feat: implement agent with smolagents and add caching of answers and results

README.md

@@ -1,5 +1,5 @@
 ---
-title: Template Final Assignment
+title: HF Agent Course Final Assignment
 emoji: 🕵🏻‍♂️
 colorFrom: indigo
 colorTo: indigo

app.py

@@ -3,21 +3,15 @@ import gradio as gr
 import requests
 import inspect
 import pandas as pd
+from dotenv import load_dotenv
+from src.agent import BasicAgent
+
+load_dotenv()
 
 # (Keep Constants as is)
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-# --- Basic Agent Definition ---
-# ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
-class BasicAgent:
-    def __init__(self):
-        print("BasicAgent initialized.")
-    def __call__(self, question: str) -> str:
-        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
 
 def run_and_submit_all( profile: gr.OAuthProfile | None):
     """
@@ -28,7 +22,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code
 
     if profile:
-        username= f"{profile.username}"
+        username = f"{profile.username}"
         print(f"User logged in: {username}")
     else:
         print("User not logged in.")
@@ -69,16 +63,45 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
         print(f"An unexpected error occurred fetching questions: {e}")
         return f"An unexpected error occurred fetching questions: {e}", None
 
+    # 2.1 Load cached results if available
+    cached_results = None
+    cached_answers = None
+    if os.path.exists("cached_results.csv"):
+        try:
+            cached_results = pd.read_csv("cached_results.csv")
+            print(f"Loaded cached results from cached_results.csv with {len(cached_results)} entries.")
+        except Exception as e:
+            print(f"Error loading cached results: {e}")
+            return f"Error loading cached results: {e}", None
+
+    if os.path.exists("cached_answers.csv"):
+        try:
+            cached_answers = pd.read_csv("cached_answers.csv")
+            print(f"Loaded cached answers from cached_answers.csv with {len(cached_answers)} entries.")
+        except Exception as e:
+            print(f"Error loading cached answers: {e}")
+            return f"Error loading cached answers: {e}", None
+
     # 3. Run your Agent
     results_log = []
     answers_payload = []
+
+    # If cached results or answers are available, append them to the log
+    if cached_results is not None:
+        results_log = cached_results.to_dict(orient="records")
+    if cached_answers is not None:
+        answers_payload = cached_answers.to_dict(orient="records")
+
     print(f"Running agent on {len(questions_data)} questions...")
-    for item in questions_data:
+    for item in questions_data[0:20]:
         task_id = item.get("task_id")
         question_text = item.get("question")
         if not task_id or question_text is None:
             print(f"Skipping item with missing task_id or question: {item}")
             continue
+        if not cached_answers.empty and any(d['task_id'] == task_id for d in answers_payload):
+            print(f"Skipping task {task_id} as it is already answered in cached answers.")
+            continue
         try:
             submitted_answer = agent(question_text)
             answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
@@ -91,6 +114,14 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
         print("Agent did not produce any answers to submit.")
         return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
 
+    # 3.1 Store results to file system
+    results_df = pd.DataFrame(results_log)
+    results_df.to_csv("cached_results.csv", index=False)
+
+    answers_df = pd.DataFrame(answers_payload)
+    answers_df.to_csv("cached_answers.csv", index=False)
+    print("Results saved to file system")
+
     # 4. Prepare Submission 
     submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
     status_update = f"Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..."

src/agent.py

@@ -0,0 +1,27 @@
+import os
+
+import yaml
+from smolagents import OpenAIServerModel, CodeAgent, DuckDuckGoSearchTool, VisitWebpageTool
+
+
+model = OpenAIServerModel(
+    model_id="claude-3-5-haiku-20241022",
+    api_base="https://api.anthropic.com/v1/",
+    api_key=os.environ["ANTROPHIC_API_KEY"],
+)
+
+agent = CodeAgent(
+    model=model,
+    tools=[DuckDuckGoSearchTool(), VisitWebpageTool()],
+    max_steps=10,
+    additional_authorized_imports=["time", "numpy", "pandas"]
+)
+
+class BasicAgent:
+    def __init__(self):
+        print("BasicAgent initialized.")
+    def __call__(self, question: str) -> str:
+        print(f"Agent received question (first 50 chars): {question[:50]}...")
+        answer = agent.run(question)
+        print(f"Agent returning answer: {answer}")
+        return answer