BatchalignHK 0.7.20.post11__tar.gz → 0.7.20.post12__tar.gz

{batchalignhk-0.7.20.post11 → batchalignhk-0.7.20.post12}/BatchalignHK.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: BatchalignHK
-Version: 0.7.20.post11
+Version: 0.7.20.post12
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu

{batchalignhk-0.7.20.post11 → batchalignhk-0.7.20.post12}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: BatchalignHK
-Version: 0.7.20.post11
+Version: 0.7.20.post12
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu

{batchalignhk-0.7.20.post11 → batchalignhk-0.7.20.post12}/batchalign/document.py

@@ -206,8 +206,8 @@ class Utterance(BaseModel):
     def __len__(self):
         return len(self.content)
 
-    def __str__(self):
-        if self.text != None:
+    def tostring(self, always_detokenize=False):
+        if self.text != None and not always_detokenize:
             t = self.text
         else:
             t = self._detokenize()
@@ -231,6 +231,9 @@ class Utterance(BaseModel):
 
         return t
 
+    def __str__(self):
+        return self.tostring()
+
     def __repr__(self):
         return str(self)
 

{batchalignhk-0.7.20.post11 → batchalignhk-0.7.20.post12}/batchalign/formats/chat/lexer.py

@@ -79,14 +79,16 @@ class UtteranceLexer:
             # self.__clauses.append((form.strip(), TokenType.FEAT))
         elif form.strip() in NORMAL_GROUP_MARKS:
             # basically ignore the form
-            popped = self.__clauses.pop(-1)[0]
-            if not isinstance(popped, str):
-                for i in popped:
-                    if i[0] not in CHAT_IGNORE and i[0] != "&":
-                        self.__clauses.append(i)
-            else:
-                if popped not in CHAT_IGNORE and popped[0] != "&":
-                    self.__clauses.append((popped, TokenType.REGULAR))
+            o = self.__clauses.pop(-1)
+            if len(o) <= 1 or o[1] != TokenType.FP:
+                popped = o[0]
+                if not isinstance(popped, str):
+                    for i in popped:
+                        if i[0] not in CHAT_IGNORE and i[0] != "&":
+                            self.__clauses.append(i)
+                else:
+                    if popped not in CHAT_IGNORE and popped[0] != "&":
+                        self.__clauses.append((popped, TokenType.REGULAR))
             # if isinstance(popped, str) and :
             # pass
             # self.__clauses.append((form.strip(), TokenType.FEAT))

{batchalignhk-0.7.20.post11 → batchalignhk-0.7.20.post12}/batchalign/models/resolve.py

@@ -11,7 +11,7 @@ resolver = {
         "yue": "PolyU-AngelChanLab/Cantonese-Utterance-Segmentation",
     },
     "whisper": {
-        'eng': ("talkbank/CHATWhisper-en-large-v1", "openai/whisper-large-v2"),
+        'eng': ("talkbank/CHATWhisper-en", "openai/whisper-large-v2"),
         'yue': ("alvanlii/whisper-small-cantonese", "alvanlii/whisper-small-cantonese"),
         "heb": ("ivrit-ai/whisper-large-v3", "ivrit-ai/whisper-large-v3")
     }

{batchalignhk-0.7.20.post11 → batchalignhk-0.7.20.post12}/batchalign/models/whisper/infer_asr.py

@@ -93,7 +93,7 @@ class WhisperASRModel(object):
                 stride_length_s=3,
                 device=DEVICE,
                 torch_dtype=torch.bfloat16,
-                return_timestamps="word",
+                return_timestamps=True,
             )
         except TypeError:
             self.pipe = pipeline(
@@ -104,7 +104,7 @@ class WhisperASRModel(object):
                 stride_length_s=3,
                 device=DEVICE,
                 torch_dtype=torch.float16,
-                return_timestamps="word",
+                return_timestamps=True,
             )
         L.debug("Done, initalizing processor and config...")
         processor = WhisperProcessor.from_pretrained(base)

{batchalignhk-0.7.20.post11 → batchalignhk-0.7.20.post12}/batchalign/pipelines/morphosyntax/ud.py

@@ -729,7 +729,7 @@ def morphoanalyze(doc: Document, retokenize:bool, skipmultilang:bool, status_hoo
                 pass
 
 
-# pycountry.languages.get(alpha_3=i).alpha_2 for i in lang
+    # pycountry.languages.get(alpha_3=i).alpha_2 for i in lang
 
     config = {"processors": {"tokenize": "default",
                              "pos": "default",
@@ -813,6 +813,9 @@ def morphoanalyze(doc: Document, retokenize:bool, skipmultilang:bool, status_hoo
             line_cut = i.strip(join_with_spaces=True)
         else:
             line_cut = i.strip(join_with_spaces=True)[:-len(ending)].strip()
+
+        # import ipdb
+        # ipdb.set_trace()
             # ending = ending.replace("+//", "")
 
         # if we don't have anything in line cut, just take the original

batchalignhk-0.7.20.post12/batchalign/version

@@ -0,0 +1,3 @@
+0.7.20-post.12
+July 30th, 2025
+even more parsing patches?

batchalignhk-0.7.20.post11/batchalign/version

@@ -1,3 +0,0 @@
-0.7.20-post.11
-July 30th, 2025
-even more parsing patches?