more slapdash asr fixes

batchalign/document.py

@@ -198,9 +198,14 @@ def __str__(self):
             t = self._detokenize()
 
         t = t.replace(". . .", "+...")
+        t = t.replace(" ' ", "'")
         t = t.replace("¿", "").replace("¡", "")
         t = re.sub(r"^\+\.\.\.", "", t.strip()).strip()
-        t = re.sub(r"^\W+", "", t.strip()).strip()
+        # this is here thrice to prevent stuff from not
+        # matching once because .sub seems to only match once
+        t = re.sub(r"^[^\w\d\s<]+", "", t.strip()).strip()
+        t = re.sub(r"^[^\w\d\s<]+", "", t.strip()).strip()
+        t = re.sub(r"^[^\w\d\s<]+", "", t.strip()).strip()
         t = re.sub(r",", " , ", t.strip()).strip()
         t = re.sub(r" +", " ", t.strip()).strip()
         return t

batchalign/version

@@ -1,3 +1,3 @@
-0.7.3-beta.8
-June 14, 2024
+0.7.3-beta.9
+June 17, 2024
 more asr fixes

scratchpad.py

@@ -23,14 +23,14 @@
 ########### The Batchalign Core Test Harness ###########
 from batchalign.formats.chat.parser import chat_parse_utterance
 
-# ng = NgramRetraceEngine()
-# disf = DisfluencyReplacementEngine()
-# doc = Document.new("tu dois la manger maman, maman, à fin", lang="eng")
+ng = NgramRetraceEngine()
+disf = DisfluencyReplacementEngine()
+doc = Document.new("I ' m such an idiot", lang="eng")
 # # doc[0].content[4].text = "maman,"
 # # doc[0].content[5].text = "maman,"
-# pipe = BatchalignPipeline(ng, disf)
-# tmp = pipe(doc)
-# tmp
+pipe = BatchalignPipeline(ng, disf)
+tmp = pipe(doc)
+tmp
 
 
 # tmp[0].content