add agent

agent.py

@@ -1,17 +1,38 @@
-import os
-from typing import TypedDict, Annotated
-from dotenv import load_dotenv
-from langgraph.graph.message import add_messages
-from langchain_core.messages import AnyMessage, HumanMessage, AIMessage
 from langgraph.prebuilt import ToolNode, tools_condition
 from langgraph.graph import START, StateGraph, MessagesState
-from langchain_huggingface import HuggingFaceEndpoint, ChatHuggingFace
-from tools import wiki_search, tavily_search, arxiv_search, add, subtract, multiply, divide, power, sqrt, modulus
+from langchain_openai import AzureChatOpenAI
 
+from config import (
+    MODEL_ENDPOINT,
+    MODEL_KEY,
+    MODEL_NAME,
+    MODEL_API_VERSION,
+)
 
-load_dotenv()
-HF_TOKEN = os.getenv("HF_TOKEN")
+from tools import (
+    wiki_search,
+    tavily_search,
+    arxiv_search,
+    add,
+    subtract,
+    multiply,
+    divide,
+    power,
+    sqrt,
+    modulus,
+    scrape_webpage,
+    analyze_image,
+    is_commutative,
+    commutativity_counterexample_pairs,
+    commutativity_counterexample_elements,
+    find_identity_element,
+    find_inverses,
+    transcribe_audio,
+    execute_source_file,
+    interact_tabular,
+)
 
+# Define tools
 TOOLS = [
     wiki_search,
     tavily_search,
@@ -22,31 +43,48 @@ TOOLS = [
     divide,
     power,
     sqrt,
-    modulus
+    modulus,
+    scrape_webpage,
+    analyze_image,
+    is_commutative,
+    commutativity_counterexample_pairs,
+    commutativity_counterexample_elements,
+    find_identity_element,
+    find_inverses,
+    transcribe_audio,
+    execute_source_file,
+    interact_tabular
 ]
 
-def build_agent():
-    # Define llm from Hugging Face
-    llm = HuggingFaceEndpoint(
-        repo_id="Qwen/Qwen2.5-Coder-32B-Instruct",
-        huggingfacehub_api_token=HF_TOKEN
+
+def build_agent() -> StateGraph:
+    """
+    Build the agent.
+    Returns:
+        StateGraph: The agent graph.
+    """
+    llm = AzureChatOpenAI(
+        azure_deployment=MODEL_NAME,
+        api_version=MODEL_API_VERSION,
+        azure_endpoint=MODEL_ENDPOINT,
+        api_key=MODEL_KEY,
     )
 
-    # Define chat interface and the tools
-    chat = ChatHuggingFace(llm=llm, verbose=True)
-    chat_w_tools = chat.bind_tools(TOOLS)
+    chat_w_tools = llm.bind_tools(TOOLS)
 
-    # Node
+    # Assistant node
     def assistant(state: MessagesState):
         """Assistant node"""
         return {"messages": [chat_w_tools.invoke(state["messages"])]}
 
-
+    # Build graph
     builder = StateGraph(MessagesState)
 
+    # Add nodes
     builder.add_node("assistant", assistant)
     builder.add_node("tools", ToolNode(TOOLS))
 
+    # Add edges
     builder.add_edge(START, "assistant")
     builder.add_conditional_edges(
         "assistant",
@@ -54,5 +92,5 @@ def build_agent():
     )
     builder.add_edge("tools", "assistant")
 
-    # Compile graph
-    return builder.compile()
+    # Compile graph and return it
+    return builder.compile()

app.py

@@ -1,15 +1,30 @@
 import os
-import gradio as gr
+import tempfile
+import atexit
 import requests
-import inspect
+import gradio as gr
 import pandas as pd
 from agent import build_agent
 from config import SYSTEM_PROMPT, SPACE_ID
 from langchain_core.messages import SystemMessage, HumanMessage
 
+
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
+TEMP_FILES = []
+
+
+def cleanup_temp_files():
+    for path in TEMP_FILES:
+        try:
+            os.remove(path)
+        except Exception as e:
+            print(f"Could not delete temp file {path}: {e}")
+
+
+atexit.register(cleanup_temp_files)
+
 
 def get_file(task_id: str) -> requests.Response:
     """I
@@ -20,6 +35,7 @@ def get_file(task_id: str) -> requests.Response:
     response.raise_for_status()
     return response
 
+
 def get_question_data(elem: dict) -> tuple[str, str]:
     """
     Fetches question text and file path if there are any.
@@ -30,15 +46,18 @@ def get_question_data(elem: dict) -> tuple[str, str]:
     """
     question_text = elem["question"]
     file_name = elem["file_name"]
+    file_path = None
 
     if file_name != "":
         task_id = elem["task_id"]
         response = get_file(task_id=task_id)
 
-        file_path = f"data/{file_name}"
+        temp_dir = tempfile.gettempdir()
+        file_path = os.path.join(temp_dir, file_name)
         with open(file_path, "wb") as f:
             f.write(response.content)
-    
+        TEMP_FILES.append(file_path)
+
     return file_path, question_text
 
 
@@ -47,7 +66,8 @@ class BasicAgent:
     def __init__(self):
         self.agent = build_agent()
         print("BasicAgent initialized.")
-    def __call__(self, question: str, file_path: str=None) -> str:
+
+    def __call__(self, question: str, file_path: str = None) -> str:
         messages = [
             SystemMessage(content=SYSTEM_PROMPT),
         ]
@@ -65,17 +85,17 @@ class BasicAgent:
 
         return final_answer
 
-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 = SPACE_ID
 
     if profile:
-        username= f"{profile.username}"
+        username = f"{profile.username}"
         print(f"User logged in: {username}")
     else:
         print("User not logged in.")
@@ -102,16 +122,16 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
         response.raise_for_status()
         questions_data = response.json()
         if not questions_data:
-             print("Fetched questions list is empty.")
-             return "Fetched questions list is empty or invalid format.", None
+            print("Fetched questions list is empty.")
+            return "Fetched questions list is empty or invalid format.", None
         print(f"Fetched {len(questions_data)} questions.")
     except requests.exceptions.RequestException as e:
         print(f"Error fetching questions: {e}")
         return f"Error fetching questions: {e}", None
     except requests.exceptions.JSONDecodeError as e:
-         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
+        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 Exception as e:
         print(f"An unexpected error occurred fetching questions: {e}")
         return f"An unexpected error occurred fetching questions: {e}", None
@@ -132,14 +152,14 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
             results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
             print(f"Task ID: {task_id}, Question: {question_text}, Submitted Answer: {submitted_answer}")
         except Exception as 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}"})
+            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:
         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}'..."
     print(status_update)
@@ -219,11 +239,12 @@ with gr.Blocks() as demo:
         outputs=[status_output, results_table]
     )
 
+
 if __name__ == "__main__":
-    print("\n" + "-"*30 + " App Starting " + "-"*30)
+    print("\n" + "-" * 30 + " App Starting " + "-" * 30)
     # Check for SPACE_HOST and SPACE_ID at startup for information
     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}")
@@ -231,14 +252,15 @@ 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")
+        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("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 Basic Agent Evaluation...")
-    demo.launch(debug=True, share=False)
+    demo.launch(debug=True, share=False)

config.py

@@ -10,4 +10,10 @@ SPACE_ID = os.getenv("SPACE_ID")
 
 with open("system_prompt.yaml", "r") as f:
     SYSTEM_PROMPT = yaml.safe_load(f)
-    SYSTEM_PROMPT = SYSTEM_PROMPT["system_prompt"]
+    SYSTEM_PROMPT = SYSTEM_PROMPT["system_prompt"]
+
+HF_TOKEN = os.getenv("HF_TOKEN")
+MODEL_ENDPOINT = os.getenv("MODEL_ENDPOINT")
+MODEL_KEY = os.getenv("MODEL_KEY")
+MODEL_NAME = os.getenv("MODEL_NAME")
+MODEL_API_VERSION = os.getenv("MODEL_API_VERSION")

tools.py

@@ -1,28 +1,35 @@
-from langchain_core.tools import tool
-from langchain_community.tools.tavily_search import TavilySearchResults
-from langchain_community.document_loaders import WikipediaLoader
-from langchain_community.document_loaders import ArxivLoader
-from config import TAVILY_API_KEY
-import requests
-from bs4 import BeautifulSoup
-from PIL import Image
-from pathlib import Path
 import base64
-from openai import AzureOpenAI
-from config import MODEL_NAME, MODEL_API_VERSION, MODEL_ENDPOINT, MODEL_KEY
-from faster_whisper import WhisperModel
-from typing import Dict
+import io
+import json
 import shutil
 import subprocess as sp
 import tempfile
-import pandas as pd
 import textwrap
-import io
-import json
+from pathlib import Path
+from typing import Dict
+import pandas as pd
+import requests
+from bs4 import BeautifulSoup
+
+from config import (
+    TAVILY_API_KEY,
+    MODEL_NAME,
+    MODEL_API_VERSION,
+    MODEL_ENDPOINT,
+    MODEL_KEY,
+)
 
-#=========================================
+from langchain_core.tools import tool
+from langchain_community.tools.tavily_search import TavilySearchResults
+from langchain_community.document_loaders import WikipediaLoader, ArxivLoader
+from openai import AzureOpenAI
+from faster_whisper import WhisperModel
+
+# =========================================
 # Search Tools
-#=========================================
+# =========================================
+
+
 @tool
 def wiki_search(query: str) -> str:
     """
@@ -38,14 +45,14 @@ def wiki_search(query: str) -> str:
     for doc in docs:
         # Get the standard wiki summary
         wiki_summary = f"\nTitle: {doc.metadata.get('title')}\nURL: {doc.metadata.get('source')}\n\n"
-        
+
         # Scrape and clean the full webpage
         try:
             headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36'}
             response = requests.get(doc.metadata.get('source'), headers=headers)
             response.raise_for_status()
             soup = BeautifulSoup(response.text, 'html.parser')
-            
+
             # Remove unwanted elements
             unwanted_elements = [
                 '.mw-jump-link', '.mw-editsection', '.reference',  # Wiki specific
@@ -56,7 +63,7 @@ def wiki_search(query: str) -> str:
             ]
             for element in soup.select(','.join(unwanted_elements)):
                 element.decompose()
-            
+
             # Get main content area
             content_div = soup.select_one('#mw-content-text')
             if content_div:
@@ -67,18 +74,19 @@ def wiki_search(query: str) -> str:
             else:
                 full_text = soup.get_text(separator='\n', strip=True)
 
-            
             # Combine wiki summary with cleaned webpage content
             combined_result = f"{wiki_summary}\n### Full Article Content ###\n{full_text}"
             results.append(combined_result)
-            
+
         except Exception as e:
+            print(f"Error scraping Wikipedia page: {e}")
             results.append(wiki_summary)
 
     # Join all results with clear separators
-    formatted_results = "\n\n" + "="*20 + "\n\n".join(results)
+    formatted_results = "\n\n" + "=" * 20 + "\n\n".join(results)
     return formatted_results
 
+
 @tool
 def tavily_search(query: str) -> str:
     """
@@ -101,6 +109,7 @@ def tavily_search(query: str) -> str:
 
     return formatted_results
 
+
 @tool
 def arxiv_search(query: str) -> str:
     """
@@ -123,6 +132,7 @@ def arxiv_search(query: str) -> str:
 
     return formatted_results
 
+
 @tool
 def scrape_webpage(url: str) -> str:
     """
@@ -137,20 +147,23 @@ def scrape_webpage(url: str) -> str:
         response = requests.get(url, headers=headers)
         response.raise_for_status()
         soup = BeautifulSoup(response.text, 'html.parser')
-        
+
         # Remove script and style elements
         for script in soup(['script', 'style']):
             script.decompose()
-            
+
         # Get text content
         text = soup.get_text(separator='\n', strip=True)
         return text
     except Exception as e:
         return f"Error scraping webpage: {str(e)}"
 
-#=========================================
+
+# =========================================
 # Math Tools
-#=========================================
+# =========================================
+
+
 @tool
 def add(x: float, y: float) -> float:
     """
@@ -163,6 +176,7 @@ def add(x: float, y: float) -> float:
     """
     return x + y
 
+
 @tool
 def subtract(x: float, y: float) -> float:
     """
@@ -175,6 +189,7 @@ def subtract(x: float, y: float) -> float:
     """
     return x - y
 
+
 @tool
 def multiply(x: float, y: float) -> float:
     """
@@ -187,6 +202,7 @@ def multiply(x: float, y: float) -> float:
     """
     return x * y
 
+
 @tool
 def divide(x: float, y: float) -> float:
     """
@@ -201,6 +217,7 @@ def divide(x: float, y: float) -> float:
         raise ValueError("Cannot divide by zero.")
     return x / y
 
+
 @tool
 def power(x: float, y: float) -> float:
     """
@@ -213,6 +230,7 @@ def power(x: float, y: float) -> float:
     """
     return x ** y
 
+
 @tool
 def sqrt(x: float) -> float:
     """
@@ -226,6 +244,7 @@ def sqrt(x: float) -> float:
         raise ValueError("Cannot calculate square root of a negative number.")
     return x ** 0.5
 
+
 @tool
 def modulus(x: float, y: float) -> float:
     """
@@ -238,6 +257,7 @@ def modulus(x: float, y: float) -> float:
     """
     return x % y
 
+
 @tool
 def is_commutative(set_elements: list, operation_table: list) -> bool:
     """
@@ -255,6 +275,7 @@ def is_commutative(set_elements: list, operation_table: list) -> bool:
                 return False
     return True
 
+
 @tool
 def commutativity_counterexample_pairs(set_elements: list, operation_table: list) -> list:
     """
@@ -273,6 +294,7 @@ def commutativity_counterexample_pairs(set_elements: list, operation_table: list
                 pairs.append((set_elements[i], set_elements[j]))
     return pairs
 
+
 @tool
 def commutativity_counterexample_elements(set_elements: list, operation_table: list) -> str:
     """
@@ -292,6 +314,7 @@ def commutativity_counterexample_elements(set_elements: list, operation_table: l
                 involved.add(set_elements[j])
     return ",".join(sorted(involved))
 
+
 @tool
 def is_associative(set_elements: list, operation_table: list) -> bool:
     """
@@ -317,6 +340,7 @@ def is_associative(set_elements: list, operation_table: list) -> bool:
                     return False
     return True
 
+
 @tool
 def find_identity_element(set_elements: list, operation_table: list) -> str:
     """
@@ -339,6 +363,7 @@ def find_identity_element(set_elements: list, operation_table: list) -> str:
             return candidate
     return ""
 
+
 @tool
 def find_inverses(set_elements: list, operation_table: list) -> dict:
     """
@@ -353,8 +378,6 @@ def find_inverses(set_elements: list, operation_table: list) -> dict:
     identity = find_identity_element(set_elements, operation_table)
     if not identity:
         return {e: None for e in set_elements}
-    idx = {e: i for i, e in enumerate(set_elements)}
-    identity_idx = idx[identity]
     inverses = {}
     for i in range(n):
         found = None
@@ -365,9 +388,12 @@ def find_inverses(set_elements: list, operation_table: list) -> dict:
         inverses[set_elements[i]] = found
     return inverses
 
-#=========================================
+
+# =========================================
 # Image Tools
-#=========================================
+# =========================================
+
+
 @tool
 def analyze_image(question: str, path: str) -> str:
     """
@@ -387,7 +413,7 @@ def analyze_image(question: str, path: str) -> str:
     p = Path(path).expanduser().resolve()
     if not p.exists():
         raise ValueError(f"Image file does not exist: {p}")
-    
+
     mime = "image/png" if p.suffix.lower() == ".png" else "image/jpeg"
     with open(p, "rb") as f:
         base64_image = f"data:{mime};base64,{base64.b64encode(f.read()).decode('utf-8')}"
@@ -407,9 +433,12 @@ def analyze_image(question: str, path: str) -> str:
 
     return response.choices[0].message.content.strip()
 
-#=========================================
+
+# =========================================
 # Audio Tools
-#=========================================
+# =========================================
+
+
 @tool
 def transcribe_audio(path: str) -> str:
     """
@@ -433,21 +462,24 @@ def transcribe_audio(path: str) -> str:
     text = "".join(seg.text for seg in segments).strip()
     return text
 
-#=========================================
+
+# =========================================
 # Code Tools
-#=========================================
+# =========================================
+
 LANG_COMMANDS: Dict[str, callable] = {
-    ".py": lambda s, _:[["python3", s.name]],
-    ".js": lambda s, _:[["node", s.name]],
-    ".ts": lambda s, _:[["deno", "run", "-A", s.name]],
-    ".sh": lambda s, _:[["bash", s.name]],
-    ".rb": lambda s, _:[["ruby", s.name]],
-    ".php": lambda s, _:[["php", s.name]],
-    ".go": lambda s, _:[["go", "run", s.name]]
+    ".py": lambda s, _: [["python3", s.name]],
+    ".js": lambda s, _: [["node", s.name]],
+    ".ts": lambda s, _: [["deno", "run", "-A", s.name]],
+    ".sh": lambda s, _: [["bash", s.name]],
+    ".rb": lambda s, _: [["ruby", s.name]],
+    ".php": lambda s, _: [["php", s.name]],
+    ".go": lambda s, _: [["go", "run", s.name]]
 }
 
+
 @tool
-def execute_source_file(path: str, timeout: int=10) -> str:
+def execute_source_file(path: str, timeout: int = 10) -> str:
     """
     Run the program contained in *path*
     Returns a newline-separated string:
@@ -463,7 +495,7 @@ def execute_source_file(path: str, timeout: int=10) -> str:
     src = Path(path).expanduser().resolve(strict=True)
     if src.suffix not in LANG_COMMANDS:
         raise ValueError(f"Unsupported file extension: {src.suffix}")
-    
+
     # Temp work dir for the program
     work = Path(tempfile.mkdtemp(prefix="exec_tool_"))
     shutil.copy(src, work / src.name)
@@ -490,15 +522,18 @@ def execute_source_file(path: str, timeout: int=10) -> str:
             f"STDOUT: {full_out}\n"
             f"STDERR: {full_err}"
         )
-    
+
     finally:
         shutil.rmtree(work)
 
-#=========================================
+
+# =========================================
 # Tabular data tools
-#=========================================
+# =========================================
+
 MAX_BYTES_RETURN = 200000
 
+
 # Helper functions
 def _load_table(path: Path, sheet: str) -> pd.DataFrame:
     """
@@ -518,6 +553,7 @@ def _load_table(path: Path, sheet: str) -> pd.DataFrame:
         return pd.read_parquet(path)
     raise ValueError(f"Unsupported file extension: {ext}")
 
+
 def _safe_truncate(text: str, limit: int = MAX_BYTES_RETURN) -> tuple[str, bool]:
     """
     Truncate text to a given limit.
@@ -580,7 +616,7 @@ def interact_tabular(file_path: str, operation: str = "summary", sheet: str = "S
         result = buf.getvalue()
     else:
         raise ValueError(f"Unsupported operation: {operation}")
-    
+
     result, truncated = _safe_truncate(result)
 
     info = {