generalised parsing of 'features' (in clustering output) during span labelling

git-svn-id: https://ws10smt.googlecode.com/svn/trunk@269 ec762483-ff6d-05da-a07a-a48fb63a330f

1 files changed, 12 insertions, 10 deletions

diff --git a/gi/pyp-topics/scripts/spans2labels.py b/gi/pyp-topics/scripts/spans2labels.py
index 3dc60835..f3968616 100755
--- a/gi/pyp-topics/scripts/spans2labels.py
+++ b/gi/pyp-topics/scripts/spans2labels.py
@@ -33,15 +33,14 @@ for line in file(sys.argv[1], 'r'):
   if len(contexts) == 1: continue
   assert len(contexts) % 2 == 0
   for i in range(0, len(contexts), 2):
-    #parse contexts[i+1] = " C=1 P=0.8 "
-    features = contexts[i+1].split()
-    category = features[0].split("=")[1].strip()
-    prob = float(features[1].split("=")[1].strip())
-    if prob >= threshold:
-      phrase_context_index[(phrase,contexts[i])] = category
-    else:
-      phrase_context_index[(phrase,contexts[i])] = cutoff_cat
-#      print (phrase,contexts[i]), category, prob
+    #parse contexts[i+1] = " C=1 P=0.8 ... "
+    features=dict([ keyval.split('=') for keyval in contexts[i+1].split()])
+    category = features['C']    
+    if features.has_key('P') and float(features['P']) < threshold:
+        category = cutoff_cat
+    
+    phrase_context_index[(phrase,contexts[i])] = category 
+#   print (phrase,contexts[i]), category, prob
 
 for line in sys.stdin:
   line_segments = line.split('|||')
@@ -87,5 +86,8 @@ for line in sys.stdin:
         context = contextt
 
     label = phrase_context_index.get((phrase,context), "<UNK>")
-    print "%d-%d-%d-%d:X%s" % (s1-order,s2-order,t1-order,t2-order,label),
+    if label != cutoff_cat: #cutoff'd spans are left unlabelled
+      print "%d-%d-%d-%d:X%s" % (s1-order,s2-order,t1-order,t2-order,label),
   print
+
+