Upload 3 files

agent.py

@@ -6,9 +6,9 @@ from __future__ import annotations
 
 import asyncio
 import os
-import time
-import datetime
-from typing import Any, Sequence, Callable, List, Optional
+from typing import Any, Sequence, Callable, List
+from datetime import datetime
+from agents import RunHooks  # for lifecycle hooks
 
 from dotenv import load_dotenv
 from agents import Agent, Runner, FunctionTool, Tool
@@ -23,12 +23,6 @@ from tools import (
     duckduckgo_search,
 )
 
-# ---------------------------------------------------------------------------
-# Logging Utility
-# ---------------------------------------------------------------------------
-def log(msg):
-    print(f"[{datetime.datetime.now():%Y-%m-%d %H:%M:%S}] {msg}")
-
 # ---------------------------------------------------------------------------
 # Load the added system prompt
 # ---------------------------------------------------------------------------
@@ -85,37 +79,67 @@ def _build_agent(extra_tools: Sequence[FunctionTool] | None = None) -> Agent:
     )
 
 
+class LoggingHooks(RunHooks):
+    """RunHooks to log question start, model used, and each tool‐call step."""
+    def __init__(self):
+        self.step_counter = 0
+
+    async def on_agent_start(self, context, agent):
+        qnum = context.context.get("question_number")
+        qtext = context.context.get("question_text")
+        model = agent.model
+        ts = datetime.now().isoformat()
+        print(f"[{ts}] [Question {qnum}] Starting agent (model={model}) for question: '{qtext}'")
+
+    async def on_tool_start(self, context, agent, tool):
+        self.step_counter += 1
+        qnum = context.context.get("question_number")
+        ts = datetime.now().isoformat()
+        print(f"[{ts}] [Question {qnum}] Step {self.step_counter}: Invoking tool '{tool.name}'")
+
+    async def on_tool_end(self, context, agent, tool, result):
+        qnum = context.context.get("question_number")
+        ts = datetime.now().isoformat()
+        print(f"[{ts}] [Question {qnum}] Step {self.step_counter}: Tool '{tool.name}' completed")
+
+
 class GAIAAgent:
     """Thin synchronous wrapper around an asynchronous Agents SDK agent."""
 
     def __init__(self, *, extra_tools: Sequence[FunctionTool] | None = None):
         self._agent = _build_agent(extra_tools=extra_tools)
-        # Store the model id for logging
-        self.model_id = _select_model()
-
-    async def _arun(self, question: str, q_index: Optional[int] = None) -> str:
-        q_num = q_index + 1 if q_index is not None else "?"
-        log(f"Answering question {q_num}:")
-        log(f"    Question: {question!r}")
-        log(f"    Model: {self.model_id}")
 
-        t0 = time.time()
-        try:
+    async def _arun(self, question: str, context_data=None, hooks=None) -> str:
+        # Pass context and hooks to Runner.run if provided
+        if context_data is not None and hooks is not None:
+            result = await Runner.run(
+                self._agent,
+                question,
+                context=context_data,
+                hooks=hooks
+            )
+        else:
             result = await Runner.run(self._agent, question)
-            duration = time.time() - t0
-            log(f"    Total duration: {duration:.2f} seconds.")
-        except Exception as e:
-            log(f"    Error during answer: {e}")
-            raise
         return str(result.final_output).strip()
 
-    def __call__(self, question: str, q_index: Optional[int] = None, **kwargs: Any) -> str:
+    def __call__(self, question: str, question_number: int | None = None, **_kwargs) -> str:
+        # Prepare logging context if a question_number is given
+        context_data = None
+        hooks = None
+        if question_number is not None:
+            context_data = {
+                "question_number": question_number,
+                "question_text": question
+            }
+            hooks = LoggingHooks()
+
         try:
             loop = asyncio.get_running_loop()
         except RuntimeError:
-            return asyncio.run(self._arun(question, q_index=q_index))
+            # No running loop: use asyncio.run
+            return asyncio.run(self._arun(question, context_data, hooks))
         else:
-            return loop.run_until_complete(self._arun(question, q_index=q_index))
+            return loop.run_until_complete(self._arun(question, context_data, hooks))
 
 
 def gaia_agent(*, extra_tools: Sequence[FunctionTool] | None = None) -> GAIAAgent:

app.py

@@ -2,34 +2,30 @@ import os
 import gradio as gr
 import requests
 import pandas as pd
-import datetime
 
 # --- Our Agent ---
 from agent import gaia_agent
 
-# Logging utility
-def log(msg):
-    print(f"[{datetime.datetime.now():%Y-%m-%d %H:%M:%S}] {msg}")
-
 # Debugging level. If DEBUG=0 then DEBUG will be False. If DEBUG=1 then DEBUG will be True.
 DEBUG = os.getenv("DEBUG", "0") == "1"
 
+# (Keep Constants as is)
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-def run_and_submit_all(profile: gr.OAuthProfile | None):
+def run_and_submit_all( profile: gr.OAuthProfile | None):
     """
     Fetches all questions, runs the BasicAgent on them, submits all answers,
     and displays the results.
     """
     # --- Determine HF Space Runtime URL and Repo URL ---
-    space_id = os.getenv("SPACE_ID")  # Get the SPACE_ID for sending link to the code
+    space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code
 
     if profile:
-        username = f"{profile.username}"
-        log(f"User logged in: {username}")
+        username= f"{profile.username}"
+        print(f"User logged in: {username}")
     else:
-        log("User not logged in.")
+        print("User not logged in.")
         return "Please Login to Hugging Face with the button.", None
 
     api_url = DEFAULT_API_URL
@@ -39,16 +35,16 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     # 1. Instantiate Agent (now using OpenAI Agents SDK)
     try:
         agent = gaia_agent()
-        log("OpenAI Agent instantiated successfully.")
+        print("OpenAI Agent instantiated successfully.")
     except Exception as e:
-        log(f"Error instantiating agent: {e}")
+        print(f"Error instantiating agent: {e}")
         return f"Error initializing agent: {e}", None
-
+    # In the case of an app running as a hugging Face space, this link points toward your codebase ( useful for others so please keep it public)
     agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
-    log(agent_code)
+    print(agent_code)
 
     # 2. Fetch Questions
-    log(f"Fetching questions from: {questions_url}")
+    print(f"Fetching questions from: {questions_url}")
     import json
 
     try:
@@ -56,55 +52,70 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         response.raise_for_status()
         questions_data = response.json()
         if not questions_data:
-            log("Fetched questions list is empty.")
+            print("Fetched questions list is empty.")
             return "Fetched questions list is empty or invalid format.", None
-        log(f"Fetched {len(questions_data)} GAIA questions.")
+        print(f"Fetched {len(questions_data)} questions.")
     except json.JSONDecodeError as e:
-        log(f"Error decoding JSON response from questions endpoint: {e}")
-        log(f"Response text: {response.text[:500]}")
+        print(f"Error decoding JSON response from questions endpoint: {e}")
+        print(f"Response text: {response.text[:500]}")
         return f"Error decoding server response for questions: {e}", None
     except requests.exceptions.RequestException as e:
-        log(f"Error fetching questions: {e}")
+        print(f"Error fetching questions: {e}")
         return f"Error fetching questions: {e}", None
     except Exception as e:
-        log(f"An unexpected error occurred fetching questions: {e}")
+        print(f"An unexpected error occurred fetching questions: {e}")
         return f"An unexpected error occurred fetching questions: {e}", None
 
     # 3. Run the Agent
     results_log = []
     answers_payload = []
-    log(f"Running agent on {len(questions_data)} questions...")
-    for idx, item in enumerate(questions_data):
+    print(f"Running agent on {len(questions_data)} questions...")
+    for idx, item in enumerate(questions_data, start=1):
         task_id = item.get("task_id")
         question_text = item.get("question")
         if not task_id or question_text is None:
-            log(f"Skipping item with missing task_id or question: {item}")
+            print(f"Skipping item with missing task_id or question: {item}")
             continue
         try:
-            submitted_answer = agent(question_text, q_index=idx)
+            # pass in question_number for logging hooks
+            submitted_answer = agent(question_text, question_number=idx)
+
+            # --- DEBUG LOGGING ---
             if DEBUG:
-                log(f"[DEBUG] Task {task_id}: Answer type: {type(submitted_answer)}, Value: {repr(submitted_answer)}")
+                print(f"[DEBUG] Task {task_id}: Answer type: {type(submitted_answer)}, Value: {repr(submitted_answer)}")
             else:
-                log(f"[{task_id}] {question_text[:50]}... → {submitted_answer[:40]}")
+                print(f"[{task_id}] {question_text[:50]}... → {submitted_answer[:40]}")
 
+            # Force string type here just in case (defensive)
             submitted_answer = str(submitted_answer).strip()
-            answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
-            results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
+            answers_payload.append({
+                "task_id": task_id,
+                "submitted_answer": submitted_answer
+            })
+            results_log.append({
+                "Task ID": task_id,
+                "Question": question_text,
+                "Submitted Answer": submitted_answer
+            })
         except Exception as e:
-            log(f"Error running agent on task {task_id}: {e}")
-            results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})
+            print(f"Error running agent on task {task_id}: {e}")
+            results_log.append({
+                "Task ID": task_id,
+                "Question": question_text,
+                "Submitted Answer": f"AGENT ERROR: {e}"
+            })
 
     if not answers_payload:
-        log("Agent did not produce any answers to submit.")
+        print("Agent did not produce any answers to submit.")
         return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
 
-    # 4. Prepare Submission
+    # 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}'..."
-    log(status_update)
+    print(status_update)
 
     # 5. Submit
-    log(f"Submitting {len(answers_payload)} answers to: {submit_url}")
+    print(f"Submitting {len(answers_payload)} answers to: {submit_url}")
     try:
         response = requests.post(submit_url, json=submission_data, timeout=60)
         response.raise_for_status()
@@ -116,7 +127,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
             f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
             f"Message: {result_data.get('message', 'No message received.')}"
         )
-        log("Submission successful.")
+        print("Submission successful.")
         results_df = pd.DataFrame(results_log)
         return final_status, results_df
     except requests.exceptions.HTTPError as e:
@@ -127,22 +138,22 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         except requests.exceptions.JSONDecodeError:
             error_detail += f" Response: {e.response.text[:500]}"
         status_message = f"Submission Failed: {error_detail}"
-        log(status_message)
+        print(status_message)
         results_df = pd.DataFrame(results_log)
         return status_message, results_df
     except requests.exceptions.Timeout:
         status_message = "Submission Failed: The request timed out."
-        log(status_message)
+        print(status_message)
         results_df = pd.DataFrame(results_log)
         return status_message, results_df
     except requests.exceptions.RequestException as e:
         status_message = f"Submission Failed: Network error - {e}"
-        log(status_message)
+        print(status_message)
         results_df = pd.DataFrame(results_log)
         return status_message, results_df
     except Exception as e:
         status_message = f"An unexpected error occurred during submission: {e}"
-        log(status_message)
+        print(status_message)
         results_df = pd.DataFrame(results_log)
         return status_message, results_df
 
@@ -180,7 +191,7 @@ with gr.Blocks() as demo:
 if __name__ == "__main__":
     print("\n" + "-"*30 + " App Starting " + "-"*30)
     space_host_startup = os.getenv("SPACE_HOST")
-    space_id_startup = os.getenv("SPACE_ID")  # Get SPACE_ID at startup
+    space_id_startup = os.getenv("SPACE_ID") # Get SPACE_ID at startup
 
     if space_host_startup:
         print(f"✅ SPACE_HOST found: {space_host_startup}")
@@ -188,14 +199,14 @@ if __name__ == "__main__":
     else:
         print("ℹ️  SPACE_HOST environment variable not found (running locally?).")
 
-    if space_id_startup:  # Print repo URLs if SPACE_ID is found
+    if space_id_startup: # Print repo URLs if SPACE_ID is found
         print(f"✅ SPACE_ID found: {space_id_startup}")
         print(f"   Repo URL: https://huggingface.co/spaces/{space_id_startup}")
         print(f"   Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main")
     else:
         print("ℹ️  SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.")
 
-    print("-" * (60 + len(" App Starting ")) + "\n")
+    print("-"*(60 + len(" App Starting ")) + "\n")
 
     print("Launching Gradio Interface for Agent Evaluation…")
     demo.launch(debug=True, share=False)

tools.py

@@ -7,41 +7,12 @@ from __future__ import annotations
 import contextlib
 import io
 import os
-import time
-import datetime
-from typing import TypedDict, List, Union
+from typing import List, Dict
 
 from agents import function_tool
 
-class DuckDuckGoResult(TypedDict):
-    title: str
-    href: str
-    body: str
-
-class SpreadsheetRow(TypedDict):
-    # If you don't know the columns, leave this empty,
-    # but ideally, define them.
-    pass
-
-def log(msg):
-    print(f"[{datetime.datetime.now():%Y-%m-%d %H:%M:%S}] {msg}")
-
-def log_tool_call(func):
-    def wrapper(*args, **kwargs):
-        t0 = time.time()
-        log(f"Step: {func.__name__} started.")
-        try:
-            result = func(*args, **kwargs)
-            log(f"Step: {func.__name__} completed in {time.time() - t0:.2f}s.")
-            return result
-        except Exception as e:
-            log(f"Step: {func.__name__} error: {e}")
-            raise
-    return wrapper
-
 # 1. --------------------------------------------------------------------
 @function_tool
-@log_tool_call
 def python_run(code: str) -> str:
     """Execute trusted Python code and return the captured stdout together with
     the repr() of the last expression (or `_result` variable if set).
@@ -65,8 +36,7 @@ def python_run(code: str) -> str:
 
 # 2. --------------------------------------------------------------------
 @function_tool
-@log_tool_call
-def load_spreadsheet(path: str, sheet: Union[str, int, None] = None) -> List[SpreadsheetRow]:
+def load_spreadsheet(path: str, sheet: str | int | None = None) -> list[Dict[str, str]]:
     """Read .csv, .xls or .xlsx from disk and return rows as list of dictionaries.
 
     Args:
@@ -95,7 +65,6 @@ def load_spreadsheet(path: str, sheet: Union[str, int, None] = None) -> List[Spr
 
 # 3. --------------------------------------------------------------------
 @function_tool
-@log_tool_call
 def youtube_transcript(url: str, lang: str = "en") -> str:
     """Fetch the subtitles of a YouTube video.
 
@@ -115,7 +84,6 @@ def youtube_transcript(url: str, lang: str = "en") -> str:
 
 # 4. --------------------------------------------------------------------
 @function_tool
-@log_tool_call
 def transcribe_audio(path: str, model: str = "whisper-1") -> str:
     """Transcribe an audio file using OpenAI Whisper.
 
@@ -136,7 +104,6 @@ def transcribe_audio(path: str, model: str = "whisper-1") -> str:
 
 # 5. --------------------------------------------------------------------
 @function_tool
-@log_tool_call
 def image_ocr(path: str) -> str:
     """Perform OCR on an image using Tesseract.
 
@@ -153,8 +120,7 @@ def image_ocr(path: str) -> str:
 
 # 6. --------------------------------------------------------------------
 @function_tool
-@log_tool_call
-def duckduckgo_search(query: str, max_results: int = 5) -> List[DuckDuckGoResult]:
+def duckduckgo_search(query: str, max_results: int = 5) -> List[Dict[str, str]]:
     """Search DuckDuckGo and return a list of result dicts with title, href and body.
 
     Args:
@@ -173,4 +139,4 @@ def duckduckgo_search(query: str, max_results: int = 5) -> List[DuckDuckGoResult
                     "body": r.get("body", ""),
                 }
             )
-    return results
+    return results