feat: gemini-model support

.gradio/certificate.pem

@@ -0,0 +1,31 @@
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----

agent.py

@@ -0,0 +1,60 @@
+from smolagents import (
+    CodeAgent,
+    DuckDuckGoSearchTool,
+    FinalAnswerTool,
+    PythonInterpreterTool,
+    VisitWebpageTool,
+    WikipediaSearchTool,
+    tool,
+)
+from smolagents.models import Model
+import google.generativeai as genai
+from dotenv import load_dotenv
+import os
+
+load_dotenv()
+
+class GeminiModel(Model):
+    def __init__(self):
+        self.api_key = os.getenv("GEMINI_API_KEY")
+        self.model_name = os.getenv("GEMINI_MODEL", "gemini-pro")
+
+        if not self.api_key:
+            raise ValueError("GEMINI_API_KEY not found in .env")
+
+        genai.configure(api_key=self.api_key)
+        self.model = genai.GenerativeModel(self.model_name)
+
+    def complete(self, prompt: str) -> str:
+        try:
+            response = self.model.generate_content(prompt)
+            return response.text.strip() if hasattr(response, "text") else str(response)
+        except Exception as e:
+            return f"Error generating content: {e}"
+
+class GaiaAgent:
+    """
+    An agent designed to answer questions using a combination of tools,
+    including search engines, web page access, a Python interpreter, and more.
+    """
+
+    def __init__(self):
+        print("GaiaAgent initialized with tools.")
+
+        gemini_model = GeminiModel()
+
+        tools = [
+            DuckDuckGoSearchTool(),
+            VisitWebpageTool(),
+            WikipediaSearchTool(),
+            PythonInterpreterTool(),
+            FinalAnswerTool(),
+        ]
+
+        self.agent = CodeAgent(model=gemini_model, tools=tools)
+
+    def __call__(self, task_id: str, question: str) -> str:
+        print(f"Agent received {task_id=}\n{question[:50]=}...")
+        answer = self.agent.run(question)
+        print(f"Agent returning answer: {answer}")
+        return answer

app.py

@@ -1,6 +1,6 @@
 import os
 
-import gaia_agent
+import agent
 import gradio as gr
 import logic
 import pandas as pd
@@ -9,7 +9,9 @@ from dotenv import load_dotenv
 load_dotenv()
 
 
-def run_and_submit_all(profile: gr.OAuthProfile | None):
+def run_and_submit_all(
+    profile: gr.OAuthProfile | None,
+) -> tuple[str, pd.DataFrame | None]:
     """Fetches all questions, runs the BasicAgent on them, submits all answers,
     and displays the results.
 
@@ -39,7 +41,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
 
     # 1. Instantiate Agent
     try:
-        agent = gaia_agent.GaiaAgent()
+        gaia_agent = agent.GaiaAgent()
     except Exception as e:
         print(f"Error instantiating agent: {e}")
         return f"Error initializing agent: {e}", None
@@ -51,7 +53,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         return str(e), None
 
     # 3. Run the Agent
-    results_log, answers_payload = logic.run_agent(agent, questions_data)
+    results_log, answers_payload = logic.run_agent(gaia_agent, questions_data)
     if not answers_payload:
         print("Agent did not produce any answers to submit.")
         return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
@@ -136,4 +138,4 @@ if __name__ == "__main__":
     print("-" * (60 + len(" App Starting ")) + "\n")
 
     print("Launching Gradio Interface for Basic Agent Evaluation...")
-    gaia_ui.launch(debug=True, share=False)
+    gaia_ui.launch(debug=True, share=True)

gaia_agent.py

@@ -1,33 +0,0 @@
-class GaiaAgent:
-    """
-    A basic agent that receives a question and returns a fixed answer.
-
-    This class serves as a placeholder or a simple baseline agent for testing
-    and demonstration purposes. It does not perform any sophisticated
-    reasoning or information retrieval.
-    """
-
-    def __init__(self):
-        """
-        Initializes the GaiaAgent.
-
-        Currently, this constructor simply prints a message to the console.
-        In a more complex implementation, this method might load a model,
-        connect to a database, or perform other setup tasks.
-        """
-        print("BasicAgent initialized.")
-
-    def __call__(self, question: str) -> str:
-        """
-        Processes a question and returns a fixed answer.
-
-        Args:
-            question: The question to be processed.
-
-        Returns:
-            A fixed string representing the agent's answer.
-        """
-        print(f"Agent received question (first 50 chars): {question[:50]}...")
-        fixed_answer = "This is a default answer."
-        print(f"Agent returning fixed answer: {fixed_answer}")
-        return fixed_answer

logic.py

@@ -2,7 +2,7 @@ from typing import Dict, List, Tuple
 
 import pandas as pd
 import requests
-from gaia_agent import GaiaAgent
+from agent import GaiaAgent
 from pandas import DataFrame
 
 # --- Constants ---
@@ -113,8 +113,9 @@ def submit_answers(submission_data: dict, results_log: list) -> Tuple[str, DataF
         return status_message, results_df
 
 
-def run_agent(agent: GaiaAgent, 
-              questions_data: List[Dict]) -> Tuple[List[Dict], List[Dict]]:
+def run_agent(
+    gaia_agent: GaiaAgent, questions_data: List[Dict]
+) -> Tuple[List[Dict], List[Dict]]:
     """Runs the agent on a list of questions and returns the results and answers.
 
     This function iterates through a list of questions, runs the provided agent on each
@@ -122,7 +123,7 @@ def run_agent(agent: GaiaAgent,
     agent execution and returns the results log and the answers payload.
 
     Args:
-        agent (GaiaAgent): An instance of the GaiaAgent class, which is responsible for
+        gaia_agent (GaiaAgent): An instance of the GaiaAgent class, which is responsible for
             generating answers to the questions.
         questions_data (List[Dict]): A list of dictionaries, where each dictionary
             represents a question and contains at least the 'task_id' and 'question' keys.
@@ -145,7 +146,7 @@ def run_agent(agent: GaiaAgent,
             print(f"⚠️ Skipping invalid item (missing task_id or question): {item}")
             continue
         try:
-            submitted_answer = agent(question_text)
+            submitted_answer = gaia_agent(task_id, question_text)
             answers_payload.append(
                 {"task_id": task_id, "submitted_answer": submitted_answer}
             )

requirements.txt

@@ -2,4 +2,7 @@ gradio
 gradio[oauth]
 requests
 python-dotenv
-pandas
+pandas
+smolagents
+wikipedia-api
+google-generativeai