weird missing dp logic

batchalign/pipelines/utterance/ud_utterance.py

@@ -84,6 +84,7 @@ def parse_tree(subtree):
                      for i in stack]
 
 def process_ut(ut, nlp):
+
     # remove punct
     if (ut.content[-1].type == TokenType.PUNCT or
         ut.content[-1].text in ENDING_PUNCT):
@@ -142,7 +143,7 @@ def process_ut(ut, nlp):
         if isinstance(i, Match):
             matches.append(i)
         elif i.extra_type == ExtraType.REFERENCE:
-            new_refs.append(ReferenceTarget(key=i.key, payload=i.payload))
+            new_refs.append(ReferenceTarget(key=i.key, payload=i.payload if i.payload else -1))
 
     # we now sort the references based on their orignial utterance order
     matches = matches + new_refs

batchalign/version

@@ -1,3 +1,3 @@
-0.7.7-alpha.1
+0.7.7-alpha.2
 December 06, 2024
-Changing some benchmarking form handling logic
+DP logic

scratchpad.py

@@ -54,23 +54,23 @@
 
 # # ng = NgramRetraceEngine()
 # # disf = DisfluencyReplacementEngine()
-# doc = Document.new("in general, we have hair here.", lang="eng")
+doc = Document.new("I am a very large chicken indeed.", lang="eng")
 
-# forms, delim = chat_parse_utterance("in general, we have hair(stuff) here.", None, None, None, None)
-# utterance = Utterance(content=forms, delim=delim)
-# gold = Document(content=[utterance], langs=["eng"])
+forms, delim = chat_parse_utterance("I am a very large chicken indeed.", None, None, None, None)
+utterance = Utterance(content=forms, delim=delim)
+gold = Document(content=[utterance], langs=["eng"])
 
-# pipeline = BatchalignPipeline(EvaluationEngine())
-# result = pipeline(doc, gold=gold)
+pipeline = BatchalignPipeline(EvaluationEngine())
+result = pipeline(doc, gold=gold)
 
-# # pipeline = BatchalignPipeline.new("morphosyntax")
-# # result2 = pipeline(gold)
+# pipeline = BatchalignPipeline.new("morphosyntax")
+# result2 = pipeline(gold)
 
-# # print(str(CHATFile(doc=result2)))
+# print(str(CHATFile(doc=result2)))
 
 
-# result
-# print(result["diff"])
+result
+print(result["diff"])
 
 
 # # # # doc[0].content[4].text = "maman,"