Enhance impact analysis and configuration for vector persistence and API integration. Update main.spec to clarify data handling for ChromaDB, ensuring vector directories are excluded from packaging. Modify Text2SQLOrchestrator to unify question normalization for consistent SQL generation across languages. Introduce SchemaIndexer improvements for persistent vector storage and optimize embedding retrieval processes. Update documentation and comments for clarity on configuration changes and behavior adjustments.

backend/llm/deepseek_client.py

@@ -325,25 +325,34 @@ class DeepSeekClient:
         kwargs.setdefault("top_p", 1.0)
         return self.chat_with_json(messages, **kwargs)
 
-    def translate_nl_question_to_zh(self, question: str) -> str:
+    def normalize_nl_question_for_text2sql(self, question: str) -> str:
         """
-        将主要为英文的自然语言分析问题译为中文，便于与中文 Schema 注释 / 向量索引对齐。
+        将中文/英文问句归一为**一句**标准中文（temperature=0），使同一语义的中英表述
+        走同一套向量检索与 SQL 生成路径。
         """
-        from config.prompts import TRANSLATE_NL_TO_ZH_SYSTEM, TRANSLATE_NL_TO_ZH_USER
+        from config.prompts import (
+            CANONICALIZE_NL_FOR_TEXT2SQL_SYSTEM,
+            CANONICALIZE_NL_FOR_TEXT2SQL_USER,
+        )
 
         q = (question or "").strip()
         if not q:
             return ""
         messages = [
-            {"role": "system", "content": TRANSLATE_NL_TO_ZH_SYSTEM},
-            {"role": "user", "content": TRANSLATE_NL_TO_ZH_USER.format(question=q)},
+            {"role": "system", "content": CANONICALIZE_NL_FOR_TEXT2SQL_SYSTEM},
+            {"role": "user", "content": CANONICALIZE_NL_FOR_TEXT2SQL_USER.format(question=q)},
         ]
         msg = self.chat(messages, temperature=0.0, top_p=1.0, max_tokens=512)
         text = (msg.content or "").strip()
-        # 只取首行，避免模型附加说明
         line = text.splitlines()[0].strip() if text else ""
         return line.strip("「」\"'“”")
 
+    def translate_nl_question_to_zh(self, question: str) -> str:
+        """
+        兼容旧名：与 :meth:`normalize_nl_question_for_text2sql` 相同（不再仅限英文）。
+        """
+        return self.normalize_nl_question_for_text2sql(question)
+
 
 class AsyncDeepSeekClient:
     """