BatchalignHK 0.7.22.post20__tar.gz → 0.7.22.post23__tar.gz

{batchalignhk-0.7.22.post20 → batchalignhk-0.7.22.post23}/BatchalignHK.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: BatchalignHK
-Version: 0.7.22.post20
+Version: 0.7.22.post23
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu

{batchalignhk-0.7.22.post20 → batchalignhk-0.7.22.post23}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: BatchalignHK
-Version: 0.7.22.post20
+Version: 0.7.22.post23
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu

{batchalignhk-0.7.22.post20 → batchalignhk-0.7.22.post23}/batchalign/pipelines/asr/oai_whisper.py

@@ -8,6 +8,10 @@ import pycountry
 import logging
 L = logging.getLogger("batchalign")
 
+from opencc import OpenCC
+cc_t2s = OpenCC('t2s')
+cc_s2hk = OpenCC('s2hk')
+
 from batchalign.utils.utils import correct_timing
 from batchalign.models import resolve
 
@@ -67,11 +71,16 @@ class OAIWhisperEngine(BatchalignEngine):
         for i in res["segments"]:
             turn = []
             for j in i["words"]:
+                word = j["word"]
+                if self.__lang == "zho":
+                    word = cc_t2s.convert(word)
+                elif self.__lang == "yue":
+                    word = cc_s2hk.convert(word)
                 turn.append({
                     "type": "text",
                     "ts": j["start"],
                     "end_ts": j["end"],
-                    "value": j["word"]
+                    "value": word
                 })
             turns.append({
                 "elements": turn,

{batchalignhk-0.7.22.post20 → batchalignhk-0.7.22.post23}/batchalign/pipelines/asr/utils.py

@@ -1,6 +1,7 @@
 import re
 from batchalign.document import *
 from batchalign.utils import *
+from batchalign.pipelines.asr.num2lang import NUM2LANG
 
 from batchalign.constants import ENDING_PUNCT
 from batchalign.pipelines.asr.num2chinese import num2chinese
@@ -238,6 +239,10 @@ def process_generation(output, lang="eng", utterance_engine=None):
                     elif lang == "yue":
                         return num2chinese(i, simp=False)
                     else:
+                        if NUM2LANG.get(lang) is not None:
+                            n2l = NUM2LANG.get(lang)
+                            for a,b in list(reversed(n2l.items())):
+                                i = i.replace(a,b).strip()
                         return i
                 except:
                     return i

{batchalignhk-0.7.22.post20 → batchalignhk-0.7.22.post23}/batchalign/pipelines/dispatch.py

@@ -144,7 +144,7 @@ def dispatch_pipeline(pkg_str, lang, num_speakers=None, **arg_overrides):
         elif engine == "gtrans":
             engines.append(GoogleTranslateEngine())
         elif engine == "whisper_oai":
-            engines.append(OAIWhisperEngine())
+            engines.append(OAIWhisperEngine(lang=lang))
         elif engine == "aliyun":
             engines.append(AliyunEngine())
         elif engine == "funaudio":

batchalignhk-0.7.22.post23/batchalign/version

@@ -0,0 +1,3 @@
+0.7.22-post.23
+Novmeber 5th, 2025
+OpenCC Reimelementations

batchalignhk-0.7.22.post20/batchalign/version

@@ -1,3 +0,0 @@
-0.7.22-post.20
-October 21st, 2025
-oai whisper for benchmakring