Pine

An intuitive web interface for Treebanking.

Terminology

Parent, Head - The word depended on by a child/daughter/descendent through a dependency relation.

Child, Daughter, Descendent = The dependent of a parent/head word.

Artificial Root = The topmost node of a dependency tree. This is not a word in the sentence but is simply drawn to respect the rule that every word must have exactly one parent. See the Universal Dependencies explanation

Root Descendent = The first word descending from the artificial root, usually a verb.

File Support

Pine supports uploading treebanks in the Universal Dependencies CoNLL-U format (.conllu).

When importing a treebank with partial or invalid sentences, some words may be altered to match the CoNLL-U specification. In particular:

any orphan words (words with no parent specified) are attached to the root descendent
if multiple words share the artificial root, the first is chosen as the root descendent and all others will descend from this

Multiword Tokens

Multiword tokens are not supported. If a file is imported that contains a multiword token, it will be imported as multiple ordinary words. In this example from Universal Dependencies: 'Vámonos' would be ignored, while 'vamos' and 'nos' would be imported as separate words.

1-2    vámonos   _
1      vamos     ir
2      nos       nosotros

Empty Nodes

Empty nodes (ellipsis) have limited support. They can be imported and exported but not viewed or edited with Pine. This means you may upload and edit a CoNLL-U file containing empty nodes and they will appear in the outputted CoNLL-U file when you come to export.

If you delete a word which precedes 1 or more empty nodes, the empty nodes will be deleted as well. To use an example from Universal Dependencies:

1      Sue       Sue
2      likes     like
3      coffee    coffee
4      and       and
5      Bill      Bill
5.1    likes     like
6      tea       tea

Deleting word 5 'Bill' will also delete empty node 5.1 'likes'.

Annotation Data

Pine currently has limited support for the following annotations: Features (FEATS), Enhanced Dependencies (DEPS), Miscellaneous (MISC). These fields will be imported and exported with existing CoNLL-U data but cannot be edited.

Introduction