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

Begin refactor of tests to make new tests easier to write

Browse Source 
Handle file naming logic centrally rather than requiring a dedicated
class per input file.

Signed-off-by: Will Beason <willbeason@gmail.com>
This commit is contained in:
Will Beason 2025-05-28 09:11:36 -05:00
parent c8b14c3303
commit f3e6cc9392
1 changed files with 38 additions and 29 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

67
test/Wikiq_Unit_Test.py
View File

@ -14,6 +14,7 @@ WIKIQ: Final[str] = os.path.join(os.path.dirname(TEST_DIR), "wikiq")
TEST_OUTPUT_DIR: Final[str] = os.path.join(TEST_DIR, "test_output")
BASELINE_DIR: Final[str] = os.path.join(TEST_DIR, "baseline_output")
IKWIKI: Final[str] = "ikwiki-20180301-pages-meta-history"
def setup():
tracemalloc.start()
@ -34,6 +35,23 @@ def setup():
setup()
class WikiqTester:
def __init__(self,
wiki: str,
case_name: str,
out_format: str = "tsv",
):
self.input_file = os.path.join(TEST_DIR, "dumps", "{0}.xml.bz2".format(wiki))
self.wikiq_out_name = "{0}.{1}".format(wiki, out_format)
self.call_output = os.path.join(TEST_OUTPUT_DIR, self.wikiq_out_name)
self.baseline_file = os.path.join(BASELINE_DIR, "{0}_{1}".format(case_name, self.wikiq_out_name))
self.test_file = os.path.join(TEST_OUTPUT_DIR, "{0}_{1}".format(case_name, self.wikiq_out_name))
if os.path.exists(self.test_file):
os.remove(self.test_file)
def call_wikiq(input_file: str, *args: str, out: bool = True):
if out:
call = ' '.join([WIKIQ, input_file, "-o", TEST_OUTPUT_DIR, *args])
@ -69,67 +87,58 @@ def tmp_test_file(name: str) -> (str, str):
# wikia and wikipedia data DONE
# malformed xmls DONE
class Test_Wikipedia(unittest.TestCase):
def setUp(self):
wiki = 'ikwiki-20180301-pages-meta-history'
self.wikiq_out_name = "{0}.tsv".format(wiki)
self.call_output = os.path.join(TEST_OUTPUT_DIR, self.wikiq_out_name)
infile = "{0}.xml.bz2".format(wiki)
input_dir = os.path.join(TEST_DIR, "dumps")
self.input_file = os.path.join(TEST_DIR, input_dir, infile)
class TestWikipedia(unittest.TestCase):
def test_WP_url_encode(self):
baseline_file, test_file = tmp_test_file("url-encode_" + self.wikiq_out_name)
tester = WikiqTester(IKWIKI, "url-encode")
try:
call_wikiq(self.input_file, "--url-encode")
call_wikiq(tester.input_file, "--url-encode")
except subprocess.CalledProcessError as exc:
self.fail(exc.stderr.decode("utf8"))
copyfile(self.call_output, test_file)
copyfile(tester.call_output, tester.test_file)
# 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)
test = pd.read_table(tester.test_file)
baseline = pd.read_table(tester.baseline_file)
assert_frame_equal(test, baseline, check_like=True)
def test_WP_namespaces(self):
baseline_file, test_file = tmp_test_file("namespaces_" + self.wikiq_out_name)
tester = WikiqTester(IKWIKI, "namespaces")
try:
call_wikiq(self.input_file, "-n 0", "-n 1")
call_wikiq(tester.input_file, "-n 0", "-n 1")
except subprocess.CalledProcessError as exc:
self.fail(exc.stderr.decode("utf8"))
copyfile(self.call_output, test_file)
copyfile(tester.call_output, tester.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(tester.test_file)
num_wrong_ns = sum(~ test.namespace.isin({0, 1}))
self.assertEqual(num_wrong_ns, 0)
baseline = pd.read_table(baseline_file)
baseline = pd.read_table(tester.baseline_file)
assert_frame_equal(test, baseline, check_like=True)
def test_WP_revert_radius(self):
baseline_file, test_file = tmp_test_file("revert_radius_" + self.wikiq_out_name)
tester = WikiqTester(IKWIKI, "revert_radius")
try:
call_wikiq(self.input_file, "-n 0", "-n 1", "-rr 1")
call_wikiq(tester.input_file, "-n 0", "-n 1", "-rr 1")
except subprocess.CalledProcessError as exc:
self.fail(exc.stderr.decode("utf8"))
copyfile(self.call_output, test_file)
copyfile(tester.call_output, tester.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(tester.test_file)
num_wrong_ns = sum(~ test.namespace.isin({0, 1}))
self.assertEqual(num_wrong_ns, 0)
baseline = pd.read_table(baseline_file)
baseline = pd.read_table(tester.baseline_file)
assert_frame_equal(test, baseline, check_like=True)
class Test_Basic(unittest.TestCase):
class TestBasic(unittest.TestCase):
def setUp(self):
wiki = 'sailormoon'
@ -228,7 +237,7 @@ class Test_Basic(unittest.TestCase):
assert_frame_equal(test, baseline, check_like=True)
class Test_Malformed(unittest.TestCase):
class TestMalformed(unittest.TestCase):
def setUp(self):
wiki = 'twinpeaks'
@ -248,7 +257,7 @@ class Test_Malformed(unittest.TestCase):
self.fail("No exception raised, want: {}".format(want_exception))
class Test_Stdout(unittest.TestCase):
class TestStdout(unittest.TestCase):
def setUp(self):
wiki = 'sailormoon'
@ -276,7 +285,7 @@ class Test_Stdout(unittest.TestCase):
assert_frame_equal(test, baseline, check_like=True)
class Test_Regex(unittest.TestCase):
class TestRegex(unittest.TestCase):
def setUp(self):
wiki = 'regextest'
self.wikiq_out_name = wiki + '.tsv'