gemini 2.5 pro

app.py

@@ -8,12 +8,12 @@ import re
 import time
 from google.api_core import exceptions
 
-# --- Constants (No Changes) ---
+# --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 MAX_ITERATIONS = 7
 MAX_RETRIES = 5
 
-# --- Tool Definitions (No Changes) ---
+# --- Tool Definitions (Unchanged) ---
 class WebSearchTool:
     def __init__(self, api_key):
         self.api_key = api_key
@@ -23,7 +23,7 @@ class WebSearchTool:
         payload = {"model": "llama-3-sonar-small-32k-online", "messages": [{"role": "system", "content": "You are a world-class research assistant. Answer the user's query based on verifiable public information. Be precise and comprehensive."}, {"role": "user", "content": query}]}
         headers = {"accept": "application/json", "content-type": "application/json", "Authorization": f"Bearer {self.api_key}"}
         try:
-            response = requests.post(self.url, json=payload, headers=headers, timeout=30)
+            response = requests.post(self.url, json=payload, headers=headers, timeout=40)
             response.raise_for_status()
             return response.json()['choices'][0]['message']['content']
         except requests.exceptions.RequestException as e:
@@ -51,87 +51,77 @@ class GAIAAgent:
     def __init__(self, gemini_api_key: str, pplx_api_key: str, api_url: str):
         print("Initializing GAIAAgent...")
         genai.configure(api_key=gemini_api_key)
-        self.model = genai.GenerativeModel('gemini-1.5-flash-latest')
+        
+        # --- CORRECTING THE MODEL TO THE USER'S SPECIFICATION ---
+        self.model_name = 'gemini-2.5-pro-preview-06-05'
+        self.model = genai.GenerativeModel(self.model_name)
+        print(f"Agent equipped with user-specified model: {self.model_name}")
+        
         self.tools = {"WebSearch": WebSearchTool(api_key=pplx_api_key), "FileDownloader": FileDownloaderTool(api_url=api_url)}
 
-        # MODIFIED: Made the zero-shot prompt even stricter to prevent conversational filler.
         self.zero_shot_prompt_template = """
-You are a highly intelligent question-answering bot. Follow these instructions precisely.
-1. Analyze the user's question.
-2. If the question is simple and you are 100% certain of the answer without needing any tools, provide ONLY the answer and nothing else.
-3. If the question requires web searches, file access, or complex reasoning, respond with the single word: UNSURE.
-Do not add any explanations or introductory phrases.
+Analyze the following question. If the answer is self-contained in the question and requires no external tools, provide the answer directly and concisely. Otherwise, respond with the single word: UNSURE.
 
 Question: {question}
 Answer:"""
 
-        # MODIFIED: Added a new, explicit rule for how to fail gracefully.
         self.react_prompt_template = """
-You are a helpful assistant designed to answer questions accurately.
-
-To solve the user's question, you must use a sequence of thoughts and actions.
-You have access to the following tools:
+You are a state-of-the-art reasoning agent. Your goal is to answer the user's question by creating a plan and executing it using the tools provided.
 
-- **WebSearch[query]**: Use this to search the internet for up-to-date information.
-- **FileDownloader[task_id]**: Use this to download a file associated with the current task. The task_id is '{task_id}'.
+**Tools Available:**
+- **WebSearch[query]**: Searches the web for information. Use different queries if initial results are not satisfactory.
+- **FileDownloader[task_id]**: Downloads a file for a specific task. The task_id is '{task_id}'.
 
-Your reasoning process should follow this format:
-
-Thought: I need to figure out what information is missing. I will use a tool to find it.
-Action: ToolName[input for the tool]
+**Reasoning Format:**
+Thought: My reasoning process and plan to solve the question.
+Action: The tool I will use, in the format `ToolName[input]`.
 Observation: [The result from the tool will be inserted here]
-... (this Thought/Action/Observation cycle can repeat)
-
-Thought: I have now gathered enough information to answer the user's question.
-Final Answer: The final answer to the original question.
+... (The Thought/Action/Observation cycle can repeat)
+Thought: I have sufficient information to provide the final answer.
+Final Answer: The definitive answer to the user's question.
 
-**Important Rules:**
-1. The `Action` line must be *exactly* in the format `ToolName[input]`.
-2. The `task_id` for the current question is '{task_id}'. Use it ONLY with the FileDownloader tool.
-3. **CRITICAL RULE:** If you determine that the question cannot be answered with your tools (e.g., a required file is missing, the information is not on the web), you MUST conclude with: `Final Answer: I am unable to answer this question.` Do not make up an answer.
+**Guiding Principles:**
+1.  **Persist:** Do not give up easily. If a search fails, re-evaluate and try a more specific or different query.
+2.  **Conclude:** Once you have the answer, state it clearly with `Final Answer:`.
+3.  **Fail Gracefully:** If, after several genuine attempts, you conclude the answer is unobtainable, state `Final Answer: I am unable to answer this question.`
 
-Here is the question:
-{question}"""
-        print("GAIAAgent initialized successfully.")
+Question: {question}"""
+        print("GAIAAgent initialized with refined prompts for the specified model.")
 
     def _call_gemini_api_with_backoff(self, prompt_text):
         retries = 0
         while retries < MAX_RETRIES:
             try:
-                time.sleep(1) # Add a small base delay
+                time.sleep(2) 
                 response = self.model.generate_content(prompt_text)
                 return response.text
             except exceptions.ResourceExhausted as e:
-                wait_time = (2 ** retries)
+                wait_time = (2 ** retries) + 2
                 print(f"API Rate Limit Exceeded (429). Waiting for {wait_time}s to retry...")
                 time.sleep(wait_time)
                 retries += 1
             except Exception as e:
+                print(f"An unexpected error occurred with Gemini API: {e}")
                 return f"AGENT_ERROR: An unexpected error occurred: {e}"
         return "AGENT_ERROR: API rate limit exceeded after multiple retries."
 
     def __call__(self, question: str, task_id: str) -> str:
         print(f"\n{'='*20}\nProcessing Task ID: {task_id}\nQuestion: {question[:100]}...")
+        print(f"--- Using model: {self.model_name} ---")
 
-        print("--- Step 1: Zero-Shot Attempt ---")
+        # Step 1: Zero-Shot Attempt
         zero_shot_prompt = self.zero_shot_prompt_template.format(question=question)
         zero_shot_answer = self._call_gemini_api_with_backoff(zero_shot_prompt).strip()
-
-        if "AGENT_ERROR" in zero_shot_answer: return zero_shot_answer
-        
-        if "UNSURE" not in zero_shot_answer.upper():
+        if "AGENT_ERROR" not in zero_shot_answer and "UNSURE" not in zero_shot_answer.upper():
             print(f"Zero-shot successful! Answer: {zero_shot_answer}")
             return zero_shot_answer
 
-        print("--- Step 2: Zero-shot failed, starting ReAct loop ---")
-        
-        # CRITICAL FIX: Reset the prompt history for each question to prevent context bleed.
-        # This was the cause of the botany/bird video mix-up.
+        # Step 2: ReAct Loop
+        print("--- Zero-shot failed, starting ReAct loop ---")
         current_prompt_history = self.react_prompt_template.format(question=question, task_id=task_id)
         
         for i in range(MAX_ITERATIONS):
             print(f"\n--- ReAct Iteration {i+1} ---")
-            
             response_text = self._call_gemini_api_with_backoff(current_prompt_history)
             print(f"LLM Response:\n{response_text}")
 
@@ -147,32 +137,25 @@ Here is the question:
             if action_match:
                 tool_name = action_match.group(1).strip()
                 tool_input = action_match.group(2).strip()
-
                 if tool_name in self.tools:
                     tool = self.tools[tool_name]
                     try:
                         observation = tool.execute(task_id if tool_name == "FileDownloader" else tool_input)
-                    except Exception as e:
-                        observation = f"Error executing tool: {e}"
-                    # Append the whole thought/action/observation cycle
+                    except Exception as e: observation = f"Error executing tool: {e}"
                     current_prompt_history += f"\n{response_text}\nObservation: {observation}"
                 else:
                     current_prompt_history += f"\n{response_text}\nObservation: Error - The tool '{tool_name}' does not exist."
             else:
-                print("Error: Agent did not provide a valid Action or Final Answer. Returning last response.")
                 return response_text.strip()
-
         return "AGENT_ERROR: Agent reached max iterations."
 
-# --- Main run_and_submit_all function (No significant changes needed here, only added a longer sleep) ---
+# --- Main run_and_submit_all function ---
 def run_and_submit_all(profile: gr.OAuthProfile | None):
     space_id = os.getenv("SPACE_ID")
     if not profile: return "Please Login to Hugging Face with the button.", None
     username = f"{profile.username}"
-    print(f"User logged in: {username}")
 
-    pplx_key = os.getenv("PPLX_API_KEY")
-    gemini_key = os.getenv("GEMINI_API_KEY")
+    pplx_key, gemini_key = os.getenv("PPLX_API_KEY"), os.getenv("GEMINI_API_KEY")
     if not pplx_key or not gemini_key: return "API keys not found in Space secrets.", None
     
     api_url = DEFAULT_API_URL
@@ -192,8 +175,8 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         except Exception as e:
              results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})
         
-        print(f"--- Waiting for 8 seconds before next question to respect rate limits ---")
-        time.sleep(8) # Increased delay to be safer
+        print(f"--- Waiting for 12 seconds before next question... ---")
+        time.sleep(12)
 
     if not answers_payload: return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
     
@@ -212,18 +195,10 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     except requests.exceptions.RequestException as e:
         return f"Submission Failed: {e}", pd.DataFrame(results_log)
 
-# --- Gradio Interface (No changes) ---
+# --- Gradio Interface (Unchanged) ---
 with gr.Blocks() as demo:
     gr.Markdown("# GAIA Agent Evaluation Runner")
-    gr.Markdown("""
-        **Instructions:**
-        1.  Ensure you have added your `PPLX_API_KEY` and `GEMINI_API_KEY` to this Space's **Settings > Secrets**.
-        2.  Log in to your Hugging Face account using the button below. This uses your HF username for submission.
-        3.  Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
-        ---
-        **Disclaimers:**
-        This process is slow due to the added delays to respect API rate limits. Please be patient.
-        """)
+    gr.Markdown("Equipped with the user-specified **gemini-2.5-pro-preview-06-05** model.")
     gr.LoginButton()
     run_button = gr.Button("Run Evaluation & Submit All Answers")
     status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)