Skip to content

amitbeka/ucca

Repository files navigation

============================
README: ucca package
============================

UCCA is a novel linguistic framework for semantic annotation, whose details
are available at the following paper:
Universal Conceptual Cognitive Annotation (UCCA)
Omri Abend and Ari Rappoport, ACL 2013
http://homepages.inf.ed.ac.uk/oabend/papers/ucca_acl.pdf

This Python3-only package provides an API to the UCCA annotation and tools to
manipulate and process it. It's main features are conversion between different
representations of UCCA annotations, and rich objects for all of the linguistic
relations which appear in the theoretical framework (see core, layer0, layer1
and convert modules).

Installation (on Linux):
make dev-install  # creates soft links to the current files
make full-install  # copies the package to the user's python search path
run 'make help' for details


List of Modules
==============

1. classify -- provides interface to sklearn, mlpy and numpy which
enable training and evaluation of binary classifications related to UCCA
2. collins -- provides objects and parsing of Collins dictionary format
3. convert -- provides functions to convert between the UCCA objects (pythonic)
to site annotation XML, standard XML representation and text
4. core -- provides the basic objects of UCCA relations: Node, Edge, Layer
and Passage, which are the basic items to work with
5. features -- provides feature extraction from raw text
6. layer0 -- provides the text layer (layer 0) objects: Layer0 and Terminal
7. layer1 -- provides the foundational layer objects: Layer1, FoundationalNode,
PunctNode and Linkage
8. lex -- provides lexical utilities
9. postags -- provides basic POS tags
10. scenes -- provides utilities to extract and classify scenes and scene heads
11. token_eval -- provides utilities to evaluate token classification
12. util -- provides break2sentences (of a passage, based on annotation)
13. wikt -- provides wiktionary-related functionality

In addition, a scripts and tests packages are present, enabling unit-testing.

Author
------

Amit Beka: [email protected]


License
-------

This package is licensed under the GPLv3 or later license (see LICENSE.txt).

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages