From 2f2ba42a1453f4a3a08f9c1ecfc53c1b1c83d550 Mon Sep 17 00:00:00 2001 From: "philblunsom@gmail.com" Date: Tue, 22 Jun 2010 20:34:00 +0000 Subject: Initial ci of gi dir git-svn-id: https://ws10smt.googlecode.com/svn/trunk@5 ec762483-ff6d-05da-a07a-a48fb63a330f --- gi/pyp-topics/scripts/map-documents.py | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) create mode 100755 gi/pyp-topics/scripts/map-documents.py (limited to 'gi/pyp-topics/scripts/map-documents.py') diff --git a/gi/pyp-topics/scripts/map-documents.py b/gi/pyp-topics/scripts/map-documents.py new file mode 100755 index 00000000..703de312 --- /dev/null +++ b/gi/pyp-topics/scripts/map-documents.py @@ -0,0 +1,20 @@ +#!/usr/bin/python + +import sys + +if len(sys.argv) != 2: + print "Usage: map-documents.py vocab-file" + exit(1) + +vocab = file(sys.argv[1], 'r').readlines() +term_dict = map(lambda x: x.strip(), vocab) + +for line in sys.stdin: + tokens = line.split() + for token in tokens: + elements = token.split(':') + if len(elements) == 1: + print "%s" % (term_dict[int(elements[0])]), + else: + print "%s:%s" % (term_dict[int(elements[0])], elements[1]), + print -- cgit v1.2.3