优化翻译的逻辑

translation_utils.py

@@ -1,57 +1,108 @@
 import translators as ts
 import re
 
+# 输入长度限制
+INPUT_LIMIT = 1000
+
 def split_text_for_translation(text):
     """
     按段落和完整句子拆分文本，返回句子列表（保持原有段落结构）。
     """
     paragraphs = text.split('\n')
     split_paragraphs = []
-    # 句子分隔符：句号、问号、感叹号（中英文）
-    sentence_pattern = re.compile(r'([^。！？!?\r\n]+[。！？!?])', re.M)
+    # 句子分隔符：句号、问号、感叹号（英文和中文）
+    sentence_pattern = re.compile(r'([^.!?。！？]+[.!?。！？])', re.M)
+    
     for para in paragraphs:
         if not para.strip():
             split_paragraphs.append([''])
             continue
+        
+        # 如果段落长度小于INPUT_LIMIT，直接添加
+        if len(para) < INPUT_LIMIT:
+            split_paragraphs.append([para])
+            continue
+        
+        # 分割句子
         sentences = sentence_pattern.findall(para)
-        # 若正则未能分出句子，则整体作为一句
+        
+        # 若正则未能分出句子，则按照INPUT_LIMIT拆分
         if not sentences:
-            sentences = [para]
-        split_paragraphs.append(sentences)
+            sentences = [para[i:i+INPUT_LIMIT] for i in range(0, len(para), INPUT_LIMIT)]
+        
+        # 组合句子的逻辑
+        optimized_sentences = []
+        current_group = []
+        current_length = 0
+        
+        for sent in sentences:
+            # 计算当前句子长度
+            sent_length = len(sent)
+            
+            # 如果加入这个句子后总长度小于INPUT_LIMIT，则加入
+            if current_length + sent_length < INPUT_LIMIT:
+                current_group.append(sent)
+                current_length += sent_length
+            else:
+                # 如果当前组不为空，先保存当前组
+                if current_group:
+                    optimized_sentences.append(''.join(current_group))
+                
+                # 重置组
+                current_group = [sent]
+                current_length = sent_length
+        
+        # 添加最后一组
+        if current_group:
+            optimized_sentences.append(''.join(current_group))
+        
+        split_paragraphs.append(optimized_sentences)
+    
     return split_paragraphs
 
-def translate_text(text, to_lang='zh'):
+def translate_text(text, from_lang='en', to_lang='zh'):
     """
     自动处理长文本，按段落和句子拆分，逐句翻译，组合返回。
+    对于短文本，直接翻译。
     """
     if not text:
         return ""
+    
     try:
+        # 如果整个文本长度小于INPUT_LIMIT，直接翻译
+        if len(text) < INPUT_LIMIT:
+            return ts.translate_text(text, translator='bing', from_language=from_lang, to_language=to_lang)
+        
         # 拆分文本为段落和句子
         split_paragraphs = split_text_for_translation(text)
+
         translated_paragraphs = []
+        
         for sentences in split_paragraphs:
             if sentences == ['']:
                 translated_paragraphs.append('')
                 continue
+            
             translated_sentences = []
             for sent in sentences:
                 sent = sent.strip()
                 if not sent:
                     continue
                 # 逐句翻译
-                translated = ts.translate_text(sent, translator='bing', from_language='auto', to_language=to_lang)
+                translated = ts.translate_text(sent, translator='bing', from_language=from_lang, to_language=to_lang)
                 translated_sentences.append(translated)
             translated_paragraphs.append(''.join(translated_sentences))
+        
         return '\n'.join(translated_paragraphs)
     except Exception as e:
         return f"[翻译失败] {str(e)}"
 
+
 def translate_texts(strings, target_language):
     translated_strings = []
     for text in strings:
         try:
-            translated_string = translate_text(text, target_language)
+            translated_string = translate_text(text, 'auto', target_language)
             translated_strings.append(translated_string)
         except Exception as e:
             translated_strings.append(f"Translation error: {str(e)}")