def test_clean_file(self):
current_dir = os.path.dirname(__file__)
artifacts_dir = os.path.join(current_dir, self._artifacts_dir_name)
extensions = [p.InputFileType.json, p.InputFileType.text]
for ext in extensions:
full_input_path = os.path.join(artifacts_dir, "clean_file_sample" + ext)
raw_data = p.get_file_contents(full_input_path)
self.assertIsNotNone(raw_data)
# Test all option
check_against = self._get_test_data_for_option(raw_data)
self._test_clean_file(full_input_path, check_against)
# Test individual options
options = [
p.OPT.URL,
p.OPT.MENTION,
p.OPT.HASHTAG,
p.OPT.RESERVED,
p.OPT.EMOJI,
p.OPT.SMILEY,
p.OPT.NUMBER
]
for opt in options:
check_against = self._get_test_data_for_option(raw_data, opt)
self._test_clean_file(full_input_path, check_against, opt)
def test_write_to_text_file(self):
# Test file was created
output_path = self._write_test_contents_to_cur_dir(p.InputFileType.text)
self.assertTrue(os.path.exists(output_path))
file_contents = p.get_file_contents(output_path)
# Check the contents of written file is same as the one defined in this class
p.are_lists_equal(file_contents, self._test_file_contents)
def _test_clean_file(self, full_input_path, check_against, *options):
output_path = p.clean_file(full_input_path, True, options)
self.assertTrue(os.path.exists(output_path))
clean_content = p.get_file_contents(output_path)
p.are_lists_equal(clean_content, check_against)