def load_data_and_labels_lemonde(filepathXml):
"""
Load data and label from Le Monde XML corpus file
the format is ENAMEX-style, as follow:
<sentence id="E14">Les ventes de micro-ordinateurs en <ENAMEX type="Location" sub_type="Country"
eid="2000000003017382" name="Republic of France">France</ENAMEX> se sont ralenties en 1991. </sentence>
Returns:
tuple(numpy array, numpy array): data and labels
"""
# as we have XML mixed content, we need a real XML parser...
parser = make_parser()
handler = ENAMEXContentHandler()
parser.setContentHandler(handler)
parser.parse(filepathXml)
tokens = handler.getSents()
labels = handler.getAllLabels()
return tokens, labels
def load_data_and_labels_xml_file(filepathXml):
"""
Load data and label from an XML file
the format is as follow:
<p>
bla bla you are a <rs type="insult">CENSURED</rs>,
and I will <rs type="threat">find and kill</rs> you bla bla
</p>
only the insulting expression is labelled, and similarly only the threat
"action" is tagged
Returns:
tuple(numpy array, numpy array): data and labels
"""
# as we have XML mixed content, we need a real XML parser...
parser = make_parser()
handler = TEIContentHandler()
parser.setContentHandler(handler)
parser.parse(filepathXml)
tokens = handler.getSents()
labels = handler.getAllLabels()
return tokens, labels
