From 733e1b1507d27d4f53055f740e8098f56215ab8f Mon Sep 17 00:00:00 2001 From: Chris Dyer Date: Fri, 27 Jul 2012 17:31:00 -0400 Subject: remove old suffix array extractor (use the one in python/ instead) --- sa-extract/wrap_input.py | 37 ------------------------------------- 1 file changed, 37 deletions(-) delete mode 100755 sa-extract/wrap_input.py (limited to 'sa-extract/wrap_input.py') diff --git a/sa-extract/wrap_input.py b/sa-extract/wrap_input.py deleted file mode 100755 index e859a4fd..00000000 --- a/sa-extract/wrap_input.py +++ /dev/null @@ -1,37 +0,0 @@ -#!/usr/bin/env python -import sys -import codecs -import os -import os.path -from xml.sax.saxutils import escape - -graPrefix = sys.argv[1] - -# Second argument can be a file with observable sentence-level features, -# one set of features per line (parallel with source sentences). Features are space-delimited indicator features. -obsFeatsFile = None -if len(sys.argv) == 3: - obsFeatsFilename = sys.argv[2] - obsFeatsFile = open(obsFeatsFilename) - -sys.stdin = codecs.getreader("utf-8")(sys.stdin) -sys.stdout = codecs.getwriter("utf-8")(sys.stdout) - -i = 0 -for line in sys.stdin: - filename = "%s%d"%(graPrefix,i) - if not os.path.exists(filename): - filenameGz = filename + ".gz" - if not os.path.exists(filenameGz): - print >>sys.stderr, "Grammar file not found: ", filename, filenameGz - sys.exit(1) - else: - filename = filenameGz - - if obsFeatsFile: - obsFeats = obsFeatsFile.next().strip() - print ' '%(i,obsFeats,filename) + escape(line.strip()) + " " - else: - print ' '%(i,filename) + escape(line.strip()) + " " - i+=1 - -- cgit v1.2.3