migrate to mwxml. This completes the migration away from python-mediawiki-utilities. Except for preserving legacy persistence behavior, we can safely use the nice updates from the mediawiki-utils project.

This commit is contained in:
2018-07-05 01:16:00 -07:00
parent d77b0a4965
commit dba793c6ac
8 changed files with 49071 additions and 49083 deletions

View File

@@ -1,9 +1,10 @@
import unittest
import os
import sys
import subprocess
from shutil import copyfile
import pdb
import pandas as pd
from pandas.util.testing import assert_frame_equal
from io import StringIO
# with / without pwr DONE
# with / without url encode DONE
@@ -45,13 +46,10 @@ class Test_Wikipedia(unittest.TestCase):
copyfile(self.call_output, test_file)
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
test_lines = open(test_file)
baseline_lines = open(baseline_file)
for test, baseline in zip(test_lines, baseline_lines):
self.assertEqual(test,baseline)
test_lines.close()
baseline_lines.close()
# 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)
class Test_Basic(unittest.TestCase):
@@ -86,13 +84,10 @@ class Test_Basic(unittest.TestCase):
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
test_lines = open(test_file)
baseline_lines = open(baseline_file)
for test, baseline in zip(test_lines, baseline_lines):
self.assertEqual(test, baseline)
test = pd.read_table(test_file)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
test_lines.close()
baseline_lines.close()
def test_collapse_user(self):
test_filename = "collapse-user_" + self.wikiq_out_name
@@ -109,14 +104,9 @@ class Test_Basic(unittest.TestCase):
copyfile(self.call_output, test_file)
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
test_lines = open(test_file)
baseline_lines = open(baseline_file)
for test, baseline in zip(test_lines, baseline_lines):
self.assertEqual(test,baseline)
test_lines.close()
baseline_lines.close()
test = pd.read_table(test_file)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
def test_pwr_legacy(self):
test_filename = "persistence_legacy_" + self.wikiq_out_name
@@ -134,13 +124,30 @@ class Test_Basic(unittest.TestCase):
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
test_lines = open(test_file)
baseline_lines = open(baseline_file)
for test, baseline in zip(test_lines, baseline_lines):
self.assertEqual(test,baseline)
test = pd.read_table(test_file)
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
def test_pwr(self):
test_filename = "persistence_" + self.wikiq_out_name
test_file = os.path.join(self.test_output_dir, test_filename)
if os.path.exists(test_file):
os.remove(test_file)
call = self.base_call.format(self.input_file, self.test_output_dir)
call = call + " --persistence"
proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
proc.wait()
copyfile(self.call_output, test_file)
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_lines.close()
baseline_lines.close()
def test_url_encode(self):
test_filename = "url-encode_" + self.wikiq_out_name
@@ -156,14 +163,10 @@ class Test_Basic(unittest.TestCase):
copyfile(self.call_output, test_file)
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_lines = open(test_file)
baseline_lines = open(baseline_file)
for test, baseline in zip(test_lines, baseline_lines):
self.assertEqual(test,baseline)
test_lines.close()
baseline_lines.close()
class Test_Malformed(unittest.TestCase):
@@ -207,42 +210,14 @@ class Test_Stdout(unittest.TestCase):
call = self.base_call.format(self.input_file)
proc = subprocess.run(call,stdout=subprocess.PIPE,shell=True)
outs = proc.stdout.decode('utf-8')
outs = proc.stdout.decode("utf8")
test_file = "noargs_" + self.wikiq_out_name
baseline_file = os.path.join(".", self.baseline_output_dir, test_file)
test_lines = outs.splitlines(True)
baseline_lines = open(baseline_file)
for test, baseline in zip(test_lines, baseline_lines):
self.assertEqual(test,baseline)
def test_persistence(self):
call = self.base_call.format(self.input_file) + " --persistence"
proc = subprocess.run(call,stdout=subprocess.PIPE,shell=True)
outs = proc.stdout.decode('utf-8')
test_file = "persistence_" + self.wikiq_out_name
baseline_file = os.path.join(".", self.baseline_output_dir, test_file)
test_lines = outs.splitlines(True)
baseline_lines = open(baseline_file)
for test, baseline in zip(test_lines, baseline_lines):
self.assertEqual(test,baseline)
# test_file = "noargs_" + self.wikiq_out_name
# copyfile(self.call_output, os.path.join(self.test_output_dir, test_file))
# baseline_file = os.path.join(".", self.baseline_output_dir, test_file)
# test_lines = open(os.path.join(self.test_output_dir,test_file))
# baseline_lines = open(baseline_file)
# for test, baseline in zip(test_lines, baseline_lines):
# self.assertEqual(test,baseline)
# test_lines.close()
# baseline_lines.close()
print(baseline_file)
test = pd.read_table(StringIO(outs))
baseline = pd.read_table(baseline_file)
assert_frame_equal(test,baseline)
if __name__ == '__main__':
unittest.main()

File diff suppressed because it is too large Load Diff

File diff suppressed because it is too large Load Diff

File diff suppressed because it is too large Load Diff

File diff suppressed because it is too large Load Diff

File diff suppressed because it is too large Load Diff