def test_encode(self):
n1 = struct.Node('texttext', [struct.Node(node_type='t')],
['1', '2', '3'])
n2 = struct.Node(node_type='someType', title='Some Title')
enc = struct.NodeEncoder(sort_keys=True)
self.assertEqual(
enc.encode(n1),
enc.encode({
'node_type':
struct.Node.REGTEXT,
'text':
'texttext',
'children': [{
'node_type': 't',
'text': '',
'children': [],
'label': []
}],
'label': ['1', '2', '3']
}))
self.assertEqual(
enc.encode(n2),
enc.encode({
'node_type': 'someType',
'text': '',
'children': [],
'label': [],
'title': 'Some Title'
}))
# include the period
next_char = node_text[first_p + 1:first_p + 2]
if next_char in (')', u'”'):
first_sentence = node_text[:first_p + 2]
else:
first_sentence = node_text[:first_p + 1]
else:
first_sentence = node_text
# Key terms can't be the entire text of a leaf node
if first_sentence == node_text and not node.children:
return
words = first_sentence.split()
if (not words[-1] == part_end and not first_sentence.startswith('![')):
num_words = len(words)
# key terms are short
if num_words <= 15:
layer_element = {"key_term": first_sentence, "locations": [0]}
layer[label_id] = [layer_element]
if __name__ == "__main__":
# Use the plain text based JSON for the regulation.
tree = api_stub.get_regulation_as_json(
'/vagrant/data/stub-server/regulation/1005/2013-10604-eregs')
struct.walk(tree, generate_keyterm)
print struct.NodeEncoder().encode(layer)
def as_json(self):
""" Write out the changes. """
return struct.NodeEncoder().encode(self.changes)