From 925087356b853e2099c1b60d8b757d7aa02121a9 Mon Sep 17 00:00:00 2001 From: Chris Dyer Date: Tue, 2 Oct 2012 00:19:43 -0400 Subject: cdec cleanup, remove bayesian stuff, parsing stuff --- gi/pyp-topics/scripts/tokens2classes.py | 27 --------------------------- 1 file changed, 27 deletions(-) delete mode 100755 gi/pyp-topics/scripts/tokens2classes.py (limited to 'gi/pyp-topics/scripts/tokens2classes.py') diff --git a/gi/pyp-topics/scripts/tokens2classes.py b/gi/pyp-topics/scripts/tokens2classes.py deleted file mode 100755 index 33df255f..00000000 --- a/gi/pyp-topics/scripts/tokens2classes.py +++ /dev/null @@ -1,27 +0,0 @@ -#!/usr/bin/python - -import sys - -if len(sys.argv) != 3: - print "Usage: tokens2classes.py source_classes target_classes" - exit(1) - -source_to_topics = {} -for line in open(sys.argv[1],'r'): - term,cls = line.split() - source_to_topics[term] = cls - -target_to_topics = {} -for line in open(sys.argv[2],'r'): - term,cls = line.split() - target_to_topics[term] = cls - -for line in sys.stdin: - source, target, tail = line.split(" ||| ") - - for token in source.split(): - print source_to_topics[token], - print "|||", - for token in target.split(): - print target_to_topics[token], - print "|||", tail, -- cgit v1.2.3