更新长度检测

app.py

@@ -9,6 +9,7 @@ from pypinyin import lazy_pinyin
 from gradio_i18n import gettext, Translate
 
 from api import generate_api, get_audio
+from utils import get_length
 
 # 翻译文件位置
 trans_file = os.path.join(os.path.dirname(__file__),"i18n", "translations.json")
@@ -104,15 +105,15 @@ async def generate(selected_character = None, selected_characters = [], text = "
         elif lang == "ko":
             raise gr.Error("합성할 텍스트를 입력하세요")
     
-    if (lang == "en" and len(text.split()) > 200) or len(text) > 512:
+    if get_length(text) > 1024:
         if lang == "zh":
-            raise gr.Error("长度请控制在512个字符以内")
+            raise gr.Error("长度请控制在1024个字符以内")
         elif lang == "en":
-            raise gr.Error("The text length exceeds 200 words")
+            raise gr.Error("The text length exceeds 1024 words")
         elif lang == "ja":
-            raise gr.Error("テキストの長さが512文字を超えています")
+            raise gr.Error("テキストの長さが1024文字を超えています")
         elif lang == "ko":
-            raise gr.Error("텍스트 길이가 512자를 초과합니다")
+            raise gr.Error("텍스트 길이가 1024자를 초과합니다")
     
     audio = await generate_api(voice_ids, text)
     end_time = time.time()
@@ -218,8 +219,14 @@ def on_select(evt: gr.SelectData, characters, selected_characters, all_character
     
     selected = characters[evt.index]
     emotions = get_character_emotions(selected, all_characters)
-    default_emotion = emotions[0]["情绪"] if emotions else ""
-    default_voice_id = emotions[0]["voice_id"] if emotions else ""
+    normal_index = 0
+    for index, emotion in enumerate(emotions):
+        if emotion["情绪"] == "正常" or emotion["情绪"] == "보통" or emotion["情绪"] == "normal":
+            normal_index = index
+            break
+    
+    default_emotion = emotions[normal_index]["情绪"] if emotions else ""
+    default_voice_id = emotions[normal_index]["voice_id"] if emotions else ""
     
     character_dict = selected.copy()
     character_dict['情绪'] = default_emotion

utils.py

@@ -1,6 +1,7 @@
 import io
 import os
 import pickle
+import re
 
 import soundfile as sf
 import numpy as np
@@ -37,5 +38,26 @@ def normalize_audio_loudness(data: bytes, target_loudness: float = -23.0) -> byt
 
     return normalized_audio_bytes
 
+def get_length(text: str) -> float:
+    def calculate_string_length(text: str) -> float:
+        def split_into_words(s: str) -> list[str]:
+            return re.findall(r"\b\w+\b|[^\w\s]|\s+", s)
+
+        def calculate_effective_length(words: list[str]) -> float:
+            length = 0
+            for word in words:
+                if re.match(r"^[\u4e00-\u9fff\u3040-\u30ff\u3400-\u4dbf]+$", word):
+                    length += len(word)
+                elif re.match(r"^\w+$", word):
+                    length += 1
+                else:
+                    length += len(word) * 0.5
+            return length
+
+        words = split_into_words(text)
+        return calculate_effective_length(words)
+
+    return calculate_string_length(text)
+
 if __name__ == "__main__":
     normalize_audio_loudness()