update

backend/app/services/simulation_manager.py

@@ -334,6 +334,16 @@ class SimulationManager:
             elif state.enable_twitter:
                 realtime_output_path = os.path.join(sim_dir, "twitter_profiles.csv")
                 realtime_platform = "twitter"
+
+            # Nền tảng dùng cho metadata cost.
+            if state.enable_twitter and state.enable_reddit:
+                runtime_platform = "parallel"
+            elif state.enable_twitter:
+                runtime_platform = "twitter"
+            elif state.enable_reddit:
+                runtime_platform = "reddit"
+            else:
+                runtime_platform = None
             
             profiles = generator.generate_profiles_from_entities(
                 entities=filtered.entities,
@@ -343,6 +353,7 @@ class SimulationManager:
                 parallel_count=parallel_profile_count,  # Số dòng luồng Async
                 realtime_output_path=realtime_output_path,  # Lưu log thời gian thực
                 output_platform=realtime_platform,  # Đuôi file xuất
+                metadata_platform=runtime_platform,
                 simulation_id=simulation_id,
                 project_id=state.project_id,
             )

backend/app/services/text_processor.py

@@ -54,7 +54,7 @@ class TextProcessor:
         # Xoá các dòng trống liên tiếp (Chỉ giữ lại tối đa 2 lần xuống dòng liên tiếp)
         text = re.sub(r'\n{3,}', '\n\n', text)
         
-        # 移除行首行尾空白
+        # Loại bỏ khoảng trắng đầu và cuối
         lines = [line.strip() for line in text.split('\n')]
         text = '\n'.join(lines)
         

backend/scripts/llm_cost_patch.py

@@ -20,6 +20,36 @@ def _build_metadata(model: str) -> Dict[str, Any]:
     return metadata
 
 
+def _normalize_messages_for_qwen(messages: Any) -> Any:
+    """Ensure Qwen-compatible ordering: a single system message at index 0."""
+    if not isinstance(messages, list):
+        return messages
+
+    if not messages:
+        return messages
+
+    system_contents = []
+    non_system_messages = []
+
+    for item in messages:
+        if isinstance(item, dict) and str(item.get("role", "")).lower() == "system":
+            content = item.get("content")
+            if content is not None and str(content).strip():
+                system_contents.append(str(content))
+            continue
+        non_system_messages.append(item)
+
+    # No system messages detected, keep original payload.
+    if not system_contents:
+        return messages
+
+    merged_system = {
+        "role": "system",
+        "content": "\n\n".join(system_contents),
+    }
+    return [merged_system] + non_system_messages
+
+
 def install_openai_cost_patch(
     *,
     simulation_id: Optional[str],
@@ -58,6 +88,9 @@ def install_openai_cost_patch(
 
     def sync_create_wrapper(self, *args, **kwargs):
         model = kwargs.get("model", "unknown_model")
+        if "qwen" in str(model).lower() and "messages" in kwargs:
+            kwargs = dict(kwargs)
+            kwargs["messages"] = _normalize_messages_for_qwen(kwargs.get("messages"))
         response = original_sync_create(self, *args, **kwargs)
         try:
             record_llm_cost(
@@ -71,6 +104,9 @@ def install_openai_cost_patch(
 
     async def async_create_wrapper(self, *args, **kwargs):
         model = kwargs.get("model", "unknown_model")
+        if "qwen" in str(model).lower() and "messages" in kwargs:
+            kwargs = dict(kwargs)
+            kwargs["messages"] = _normalize_messages_for_qwen(kwargs.get("messages"))
         response = await original_async_create(self, *args, **kwargs)
         try:
             record_llm_cost(

backend/scripts/run_parallel_simulation.py

@@ -361,7 +361,7 @@ class ParallelIPCHandler:
         """
         # Nếu có chỉ định nền tảng, chỉ phỏng vấn trên nền tảng đó
         if platform in ("twitter", "reddit"):
-            result = await self._interview_single_platform(agent_id, prompt, platform)
+            result = await self._interview_single_platform(agent_id, prompt, platform)   
             
             if "error" in result:
                 self.send_response(command_id, "failed", error=result["error"])
@@ -996,11 +996,27 @@ def create_model(config: Dict[str, Any], use_boost: bool = False):
         config: Dictionary cấu hình mô phỏng
         use_boost: Có dùng cấu hình LLM tăng tốc hay không (nếu khả dụng)
     """
-    # Kiểm tra có cấu hình tăng tốc không
+    def _is_placeholder(value: str) -> bool:
+        v = (value or "").strip().lower()
+        return v in {"your_api_key_here", "your_base_url_here", "your_model_name_here"}
+
+    def _is_http_url(value: str) -> bool:
+        v = (value or "").strip().lower()
+        return v.startswith("http://") or v.startswith("https://")
+
+    # Kiểm tra có cấu hình tăng tốc hợp lệ không
     boost_api_key = os.environ.get("LLM_BOOST_API_KEY", "")
     boost_base_url = os.environ.get("LLM_BOOST_BASE_URL", "")
     boost_model = os.environ.get("LLM_BOOST_MODEL_NAME", "")
-    has_boost_config = bool(boost_api_key)
+    has_boost_config = (
+        bool(boost_api_key.strip())
+        and bool(boost_model.strip())
+        and bool(boost_base_url.strip())
+        and not _is_placeholder(boost_api_key)
+        and not _is_placeholder(boost_base_url)
+        and not _is_placeholder(boost_model)
+        and _is_http_url(boost_base_url)
+    )
     
     # Chọn LLM theo tham số và trạng thái cấu hình
     if use_boost and has_boost_config:
@@ -1011,6 +1027,8 @@ def create_model(config: Dict[str, Any], use_boost: bool = False):
         config_label = "[Boost LLM]"
     else:
         # Dùng cấu hình chung
+        if use_boost and not has_boost_config:
+            print("[Boost LLM] Invalid or placeholder boost config, fallback to [General LLM].")
         llm_api_key = os.environ.get("LLM_API_KEY", "")
         llm_base_url = os.environ.get("LLM_BASE_URL", "")
         llm_model = os.environ.get("LLM_MODEL_NAME", "")
@@ -1028,6 +1046,10 @@ def create_model(config: Dict[str, Any], use_boost: bool = False):
         raise ValueError("Missing API key config. Please set LLM_API_KEY in the project root .env file")
     
     if llm_base_url:
+        if not _is_http_url(llm_base_url):
+            raise ValueError(
+                f"Invalid LLM base URL: {llm_base_url}. It must start with http:// or https://"
+            )
         os.environ["OPENAI_API_BASE_URL"] = llm_base_url
     
     print(f"{config_label} model={llm_model}, base_url={llm_base_url[:40] if llm_base_url else 'default'}...")

frontend/src/components/Step4Report.vue

@@ -862,7 +862,7 @@ const parseInterview = (text) => {
         const quotesText = quotesMatch[1]
         // Prioritize matching > "text" format
         let quoteMatches = quotesText.match(/> "([^"]+)"/g)
-        // Backtrack: Matches > "text" or > \u201Ctext\u201D (Chinese quotation marks)
+        // Backtrack: Matches > "text" or > \u201Ctext\u201D (Vietnamese quotation marks)
         if (!quoteMatches) {
           quoteMatches = quotesText.match(/> [\u201C""]([^\u201D""]+)[\u201D""]/g)
         }
@@ -2089,9 +2089,9 @@ const extractFinalContent = (response) => {
   }
   
   // Try to find the content after "Final Answer:"
-  const chineseFinalMatch = response.match(/Final Answer[:：]\s*\n*([\s\S]*)$/i)
-  if (chineseFinalMatch) {
-    return chineseFinalMatch[1].trim()
+  const vietnameseFinalMatch = response.match(/Final Answer[:：]\s*\n*([\s\S]*)$/i)
+  if (vietnameseFinalMatch) {
+    return vietnameseFinalMatch[1].trim()
   }
   
   // Nếu bắt đầu bằng ## hoặc # hoặc >, có thể là nội dung markdown trực tiếp

frontend/src/views/Home.vue

@@ -176,7 +176,7 @@
                 <textarea
                   v-model="formData.simulationRequirement"
                   class="code-input"
-                  placeholder="// Describe your simulation or prediction request in natural language (e.g. If Wuhan University announces the revocation of disciplinary action against someone, what public opinion trends would emerge?)"
+                  placeholder="Describe your simulation or prediction request in natural language (e.g. If Wuhan University announces the revocation of disciplinary action against someone, what public opinion trends would emerge?)"
                   rows="6"
                   :disabled="loading"
                 ></textarea>

package.json

@@ -1,7 +1,7 @@
 {
   "name": "mirofish",
   "version": "0.1.0",
-  "description": "MiroFish - 简洁通用的群体智能引擎，预测万物",
+  "description": "MiroFish - A simple and versatile swarm intelligence engine that predicts everything",
   "scripts": {
     "setup": "npm install && cd frontend && npm install",
     "setup:backend": "cd backend && uv sync",