https://github.com/arne-cl/spade-parser
mirror of SPADE (Sentence-level PArsing for DiscoursE, Soricut and Marcu 2003)
https://github.com/arne-cl/spade-parser
parser rhetorical-structure-theory rst
Last synced: 10 months ago
JSON representation
mirror of SPADE (Sentence-level PArsing for DiscoursE, Soricut and Marcu 2003)
- Host: GitHub
- URL: https://github.com/arne-cl/spade-parser
- Owner: arne-cl
- License: other
- Created: 2017-12-08T22:43:14.000Z (about 8 years ago)
- Default Branch: master
- Last Pushed: 2017-12-09T15:08:44.000Z (about 8 years ago)
- Last Synced: 2025-01-06T09:12:23.783Z (12 months ago)
- Topics: parser, rhetorical-structure-theory, rst
- Language: C
- Size: 567 KB
- Stars: 1
- Watchers: 2
- Forks: 1
- Open Issues: 1
-
Metadata Files:
- Readme: README
- License: LICENSE
Awesome Lists containing this project
README
SPADE, Release Developer 0.9
(c) 2003 University of Southern California
written by Radu Soricut
SPADE (Sentence-level PArsing for DiscoursE) is a discourse parser at sentence
level written by Radu Soricut at USC/ISI. You can find details about the
approach implemented by SPADE in the paper:
Radu Soricut and Daniel Marcu (2003).
Sentence Level Discourse Parsing using Syntactic and Lexical Information.
Proceedings of the Human Language Technology and North American Association
for Computational Linguistics Conference (HLT/NAACL), May 27-June 1,
Edmonton, Canada.
SPADE assumes as input a one-sentence-per-line file, and the output is one
discourse parse tree per sentence. The algorithm uses syntactic parse trees,
and needs to call Charniak's syntactic parser. See the bottom of this page
for instructions on how to integrate Charniak's parser with SPADE.
Both Charniak's parser and SPADE use SUN executables; SPADE also needs
Perl 5.0 (or newer).
An embbeded feature of SPADE is the discourse segmenter.
You can run SPADE with the -seg-only option; in this case, the output is
one segment (elementary discourse unit) per line, with end-of-sentence
boundaries inserted at the original (input) sentence boundaries.
To run SPADE, you need to download Charniak's syntactic parser, available at
http://www.cs.brown.edu/people/ec/#software
After you install Charniak's parser on your machine, you must edit the file
spade.pl (in the bin/ directory); in line 3, set the value of $CHP to the
directory path for Charniak's parser.
IF you need to create executables for an architecure different than
SUN, you should do the following (otherwise, skip these steps):
% cd bin/
% rm dependencies edubreak
% make
After the new executables are created, you are ready to run SPADE.
Run SPADE:
% spade.pl [-seg-only] one-sentence-per-line-file
Example:
% spade.pl ../example/wsj_0616.sent