pyeeg.features.SyntacticFeatureExtractor.extract_from_tree

SyntacticFeatureExtractor.extract_from_tree(tree: None, features: List[str]) → Dict[str, List[int]]

Extract requested features from a single parse tree.

Parameters:

• tree (nltk.tree.Tree) – Constituency parse tree.

• features (list of str) –

  Feature names to extract. Valid values are:

  • "depth": depth of each leaf via depth_single_tree()

  • "opening": opening value of each leaf via opening_single_tree()

  • "closing": closing value of each leaf via closing_single_tree()

  • "tree_height": height of the whole tree, repeated once per leaf

  • "all": all of the above features

Returns:

result – Mapping from feature name to the per-leaf values, in leaf (word) order. Each list has one entry per leaf of tree. Unknown feature names are silently ignored.

Return type:

dict of str -> list of int