Add script to convert from sa-extract's unnamed format to cdec's more readable named format

1 files changed, 19 insertions, 0 deletions

diff --git a/sa-extract/sa2cdec.py b/sa-extract/sa2cdec.py
new file mode 100755
index 00000000..55fb19f3
--- /dev/null
+++ b/sa-extract/sa2cdec.py
@@ -0,0 +1,19 @@
+#!/usr/bin/env python
+import sys
+
+featNames = [ line.strip() for line in open(sys.argv[1]) if not line.startswith('#') ]
+
+for line in sys.stdin:
+  try:
+    (lhs, src, tgt, feats, align) = line.strip("\n").split(' ||| ')
+  except:
+    print >>sys.stderr, 'WARNING: No alignments:', line
+    try:
+      (lhs, src, tgt, feats) = line.strip().split(' ||| ')
+      align = ''
+    except:
+      print >>sys.stderr, "ERROR: Malformed line:", line
+      raise
+  featValues = feats.split(' ')
+  namedFeats = ' '.join( name+"="+value for (name, value) in zip(featNames, featValues) )
+  print " ||| ".join( (lhs, src, tgt, namedFeats, align) )