandas.util.testing import assert_frame_equal
from pandas.testing import assert_frame_equal
from io import StringIO
# with / without pwr DONE
@@ -49,7 +49,7 @@ class Test_Wikipedia(unittest.TestCase):
# as a test let's make sure that we get equal data frames
test = pd.read_table(test_file)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
assert_frame_equal(test,baseline, check_like=True)
def test_WP_namespaces(self):
print(os.path.abspath('.'))
@@ -71,7 +71,7 @@ class Test_Wikipedia(unittest.TestCase):
num_wrong_ns = sum(~ test.namespace.isin({0,1}))
self.assertEqual(num_wrong_ns, 0)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
assert_frame_equal(test,baseline, check_like=True)
def test_WP_revert_radius(self):
print(os.path.abspath('.'))
@@ -93,7 +93,7 @@ class Test_Wikipedia(unittest.TestCase):
num_wrong_ns = sum(~ test.namespace.isin({0,1}))
self.assertEqual(num_wrong_ns, 0)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
assert_frame_equal(test,baseline, check_like=True)
@@ -131,7 +131,7 @@ class Test_Basic(unittest.TestCase):
test = pd.read_table(test_file)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
assert_frame_equal(test,baseline, check_like=True)
def test_collapse_user(self):
@@ -151,7 +151,7 @@ class Test_Basic(unittest.TestCase):
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
test = pd.read_table(test_file)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
assert_frame_equal(test,baseline, check_like=True)
def test_pwr_segment(self):
test_filename = "persistence_segment_" + self.wikiq_out_name
@@ -171,7 +171,7 @@ class Test_Basic(unittest.TestCase):
test = pd.read_table(test_file)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
assert_frame_equal(test,baseline, check_like=True)
def test_pwr_legacy(self):
test_filename = "persistence_legacy_" + self.wikiq_out_name
@@ -191,7 +191,7 @@ class Test_Basic(unittest.TestCase):
test = pd.read_table(test_file)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
assert_frame_equal(test,baseline, check_like=True)
def test_pwr(self):
test_filename = "persistence_" + self.wikiq_out_name
@@ -211,7 +211,9 @@ class Test_Basic(unittest.TestCase):
test = pd.read_table(test_file)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
test = test.reindex(columns=sorted(test.columns))
assert_frame_equal(test,baseline, check_like=True)
def test_url_encode(self):
@@ -230,7 +232,9 @@ class Test_Basic(unittest.TestCase):
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
test = pd.read_table(test_file)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
test = test.reindex(columns=sorted(test.columns))
assert_frame_equal(test,baseline, check_like=True)
class Test_Malformed(unittest.TestCase):
@@ -241,7 +245,7 @@ class Test_Malformed(unittest.TestCase):
self.wiki = 'twinpeaks'
self.wikiq_out_name = self.wiki + ".tsv"
self.test_output_dir = os.path.join(".", "test_output")
self.call_output = os.path.join(self.test_output_dir, self.wikiq_out_name)
self.call_output = os.path.join(self.test_output_dir, self.wikiq_out_name)
self.infile = "{0}.xml.7z".format(self.wiki)
self.base_call = "../wikiq {0} -o {1}"
@@ -273,6 +277,7 @@ class Test_Stdout(unittest.TestCase):
def test_noargs(self):
call = self.base_call.format(self.input_file)
print(call)
proc = subprocess.run(call,stdout=subprocess.PIPE,shell=True)
outs = proc.stdout.decode("utf8")
@@ -281,7 +286,7 @@ class Test_Stdout(unittest.TestCase):
print(baseline_file)
test = pd.read_table(StringIO(outs))
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
assert_frame_equal(test,baseline, check_like=True)
class Test_Regex(unittest.TestCase):
@@ -367,7 +372,7 @@ class Test_Regex(unittest.TestCase):
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test, baseline)
assert_frame_equal(test, baseline, check_like=True)
print(i)
@@ -392,7 +397,7 @@ class Test_Regex(unittest.TestCase):
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test, baseline)
assert_frame_equal(test, baseline, check_like=True)
if __name__ == '__main__':