def generic_visit(self, node, *args, **kwargs) -> str:
"""Called if no explicit visitor function exists for a node."""
if node is None:
return ''
content = []
try:
for child in node.contents:
value = self.visit(child, *args, **kwargs)
content.append(value)
except TypeError as e:
logs.error(e)
return ''.join(content)
def download_url(url_path: str, file_path: str):
headers = {
'Accept': '*/*',
'User-Agent': 'curl/7.64.1',
}
req = requests.get(url_path, headers=headers, stream=True)
if req.status_code != 200:
logs.error(f'Cannot make request. Result: {req.status_code:d}')
exit(1)
with open(file_path, 'wb') as file_out:
req.raw.decode_content = True
shutil.copyfileobj(req.raw, file_out)
def ebook_generate_content(self, url_path):
content = self.download_content(url_path)
bs = BeautifulSoup(content, 'html.parser')
chapter_metadata = self.config.get('metadata', True)
title_strip = self.config.get('title_prefix_trim', '')
try:
extractor = etr.create_extractor(url_path, bs)
extractor.cleanup()
transformer = etr.create_transformer(url_path,
extractor.get_content(),
self.output_dir)
transformer.transform()
self.book_maker.render_chapter(extractor,
transformer,
url_path,
self._get_saved_file_name(url_path),
metadata=chapter_metadata,
title_strip=title_strip)
except etr.ContentNotFoundError as e:
logs.error(e, url_path)
def generate(args):
try:
Colusa.generate_book(args.input)
except ConfigurationError as e:
logs.error(e)
def init(args):
try:
Colusa.generate_new_configuration(args.output)
except ConfigurationError as e:
logs.error(e)