batchalign 0.7.6a27__tar.gz → 0.7.6a29__tar.gz

{batchalign-0.7.6a27/batchalign.egg-info → batchalign-0.7.6a29}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: batchalign
-Version: 0.7.6a27
+Version: 0.7.6a29
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu

{batchalign-0.7.6a27 → batchalign-0.7.6a29}/batchalign/pipelines/morphosyntax/ja/verbforms.py

@@ -4,6 +4,22 @@ Fix Japanese verb forms.
 """
 
 def verbform(upos, target, text):
+    if "ちゃ" in text:
+        return "sconj", "ば"
+    if "なきゃ" in text:
+        return "sconj", "なきゃ"
+    if "じゃ" in text:
+        return "sconj", "ちゃ"
+    if "れる" in text:
+        return "aux", "られる"
+    if "じゃう" in text:
+        return "aux", "ちゃう"
+    if "よう" in text:
+        return "aux", "おう"
+    if "だら" in text:
+        return "aux", "たら"
+    if "だ" in target:
+        return "aux", "た"
     if "為る" in target and 'さ' == text:
         return "part", "為る"
     if "無い" in target:

{batchalign-0.7.6a27 → batchalign-0.7.6a29}/batchalign/pipelines/morphosyntax/ud.py

@@ -258,6 +258,8 @@ def handler__VERB(word, lang=None):
     res = handler(word, lang)
     if "sconj" in res:
         return res
+    elif word.text == "ろ":
+        return res
     elif "verb" not in res and "aux" not in res:
         if word.text == "たり":
             return res+stringify_feats("Inf", "S")
@@ -893,8 +895,11 @@ def morphoanalyze(doc: Document, retokenize:bool, status_hook:callable = None, *
 
                 for i in aligned:
                     if isinstance(i, Match):
-                        if i.reference_payload not in chunks_backplate[i.payload]:
-                            chunks_backplate[i.payload].append(i.reference_payload)
+                        if not creaky:
+                            if i.reference_payload not in chunks_backplate[i.payload]:
+                                chunks_backplate[i.payload].append(i.reference_payload)
+                        else:
+                            collected += i.key
                     elif isinstance(i, Extra) and i.extra_type == ExtraType.PAYLOAD:
                         if i.key == "⁎":
                             creaky = not creaky
@@ -913,7 +918,6 @@ def morphoanalyze(doc: Document, retokenize:bool, status_hook:callable = None, *
                                 pos = "x",
                                 feats = ""
                             )]
-
                 poses = [i.morphology[0].pos.upper() for i in ut
                          if i.morphology
                          and len(i.morphology) > 0]
@@ -938,9 +942,11 @@ def morphoanalyze(doc: Document, retokenize:bool, status_hook:callable = None, *
                 retokenized_ut = retokenized_ut.replace(" ↑", "↑")
                 retokenized_ut = re.sub(r"@ ?w ?p", "@wp", retokenized_ut)
                 retokenized_ut = retokenized_ut.replace(" @", "@")
-                retokenized_ut = re.sub(r"\*[* ]*", "*", retokenized_ut)
+                # retokenized_ut = re.sub(r"\*[* ]*", "*", retokenized_ut)
                 retokenized_ut = re.sub(r"⁎[⁎ ]*(.*?)[⁎ ]*⁎", r"⁎\1⁎ ", retokenized_ut)
                 retokenized_ut = re.sub(r"\[\*(.)\]", r"[* \1]", retokenized_ut)
+                retokenized_ut = re.sub(r" +", r" ", retokenized_ut)
+
                 # pray to everyone that it works---this will simply crash and ignore
                 # the utterance if it didn't work, so we are doing this as a sanity
                 # check rather than needing the parsed result

{batchalign-0.7.6a27 → batchalign-0.7.6a29}/batchalign/version

@@ -1,3 +1,3 @@
-0.7.6-alpha.27
+0.7.6-alpha.29
 November 10, 2024
 Japanese changes + packaging

{batchalign-0.7.6a27 → batchalign-0.7.6a29/batchalign.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: batchalign
-Version: 0.7.6a27
+Version: 0.7.6a29
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu