conllreader

Build syntactic trees from CoNLL parser output

exception conllreader.ConllInvalidPositionError(bad_root, max_root)

Bases: builtins.Exception

Trying to build a subtree from a node that does not exist

Variables:

• bad_root – integer, the position from which we attempted to build a subtree
• max_root – integer, the last valid position

class conllreader.ConllSemanticAppender(syntactic_conll_file)

Bases: builtins.object

Appends semantic information at the “right” of a ConLL file.

The input is a syntactic ConLL file, and the output a so-called semantic CoNLL file.

add_frame_annotation(frame_annotation)

add_new_column(sentence_id)

dump_semantic_file(filename)

conllreader.LongestCommonSubstring(S1, S2)

class conllreader.SyntacticTreeBuilder(conll_tree)

Bases: builtins.object

Wrapper class for the building of a syntactic tree

Variables:

• node_dict – every SyntacticTreeNode available by CoNLL word id
• father_ids – every dependency relation: child id -> father id
• tree_list – every root node, that is every root subtree
• sentence – the “sentence” (words separated by spaces)

fill_begin_end(node)

Fill begin/end values of very subtree

class conllreader.SyntacticTreeNode(word_id, word, pos, deprel, begin_word)

Bases: builtins.object

A node (internal or terminal) of a syntactic tree

Variables:

• word_id – int, the CoNLL word id (starts at 1)
• word – string, the word contained by the node
• pos – part-of-speech of the node
• deprel – string, function attributed by the parser to this word
• father – SyntacticTreeBuilder, the father of this node
• children – SyntacticTreeNode list, the children of this node
• begin – int, the character position this phrase starts at (root would be 0)
• end – int, the position this phrase ends at (root would be last character)
• begin_word – int, the position this word begins at

_closest_match_as_node_lcs(arg)

closest_match(arg)

Search the closest match to arg

closest_match_as_node(arg)

contains(arg)

Search an exact argument in all subtrees

flat()

Return the tokenized sentence from the parse tree.

class conllreader.TreeBuilderTest(methodName='runTest')

Bases: unittest.case.TestCase

setUp()

test_another_flat()

test_lima_tree()

test_none_fathers()

test_tree_contains()

test_tree_flat()

test_tree_match()

test_tree_str()