Version Check (#1390)

* version check

* fix webui and symbols

* fix v1 language map

GPT_SoVITS/text/__init__.py

@@ -1,18 +1,26 @@
 import os
-if os.environ.get("version","v1")=="v1":
-  from text.symbols import symbols
-else:
-  from text.symbols2 import symbols
+# if os.environ.get("version","v1")=="v1":
+#   from text.symbols import symbols
+# else:
+#   from text.symbols2 import symbols
 
-_symbol_to_id = {s: i for i, s in enumerate(symbols)}
+from text import symbols as symbols_v1
+from text import symbols2 as symbols_v2
 
-def cleaned_text_to_sequence(cleaned_text):
+_symbol_to_id_v1 = {s: i for i, s in enumerate(symbols_v1.symbols)}
+_symbol_to_id_v2 = {s: i for i, s in enumerate(symbols_v2.symbols)}
+
+def cleaned_text_to_sequence(cleaned_text, version):
   '''Converts a string of text to a sequence of IDs corresponding to the symbols in the text.
     Args:
       text: string to convert to a sequence
     Returns:
       List of integers corresponding to the symbols in the text
   '''
-  phones = [_symbol_to_id[symbol] for symbol in cleaned_text]
+  if version == "v1":
+    phones = [_symbol_to_id_v1[symbol] for symbol in cleaned_text]
+  else:
+    phones = [_symbol_to_id_v2[symbol] for symbol in cleaned_text]
+
   return phones
 

GPT_SoVITS/text/cleaner.py

@@ -1,13 +1,17 @@
 from text import japanese, cleaned_text_to_sequence, english,korean,cantonese
 import os
-if os.environ.get("version","v1")=="v1":
-    from text import chinese
-    from text.symbols import symbols
-else:
-    from text import chinese2 as chinese
-    from text.symbols2 import symbols
+# if os.environ.get("version","v1")=="v1":
+#     from text import chinese
+#     from text.symbols import symbols
+# else:
+#     from text import chinese2 as chinese
+#     from text.symbols2 import symbols
+
+from text import symbols as symbols_v1
+from text import symbols2 as symbols_v2
+from text import chinese as chinese_v1
+from text import chinese2 as chinese_v2
 
-language_module_map = {"zh": chinese, "ja": japanese, "en": english, "ko": korean,"yue":cantonese}
 special = [
     # ("%", "zh", "SP"),
     ("￥", "zh", "SP2"),
@@ -16,13 +20,20 @@ special = [
 ]
 
 
-def clean_text(text, language):
+def clean_text(text, language, version):
+    if version == "v1":
+        symbols = symbols_v1.symbols
+        language_module_map = {"zh": chinese_v1, "ja": japanese, "en": english}
+    else:
+        symbols = symbols_v2.symbols
+        language_module_map = {"zh": chinese_v2, "ja": japanese, "en": english, "ko": korean,"yue":cantonese}
+
     if(language not in language_module_map):
         language="en"
         text=" "
     for special_s, special_l, target_symbol in special:
         if special_s in text and language == special_l:
-            return clean_special(text, language, special_s, target_symbol)
+            return clean_special(text, language, special_s, target_symbol, version)
     language_module = language_module_map[language]
     if hasattr(language_module,"text_normalize"):
         norm_text = language_module.text_normalize(text)
@@ -42,11 +53,18 @@ def clean_text(text, language):
         word2ph = None
 
     for ph in phones:
-        assert ph in symbols
+        phones = ['UNK' if ph not in symbols else ph for ph in phones]
     return phones, word2ph, norm_text
 
 
-def clean_special(text, language, special_s, target_symbol):
+def clean_special(text, language, special_s, target_symbol, version):
+    if version == "v1":
+        symbols = symbols_v1.symbols
+        language_module_map = {"zh": chinese_v1, "ja": japanese, "en": english}
+    else:
+        symbols = symbols_v2.symbols
+        language_module_map = {"zh": chinese_v2, "ja": japanese, "en": english, "ko": korean,"yue":cantonese}
+
     """
     特殊静音段sp符号处理
     """

GPT_SoVITS/text/english.py

@@ -6,10 +6,7 @@ from g2p_en import G2p
 
 from text.symbols import punctuation
 
-if os.environ.get("version","v1")=="v1":
-    from text.symbols import symbols
-else:
-    from text.symbols2 import symbols
+from text.symbols2 import symbols
 
 import unicodedata
 from builtins import str as unicode

GPT_SoVITS/text/japanese.py

@@ -4,12 +4,6 @@ import sys
 
 import pyopenjtalk
 
-
-import os
-if os.environ.get("version","v1")=="v1":
-    from text.symbols import symbols
-else:
-    from text.symbols2 import symbols
 from text.symbols import punctuation
 # Regular expression matching Japanese without punctuation marks:
 _japanese_characters = re.compile(
@@ -61,12 +55,13 @@ def post_replace_ph(ph):
         "、": ",",
         "...": "…",
     }
+
     if ph in rep_map.keys():
         ph = rep_map[ph]
-    if ph in symbols:
-        return ph
-    if ph not in symbols:
-        ph = "UNK"
+    # if ph in symbols:
+    #     return ph
+    # if ph not in symbols:
+    #     ph = "UNK"
     return ph
 
 

GPT_SoVITS/text/korean.py

@@ -2,11 +2,8 @@ import re
 from jamo import h2j, j2hcj
 import ko_pron
 from g2pk2 import G2p
-import os
-if os.environ.get("version","v1")=="v1":
-    from text.symbols import symbols
-else:
-    from text.symbols2 import symbols
+
+from text.symbols2 import symbols 
 
 # This is a list of Korean classifiers preceded by pure Korean numerals.
 _korean_classifiers = '군데 권 개 그루 닢 대 두 마리 모 모금 뭇 발 발짝 방 번 벌 보루 살 수 술 시 쌈 움큼 정 짝 채 척 첩 축 켤레 톨 통'