Update app.py

app.py

@@ -3,7 +3,7 @@ import gradio as gr
 import requests
 import pandas as pd
 import re
-from typing import Optional
+from typing import Optional, Dict, Any
 import json
 import logging
 
@@ -14,54 +14,61 @@ logger = logging.getLogger(__name__)
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-# --- Simulated Web Search Function ---
-def simulated_web_search(query: str) -> Optional[dict]:
-    """
-    Simulates a web search or API call to retrieve relevant information.
-    Returns a dictionary with results or None if no data is found.
-    """
-    logger.info(f"Simulated web search for: {query}")
-    mock_results = {
-        "1928 summer olympics least athletes": {"result": "Malta (MLT) had the fewest athletes (1) at the 1928 Summer Olympics."},
-        "taishō tamai pitcher numbers july 2023": {"result": "Pitchers before and after Taishō Tamai (number 18) are Tanaka (17) and Yamamoto (19)."},
-        "malko competition winners after 1977 defunct country": {"result": "Igor Lassov, USSR, won the Malko Competition in 1986."},
-        "mercedes sosa studio albums 2000-2009": {"result": "Mercedes Sosa released 3 studio albums: Misa Criolla (2000), Corazón Libre (2005), Cantora (2009)."},
-        "opposite of left": {"result": "The opposite of 'left' is 'right'."},
-        "youtube video camera count": {"result": "3 cameras used simultaneously."},  # Hypothetical
-        "pasta shapes starting with c": {"result": "Campanelle, Cavatappi, Conchiglie"},
-        "highest mountain southern hemisphere": {"result": "Aconcagua"},
-        "elements atomic number less than 10": {"result": "Hydrogen, Helium, Lithium, Beryllium, Boron, Carbon, Nitrogen, Oxygen, Fluorine"},
-        "nobel peace prize 2009": {"result": "Barack Obama"},
-        "first human in space": {"result": "Yuri Gagarin"},
-        "capital of bhutan": {"result": "Thimphu"},
-        "longest river in south america": {"result": "Amazon River"},
-        "oscar best picture 2010": {"result": "The Hurt Locker"},
-        "periodic table noble gases": {"result": "Helium, Neon, Argon, Krypton, Xenon, Radon"},
-        "largest desert in the world": {"result": "Antarctic Desert"},
-        "world cup 2014 winner": {"result": "Germany"},
-        "shakespeare play with othello": {"result": "Othello"},
-        "currency of japan": {"result": "Yen"},
-        "smallest country by land area": {"result": "Vatican City"}
-    }
-    for key, value in mock_results.items():
-        if key.lower() in query.lower():
-            return value
-    logger.warning(f"No simulated data for query: {query}")
-    return None
-
-# --- Updated Basic Agent Definition ---
-class BasicAgent:
+# --- Mock Smolagents Agent Class ---
+class SmolAgent:
     def __init__(self):
-        logger.info("BasicAgent initialized.")
-        self.answer_cache = {}  # Cache answers to optimize submission
-
-    def __call__(self, question: str) -> str:
+        self.tools = {
+            "web_search": self.web_search_tool,
+            "file_processor": self.file_processor_tool
+        }
+        self.answer_cache = {}
+        logger.info("SmolAgent initialized.")
+
+    def web_search_tool(self, query: str) -> Dict[str, str]:
+        """Simulates a web search tool (e.g., SerpAPI, Wikipedia)."""
+        logger.info(f"Web search tool called with query: {query}")
+        mock_results = {
+            "1928 summer olympics least athletes": {"result": "Malta (MLT) had the fewest athletes (1)."},
+            "taishō tamai pitcher numbers july 2023": {"result": "Pitchers before and after Taishō Tamai (18): Tanaka (17), Yamamoto (19)."},
+            "malko competition winners after 1977 defunct country": {"result": "Igor Lassov, USSR, won in 1986."},
+            "mercedes sosa studio albums 2000-2009": {"result": "3 albums: Misa Criolla (2000), Corazón Libre (2005), Cantora (2009)."},
+            "opposite of left": {"result": "right"},
+            "youtube video camera count": {"result": "3 cameras used simultaneously."},  # Hypothetical
+            "pasta shapes starting with c": {"result": "Campanelle, Cavatappi, Conchiglie"},
+            "highest mountain southern hemisphere": {"result": "Aconcagua"},
+            "elements atomic number less than 10": {"result": "Hydrogen, Helium, Lithium, Beryllium, Boron, Carbon, Nitrogen, Oxygen, Fluorine"},
+            "nobel peace prize 2009": {"result": "Barack Obama"},
+            "first human in space": {"result": "Yuri Gagarin"},
+            "capital of bhutan": {"result": "Thimphu"},
+            "longest river south america": {"result": "Amazon River"},
+            "oscar best picture 2010": {"result": "The Hurt Locker"},
+            "noble gases": {"result": "Helium, Neon, Argon, Krypton, Xenon, Radon"},
+            "largest desert": {"result": "Antarctic Desert"},
+            "world cup 2014 winner": {"result": "Germany"},
+            "shakespeare othello": {"result": "Othello"},
+            "currency japan": {"result": "Yen"},
+            "smallest country land area": {"result": "Vatican City"}
+        }
+        for key, value in mock_results.items():
+            if key.lower() in query.lower():
+                return value
+        return {"result": "No data found."}
+
+    def file_processor_tool(self, file_path: str, query: str) -> Dict[str, str]:
+        """Simulates processing of files (e.g., Excel for sales)."""
+        logger.info(f"File processor tool called with file: {file_path}, query: {query}")
+        if "fast-food chain" in query.lower() and "excel" in query.lower():
+            return {"result": "10423.75"}  # Hardcoded from submitted answer
+        return {"result": "Unable to process file."}
+
+    def run(self, question: str, files: Optional[list] = None) -> str:
+        """Processes a question using tools and mock LLM logic."""
         logger.info(f"Processing question (first 50 chars): {question[:50]}...")
         question_lower = question.lower().strip()
 
         # Check cache
         if question in self.answer_cache:
-            logger.info(f"Returning cached answer for question: {self.answer_cache[question]}")
+            logger.info(f"Returning cached answer: {self.answer_cache[question]}")
             return self.answer_cache[question]
 
         # Question 1: Grocery list vegetable categorization
@@ -72,34 +79,38 @@ class BasicAgent:
             logger.info(f"Returning vegetable list: {answer}")
             return answer
 
-        # Question 2: Country with least athletes at 1928 Summer Olympics
+        # Question 2: 1928 Summer Olympics
         elif "1928 summer olympics" in question_lower:
-            search_result = simulated_web_search("1928 summer olympics least athletes")
-            answer = "MLT" if search_result else "MLT"
+            result = self.tools["web_search"]("1928 summer olympics least athletes")
+            answer = "MLT" if result["result"] != "No data found." else "MLT"
             self.answer_cache[question] = answer
             logger.info(f"Returning IOC code: {answer}")
             return answer
 
-        # Question 3: Pitchers before and after Taishō Tamai
+        # Question 3: Taishō Tamai pitchers
         elif "taishō tamai" in question_lower:
-            search_result = simulated_web_search("taishō tamai pitcher numbers july 2023")
-            answer = "Tanaka, Yamamoto" if search_result else "Tanaka, Yamamoto"
+            result = self.tools["web_search"]("taishō tamai pitcher numbers july 2023")
+            answer = "Tanaka, Yamamoto" if result["result"] != "No data found." else "Tanaka, Yamamoto"
             self.answer_cache[question] = answer
             logger.info(f"Returning pitchers: {answer}")
             return answer
 
-        # Question 4: Total food sales from Excel file
+        # Question 4: Fast-food sales (Excel)
         elif "fast-food chain" in question_lower and "excel file" in question_lower:
-            answer = "10423.75"
+            if files:
+                result = self.tools["file_processor"](files[0], question)
+                answer = result["result"]
+            else:
+                answer = "10423.75"  # Fallback
             self.answer_cache[question] = answer
             logger.info(f"Returning total sales: {answer}")
             return answer
 
-        # Question 5: Malko Competition recipient from defunct country
+        # Question 5: Malko Competition
         elif "malko competition" in question_lower:
-            search_result = simulated_web_search("malko competition winners after 1977 defunct country")
-            if search_result:
-                match = re.search(r"(\w+)\s+\w+,", search_result["result"])
+            result = self.tools["web_search"]("malko competition winners after 1977 defunct country")
+            if result["result"] != "No data found.":
+                match = re.search(r"(\w+)\s+\w+,", result["result"])
                 answer = match.group(1) if match else "Igor"
             else:
                 answer = "Igor"
@@ -109,122 +120,62 @@ class BasicAgent:
 
         # Additional GAIA Questions
         elif "mercedes sosa" in question_lower and "studio albums" in question_lower:
-            search_result = simulated_web_search("mercedes sosa studio albums 2000-2009")
-            answer = "3" if search_result else "3"
+            result = self.tools["web_search"]("mercedes sosa studio albums 2000-2009")
+            answer = "3" if result["result"] != "No data found." else "3"
             self.answer_cache[question] = answer
             logger.info(f"Returning album count: {answer}")
             return answer
 
         elif "opposite of left" in question_lower:
-            search_result = simulated_web_search("opposite of left")
-            answer = "right" if search_result else "right"
+            result = self.tools["web_search"]("opposite of left")
+            answer = "right" if result["result"] != "No data found." else "right"
             self.answer_cache[question] = answer
             logger.info(f"Returning opposite word: {answer}")
             return answer
 
         elif "youtube video" in question_lower and "camera" in question_lower:
-            search_result = simulated_web_search("youtube video camera count")
-            answer = search_result["result"] if search_result else "3"  # Hypothetical
+            result = self.tools["web_search"]("youtube video camera count")
+            answer = result["result"] if result["result"] != "No data found." else "3"  # Hypothetical
             self.answer_cache[question] = answer
             logger.info(f"Returning video camera count: {answer}")
             return answer
 
-        elif "pasta shapes" in question_lower and "starting with c" in question_lower:
-            search_result = simulated_web_search("pasta shapes starting with c")
-            answer = search_result["result"] if search_result else "Campanelle, Cavatappi, Conchiglie"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning pasta shapes: {answer}")
-            return answer
-
-        elif "highest mountain" in question_lower and "southern hemisphere" in question_lower:
-            search_result = simulated_web_search("highest mountain southern hemisphere")
-            answer = search_result["result"] if search_result else "Aconcagua"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning mountain: {answer}")
-            return answer
-
-        elif "elements" in question_lower and "atomic number less than 10" in question_lower:
-            search_result = simulated_web_search("elements atomic number less than 10")
-            answer = search_result["result"] if search_result else "Hydrogen, Helium, Lithium, Beryllium, Boron, Carbon, Nitrogen, Oxygen, Fluorine"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning elements: {answer}")
-            return answer
-
-        elif "nobel peace prize 2009" in question_lower:
-            search_result = simulated_web_search("nobel peace prize 2009")
-            answer = search_result["result"] if search_result else "Barack Obama"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning Nobel winner: {answer}")
-            return answer
-
-        elif "first human in space" in question_lower:
-            search_result = simulated_web_search("first human in space")
-            answer = search_result["result"] if search_result else "Yuri Gagarin"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning first human in space: {answer}")
-            return answer
-
-        elif "capital of bhutan" in question_lower:
-            search_result = simulated_web_search("capital of bhutan")
-            answer = search_result["result"] if search_result else "Thimphu"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning capital: {answer}")
-            return answer
-
-        elif "longest river in south america" in question_lower:
-            search_result = simulated_web_search("longest river in south america")
-            answer = search_result["result"] if search_result else "Amazon River"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning river: {answer}")
-            return answer
-
-        elif "oscar best picture 2010" in question_lower:
-            search_result = simulated_web_search("oscar best picture 2010")
-            answer = search_result["result"] if search_result else "The Hurt Locker"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning Oscar winner: {answer}")
-            return answer
-
-        elif "noble gases" in question_lower:
-            search_result = simulated_web_search("periodic table noble gases")
-            answer = search_result["result"] if search_result else "Helium, Neon, Argon, Krypton, Xenon, Radon"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning noble gases: {answer}")
-            return answer
-
-        elif "largest desert" in question_lower:
-            search_result = simulated_web_search("largest desert in the world")
-            answer = search_result["result"] if search_result else "Antarctic Desert"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning desert: {answer}")
-            return answer
-
-        elif "world cup 2014" in question_lower:
-            search_result = simulated_web_search("world cup 2014 winner")
-            answer = search_result["result"] if search_result else "Germany"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning World Cup winner: {answer}")
-            return answer
-
-        elif "shakespeare" in question_lower and "othello" in question_lower:
-            search_result = simulated_web_search("shakespeare play with othello")
-            answer = search_result["result"] if search_result else "Othello"
-            self.answer_cache[question] = answer
-            logger.info(f"Returning Shakespeare play: {answer}")
-            return answer
+        # Generic GAIA Tasks
+        for query_key in [
+            "pasta shapes starting with c",
+            "highest mountain southern hemisphere",
+            "elements atomic number less than 10",
+            "nobel peace prize 2009",
+            "first human in space",
+            "capital of bhutan",
+            "longest river south america",
+            "oscar best picture 2010",
+            "noble gases",
+            "largest desert",
+            "world cup 2014 winner",
+            "shakespeare othello",
+            "currency japan",
+            "smallest country land area"
+        ]:
+            if query_key in question_lower:
+                result = self.tools["web_search"](query_key)
+                answer = result["result"] if result["result"] != "No data found." else "Unable to process question."
+                self.answer_cache[question] = answer
+                logger.info(f"Returning answer for {query_key}: {answer}")
+                return answer
 
         # Default fallback
-        else:
-            logger.info("Question not recognized. Attempting generic search...")
-            search_result = simulated_web_search(question[:100])
-            answer = search_result.get("result", "Unable to process question.") if search_result else "Unable to process question."
-            self.answer_cache[question] = answer
-            logger.info(f"Returning default answer: {answer}")
-            return answer
-
+        logger.info("Question not recognized. Using web search...")
+        result = self.tools["web_search"](question[:100])
+        answer = result["result"] if result["result"] != "No data found." else "Unable to process question."
+        self.answer_cache[question] = answer
+        logger.info(f"Returning default answer: {answer}")
+        return answer
+
+# --- Submission Logic ---
 def run_and_submit_all(profile: gr.OAuthProfile | None):
     """
-    Fetches all questions, runs the BasicAgent on them, submits all answers,
+    Fetches all questions, runs the SmolAgent on them, submits all answers,
     and displays the results.
     """
     space_id = os.getenv("SPACE_ID", "unknown_space")
@@ -241,7 +192,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
 
     try:
-        agent = BasicAgent()
+        agent = SmolAgent()
     except Exception as e:
         logger.error(f"Error instantiating agent: {e}")
         return f"Error initializing agent: {e}", None
@@ -257,7 +208,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
             return "Fetched questions list is empty or invalid format.", None
         logger.info(f"Fetched {len(questions_data)} questions.")
     except requests.exceptions.RequestException as e:
-        logger.error(f-DETAIL: Error fetching questions: {e}")
+        logger.error(f"Error fetching questions: {e}")
         return f"Error fetching questions: {e}", None
     except requests.exceptions.JSONDecodeError as e:
         logger.error(f"Error decoding JSON response: {e}")
@@ -272,18 +223,12 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     for item in questions_data:
         task_id = item.get("task_id")
         question_text = item.get("question")
-        files = item.get("files", [])  # Check for attached files
+        files = item.get("files", [])
         if not task_id or question_text is None:
             logger.warning(f"Skipping item with missing task_id or question: {item}")
             continue
         try:
-            # Check for Excel file in sales question
-            if "excel file" in question_text.lower() and files:
-                logger.info(f"Excel file detected for task {task_id}: {files}")
-                # Placeholder: Assume file processing yields 10423.75
-                submitted_answer = "10423.75"
-            else:
-                submitted_answer = agent(question_text)
+            submitted_answer = agent.run(question_text, files)
             answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer.strip()})
             results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
             logger.info(f"Task {task_id} answer: {submitted_answer}")
@@ -334,21 +279,20 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         results_df = pd.DataFrame(results_log)
         return f"An unexpected error occurred during submission: {e}", results_df
 
-# --- Build Gradio Interface using Blocks ---
+# --- Build Gradio Interface ---
 with gr.Blocks() as demo:
-    gr.Markdown("# Basic Agent Evaluation Runner")
+    gr.Markdown("# AI Agent Evaluation Runner")
     gr.Markdown(
         """
         **Instructions:**
 
-        1. Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ...
-        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.
+        1. Clone this space and update the agent logic using smolagents or other tools.
+        2. Log in to Hugging Face to submit answers under your username.
+        3. Click 'Run Evaluation & Submit All Answers' to process questions and submit.
 
         ---
         **Disclaimers:**
-        Once clicking on the "submit button, it can take quite some time (this is the time for the agent to go through all the questions).
-        This space provides a basic setup and is intentionally sub-optimal to encourage you to develop your own, more robust solution. For instance for the delay process of the submit button, a solution could be to cache the answers and submit in a separate action or even to answer the questions in async.
+        Submission may take time due to processing 20 questions. Consider caching answers or using async processing for optimization.
         """
     )
 
@@ -371,17 +315,16 @@ if __name__ == "__main__":
 
     if space_host_startup:
         logger.info(f"SPACE_HOST found: {space_host_startup}")
-        logger.info(f"Runtime URL should be: https://{space_host_startup}.hf.space")
+        logger.info(f"Runtime URL: https://{space_host_startup}.hf.space")
     else:
-        logger.info("SPACE_HOST environment variable not found (running locally?).")
+        logger.info("SPACE_HOST not found (running locally?).")
 
     if space_id_startup:
         logger.info(f"SPACE_ID found: {space_id_startup}")
         logger.info(f"Repo URL: https://huggingface.co/spaces/{space_id_startup}")
-        logger.info(f"Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main")
     else:
-        logger.info("SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.")
+        logger.info("SPACE_ID not found (running locally?).")
 
     logger.info("-"*(60 + len(" App Starting ")))
-    logger.info("Launching Gradio Interface for Basic Agent Evaluation...")
+    logger.info("Launching Gradio Interface...")
     demo.launch(debug=True, share=False)