Tokenizes table cells
(self, node)
| 63 | self.__tokens__ = [] |
| 64 | |
| 65 | def tokenize(self, node): |
| 66 | ''' Tokenizes table cells |
| 67 | ''' |
| 68 | self.__tokens__.append('<%s>' % node.tag) |
| 69 | if node.text is not None: |
| 70 | self.__tokens__ += list(node.text) |
| 71 | for n in node.getchildren(): |
| 72 | self.tokenize(n) |
| 73 | if node.tag != 'unk': |
| 74 | self.__tokens__.append('</%s>' % node.tag) |
| 75 | if node.tag != 'td' and node.tail is not None: |
| 76 | self.__tokens__ += list(node.tail) |
| 77 | |
| 78 | def load_html_tree(self, node, parent=None): |
| 79 | ''' Converts HTML tree to the format required by apted |