batchalign 0.7.8__tar.gz → 0.7.9.post0__tar.gz

{batchalign-0.7.8/batchalign.egg-info → batchalign-0.7.9.post0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: batchalign
-Version: 0.7.8
+Version: 0.7.9.post0
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu

{batchalign-0.7.8 → batchalign-0.7.9.post0}/batchalign/cli/cli.py

@@ -158,10 +158,13 @@ def transcribe(ctx, in_dir, out_dir, lang, num_speakers, **kwargs):
     def writer(doc, output):
         doc.content.insert(0, CustomLine(id="Comment", type=CustomLineType.INDEPENDENT,
                                          content=f"Batchalign {VERSION_NUMBER.strip()}, ASR Engine {asr}"))
-        CHATFile(doc=doc).write(output.lower()
+        CHATFile(doc=doc).write(output,
                                 .replace(".wav", ".cha")
+                                .replace(".WAV", ".cha")
                                 .replace(".mp4", ".cha")
+                                .replace(".MP4", ".cha")
                                 .replace(".mp3", ".cha"),
+                                .replace(".MP3", ".cha"),
                                 write_wor=kwargs.get("wor", False))
 
     if kwargs.get("diarize"):

{batchalign-0.7.8 → batchalign-0.7.9.post0}/batchalign/constants.py

@@ -1,5 +1,5 @@
 # CHAT punctuation specifications
-ENDING_PUNCT = [".", "?", "!", "+//.", "+/.", "+...", "+\"/.", "+..?", "+\".", "+//?", "+.", "+!?", "+/?", "..."]
+ENDING_PUNCT = [".", "?", "!", "+//.", "+/.", "+...", "+\"/.", "+..?", "+\".", "+//?", "+.", "+!?", "+/?", "...", "(.)"]
 MOR_PUNCT = ["‡", "„", ","]
 CHAT_IGNORE = ["xxx", "yyy", "www"]
 

{batchalign-0.7.8 → batchalign-0.7.9.post0}/batchalign/formats/chat/generator.py

@@ -41,9 +41,9 @@ def generate_chat_utterance(utterance: Utterance, special_mor=False, write_wor=T
         gras.append(i.dependency)
         if i.time:
             has_wor = True
-            wor_elems.append(f"{i.text} \x15{str(i.time[0])}_{str(i.time[1])}\x15")
+            wor_elems.append(re.sub(r"@\w+", "", f"{i.text} \x15{str(i.time[0])}_{str(i.time[1])}\x15"))
         else:
-            wor_elems.append(i.text)
+            wor_elems.append(re.sub(r"@\w+", "", i.text))
 
         if i.coreference:
             has_coref = True

{batchalign-0.7.8 → batchalign-0.7.9.post0}/batchalign/formats/chat/parser.py

@@ -64,9 +64,11 @@ def chat_parse_utterance(text, mor, gra, wor, additional):
     # fix commas for people that don't annotate commas with a space
     to_lex = to_lex.replace(",", " ,")
 
-    to_lex = re.sub(r"\([\d.:]+\)", "", to_lex)
+    to_lex = re.sub(r"\([\d.:]+\)(?!$)", "", to_lex)
     to_lex = re.sub(r"↫.*?↫", "", to_lex)
 
+    to_lex = re.sub(r"\(.\)$", r"$END_SPC$", to_lex)
+
     # if there is a punct, move it
     for end in sorted(ENDING_PUNCT, key=len, reverse=True):
         if end in to_lex:
@@ -76,6 +78,8 @@ def chat_parse_utterance(text, mor, gra, wor, additional):
     to_lex = to_lex.replace("  ", " ")
 
     tokens = lex(to_lex)
+    if tokens[-1][0] == "END_SPC":
+        tokens = tokens[:-1] + [("(.)", TokenType.PUNCT)]
 
     # correct 0 forms
     res = []

batchalign-0.7.9.post0/batchalign/version

@@ -0,0 +1,3 @@
+0.7.9.post.0
+Janurary 11th, 2025
+minor .lower() extension patch

{batchalign-0.7.8 → batchalign-0.7.9.post0/batchalign.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: batchalign
-Version: 0.7.8
+Version: 0.7.9.post0
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu

batchalign-0.7.8/batchalign/version

@@ -1,3 +0,0 @@
-0.7.8
-Janurary 7st, 2025
-Don't use options multi