Update services.py

services.py

@@ -19,125 +19,129 @@ class GeminiService:
 
     def _check_client(self):
         if not self.client:
-            raise ValueError("API Key 未設定")
+            raise ValueError("API Key 未設定，請檢查 .env 或 Hugging Face Secrets")
 
-    def search_companies(self, query: str, exclude_names: List[str] = []) -> List[Dict]:
+    # ==========================
+    # 🎓 教授搜尋相關功能
+    # ==========================
+    def search_professors(self, query: str, exclude_names: List[str] = []) -> List[Dict]:
+        self._check_client()
+        exclusion_prompt = ""
+        if exclude_names:
+            exclusion_prompt = f"IMPORTANT: Do not include: {', '.join(exclude_names)}."
+
+        # Phase 1: Search
+        search_prompt = f"""
+        Using Google Search, find 10 prominent professors in universities across Taiwan who are experts in the field of "{query}".
+        CRITICAL: FACT CHECK they are current faculty. RELEVANCE must be high.
+        {exclusion_prompt}
+        List them (Name - University - Department) in Traditional Chinese.
         """
-        Step 1: 領域探索 -> 公司列表
+        search_response = self.client.models.generate_content(
+            model=self.model_id, contents=search_prompt,
+            config=types.GenerateContentConfig(tools=[types.Tool(google_search=types.GoogleSearch())])
+        )
+        
+        # Phase 2: Extract JSON
+        extract_prompt = f"""
+        From the text below, extract professor names, universities, and departments.
+        Calculate a Relevance Score (0-100) based on query: "{query}".
+        Return ONLY a JSON array: [{{"name": "...", "university": "...", "department": "...", "relevanceScore": 85}}]
+        Text: --- {search_response.text} ---
         """
+        extract_response = self.client.models.generate_content(
+            model=self.model_id, contents=extract_prompt,
+            config=types.GenerateContentConfig(response_mime_type='application/json')
+        )
+        try: return json.loads(extract_response.text)
+        except: return []
+
+    def get_professor_details(self, professor: Dict) -> Dict:
+        self._check_client()
+        name, uni, dept = professor.get('name'), professor.get('university'), professor.get('department')
+        prompt = f"""
+        Act as an academic consultant. Investigate Professor {name} from {dept} at {uni}.
+        Find "Combat Experience":
+        1. **Key Publications (Last 5 Years)**: Find 2-3 top papers with Citation Counts.
+        2. **Alumni Directions**: Where do their graduates work?
+        3. **Industry Collaboration**: Any industry projects?
+        Format output in Markdown (Traditional Chinese).
+        """
+        response = self.client.models.generate_content(
+            model=self.model_id, contents=prompt,
+            config=types.GenerateContentConfig(tools=[types.Tool(google_search=types.GoogleSearch())])
+        )
+        return self._format_response_with_sources(response)
+
+    # ==========================
+    # 🏢 公司搜尋相關功能
+    # ==========================
+    def search_companies(self, query: str, exclude_names: List[str] = []) -> List[Dict]:
         self._check_client()
         exclusion_prompt = ""
         if exclude_names:
             exclusion_prompt = f"IMPORTANT: Do not include: {', '.join(exclude_names)}."
 
-        # Phase 1: Google Search (廣泛探索)
-        # 這裡的 Prompt 強調：如果使用者輸入的是「領域(如: AI)」，請列出該領域的台灣代表性公司。
+        # Phase 1: Search
         search_prompt = f"""
-        Using Google Search, find 5 to 10 prominent companies in Taiwan related to the query: "{query}".
-        
-        **Instructions:**
-        1. **Domain Search:** If "{query}" is an industry or technology (e.g., "AI", "Green Energy"), list the top representative Taiwanese companies in this field.
-        2. **Company Search:** If "{query}" is a specific name, list that company and its direct competitors.
-        3. **Target:** Focus on Taiwanese companies (or global companies with major R&D in Taiwan).
+        Using Google Search, find 5 to 10 prominent companies in Taiwan related to: "{query}".
+        Instructions:
+        1. If "{query}" is an industry (e.g. AI), list representative Taiwanese companies.
+        2. If "{query}" is a name, list the company and competitors.
         {exclusion_prompt}
-        
         List them (Full Name - Industry/Main Product) in Traditional Chinese.
         """
-
         search_response = self.client.models.generate_content(
-            model=self.model_id,
-            contents=search_prompt,
-            config=types.GenerateContentConfig(
-                tools=[types.Tool(google_search=types.GoogleSearch())]
-            )
+            model=self.model_id, contents=search_prompt,
+            config=types.GenerateContentConfig(tools=[types.Tool(google_search=types.GoogleSearch())])
         )
-        raw_text = search_response.text
 
-        # Phase 2: Extract JSON (結構化)
+        # Phase 2: Extract JSON
         extract_prompt = f"""
-        From the text below, extract company names and their industry/main product.
-        Calculate a Relevance Score (0-100) based on query: "{query}".
-        
-        Return ONLY a JSON array: [{{"name": "...", "industry": "...", "relevanceScore": 85}}]
-        
-        Text:
-        ---
-        {raw_text}
-        ---
+        From text, extract company names and industry.
+        Calculate Relevance Score (0-100) for query: "{query}".
+        Return ONLY JSON array: [{{"name": "...", "industry": "...", "relevanceScore": 85}}]
+        Text: --- {search_response.text} ---
         """
-
         extract_response = self.client.models.generate_content(
-            model=self.model_id,
-            contents=extract_prompt,
-            config=types.GenerateContentConfig(
-                response_mime_type='application/json'
-            )
+            model=self.model_id, contents=extract_prompt,
+            config=types.GenerateContentConfig(response_mime_type='application/json')
         )
-
-        try:
-            return json.loads(extract_response.text)
-        except Exception as e:
-            print(f"JSON Parse Error: {e}")
-            return []
+        try: return json.loads(extract_response.text)
+        except: return []
 
     def get_company_details(self, company: Dict) -> Dict:
-        """
-        Step 2: 進行商業徵信調查 (Deep Dive)
-        """
         self._check_client()
         name = company.get('name')
-        
         prompt = f"""
-        Act as a professional "Business Analyst & Investigator". 
-        Conduct a comprehensive investigation on the Taiwanese company: "{name}".
-
-        **Investigation Targets:**
-
-        1.  **Overview (基本盤)**:
-            - **Tax ID (統編)** & **Capital (資本額)**. (Try to find specific numbers)
-            - **Representative (代表人)**.
-            - **Core Business**: What specific problem do they solve? What is their "Ace" product?
-
-        2.  **Workforce & Culture (內部情報)**:
-            - **Employee Count**.
-            - **Reviews/Gossip**: Search **PTT (Tech_Job, Soft_Job)**, **Dcard**, **Qollie**.
-            - Summarize the *REAL* work vibe (e.g., "Good for juniors but low ceiling", "Free snacks but forced overtime").
-
-        3.  **Legal & Risks (排雷專區)**:
-            - Search: "{name} 勞資糾紛", "{name} 違反勞基法", "{name} 判決", "{name} 罰款".
-            - List any red flags found in government records or news.
-
-        **Format**:
-        - Use Markdown.
-        - Language: Traditional Chinese (繁體中文).
-        - Be objective but don't sugarcoat potential risks.
+        Act as a "Business Analyst". Investigate Taiwanese company: "{name}".
+        Targets:
+        1. **Overview**: Tax ID (統編), Capital (資本額), Representative.
+        2. **Workforce & Culture**: Employee count, Reviews from PTT(Tech_Job)/Dcard/Qollie (Pros & Cons).
+        3. **Legal & Risks**: Search for "{name} 勞資糾紛", "{name} 判決", "{name} 違反勞基法".
+        Format in Markdown (Traditional Chinese). Be objective.
         """
-
         response = self.client.models.generate_content(
-            model=self.model_id,
-            contents=prompt,
-            config=types.GenerateContentConfig(
-                tools=[types.Tool(google_search=types.GoogleSearch())]
-            )
+            model=self.model_id, contents=prompt,
+            config=types.GenerateContentConfig(tools=[types.Tool(google_search=types.GoogleSearch())])
         )
-        
-        # Extract Sources
+        return self._format_response_with_sources(response)
+
+    # ==========================
+    # 共用功能
+    # ==========================
+    def _format_response_with_sources(self, response):
         sources = []
         if response.candidates[0].grounding_metadata and response.candidates[0].grounding_metadata.grounding_chunks:
             for chunk in response.candidates[0].grounding_metadata.grounding_chunks:
                 if chunk.web and chunk.web.uri and chunk.web.title:
                     sources.append({"title": chunk.web.title, "uri": chunk.web.uri})
-        
         unique_sources = {v['uri']: v for v in sources}.values()
+        return {"text": response.text, "sources": list(unique_sources)}
 
-        return {
-            "text": response.text,
-            "sources": list(unique_sources)
-        }
-
-    def chat_with_ai(self, history: List[Dict], new_message: str, context: str) -> str:
+    def chat_with_ai(self, history: List[Dict], new_message: str, context: str, role_instruction: str = "Source of truth") -> str:
         self._check_client()
-        system_instruction = f"You are an expert Business Consultant. Answer based on this company report:\n{context}"
+        system_instruction = f"{role_instruction}:\n{context}"
         
         chat_history = []
         for h in history:
@@ -145,12 +149,8 @@ class GeminiService:
             chat_history.append(types.Content(role=role, parts=[types.Part(text=h["content"])]))
 
         chat = self.client.chats.create(
-            model=self.model_id,
-            history=chat_history,
-            config=types.GenerateContentConfig(
-                system_instruction=system_instruction
-            )
+            model=self.model_id, history=chat_history,
+            config=types.GenerateContentConfig(system_instruction=system_instruction)
         )
-        
         response = chat.send_message(new_message)
         return response.text