def _child_has_text(child, text, url, text_fuzz_ratio):
child_text = child.getText().strip()
if text_match(text, child_text, text_fuzz_ratio):
parent_text = child.parent.getText().strip()
if child_text == parent_text:
return False
child.wanted_attr = None
return True
if text_match(text, get_non_rec_text(child), text_fuzz_ratio):
child.is_non_rec_text = True
child.wanted_attr = None
return True
for key, value in child.attrs.items():
if not isinstance(value, str):
continue
value = value.strip()
if text_match(text, value, text_fuzz_ratio):
child.wanted_attr = key
return True
if key in {'href', 'src'}:
full_url = urljoin(url, value)
if text == full_url:
child.wanted_attr = key
child.is_full_url = True
return True
return False
def _fetch_result_from_child(child, wanted_attr, is_full_url, url, is_non_rec_text):
if wanted_attr is None:
if is_non_rec_text:
return get_non_rec_text(child)
return child.getText().strip()
if wanted_attr not in child.attrs:
return None
if is_full_url:
return urljoin(url, child.attrs[wanted_attr])
return child.attrs[wanted_attr]
