Update app.py

app.py

@@ -3,38 +3,74 @@ import time
 import gradio as gr
 import requests
 import pandas as pd
-from smolagents import ToolCallingAgent, OpenAIServerModel, DuckDuckGoSearchTool, PythonInterpreterTool, Tool
+from smolagents import ToolCallingAgent, OpenAIServerModel, PythonInterpreterTool, Tool
 
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
 # --- Custom Tools ---
 
+class WebSearchTool(Tool):
+    name = "web_search"
+    description = "Search the web for information. Use for any factual question."
+    inputs = {"query": {"type": "string", "description": "The search query"}}
+    output_type = "string"
+
+    def forward(self, query: str) -> str:
+        try:
+            from ddgs import DDGS
+            with DDGS() as ddgs:
+                results = list(ddgs.text(query, max_results=5))
+            if not results:
+                return "No results found."
+            output = ""
+            for r in results:
+                output += f"Title: {r.get('title', '')}\n"
+                output += f"URL: {r.get('href', '')}\n"
+                output += f"Summary: {r.get('body', '')}\n\n"
+            return output[:3000]
+        except Exception as e:
+            return f"Search error: {e}"
+
+
 class WikipediaTool(Tool):
     name = "wikipedia_search"
-    description = "Search Wikipedia and get article text. Use this when the question mentions Wikipedia or needs encyclopedic facts."
-    inputs = {"query": {"type": "string", "description": "The topic to search on Wikipedia"}}
+    description = "Search Wikipedia directly. Use when the question mentions Wikipedia or needs encyclopedic facts like discographies, biographies, lists."
+    inputs = {"query": {"type": "string", "description": "The Wikipedia article title or topic to search"}}
     output_type = "string"
 
     def forward(self, query: str) -> str:
         try:
+            # First search for the right article
             search_url = (
-                f"https://en.wikipedia.org/w/api.php"
-                f"?action=query&titles={query.replace(' ', '_')}"
-                f"&prop=extracts&explaintext=true&format=json"
+                "https://en.wikipedia.org/w/api.php"
+                f"?action=query&list=search&srsearch={requests.utils.quote(query)}"
+                "&format=json&srlimit=1"
             )
             r = requests.get(search_url, timeout=10)
-            pages = r.json()["query"]["pages"]
+            results = r.json()["query"]["search"]
+            if not results:
+                return "No Wikipedia article found."
+            title = results[0]["title"]
+
+            # Then fetch full article text
+            content_url = (
+                "https://en.wikipedia.org/w/api.php"
+                f"?action=query&titles={requests.utils.quote(title)}"
+                "&prop=extracts&explaintext=true&format=json"
+            )
+            r2 = requests.get(content_url, timeout=10)
+            pages = r2.json()["query"]["pages"]
             page = next(iter(pages.values()))
             text = page.get("extract", "No content found")
-            return text[:4000]
+            return f"Article: {title}\n\n{text[:5000]}"
         except Exception as e:
             return f"Wikipedia error: {e}"
 
 
 class YouTubeTranscriptTool(Tool):
     name = "youtube_transcript"
-    description = "Gets the transcript of a YouTube video. Use when the question contains a YouTube URL or asks about video content."
+    description = "Gets the transcript/captions of a YouTube video. Use when the question contains a YouTube URL."
     inputs = {"url": {"type": "string", "description": "YouTube video URL or video ID"}}
     output_type = "string"
 
@@ -48,7 +84,7 @@ class YouTubeTranscriptTool(Tool):
             else:
                 video_id = url.strip()
             transcript = YouTubeTranscriptApi.get_transcript(video_id)
-            return " ".join([t["text"] for t in transcript])[:4000]
+            return " ".join([t["text"] for t in transcript])[:5000]
         except Exception as e:
             return f"Transcript error: {e}"
 
@@ -64,12 +100,31 @@ class FileDownloadTool(Tool):
             url = f"https://agents-course-unit4-scoring.hf.space/files/{task_id}"
             r = requests.get(url, timeout=15)
             if r.status_code == 200:
-                return r.text[:4000]
+                return r.text[:5000]
             return f"No file found for task_id {task_id}"
         except Exception as e:
             return f"File download error: {e}"
 
 
+class VisitWebpageTool(Tool):
+    name = "visit_webpage"
+    description = "Fetches the full content of a webpage given its URL. Use when you have a specific URL to read."
+    inputs = {"url": {"type": "string", "description": "The URL of the webpage to visit"}}
+    output_type = "string"
+
+    def forward(self, url: str) -> str:
+        try:
+            headers = {"User-Agent": "Mozilla/5.0"}
+            r = requests.get(url, timeout=10, headers=headers)
+            # strip html tags roughly
+            import re
+            text = re.sub(r'<[^>]+>', ' ', r.text)
+            text = re.sub(r'\s+', ' ', text).strip()
+            return text[:5000]
+        except Exception as e:
+            return f"Webpage error: {e}"
+
+
 # --- Agent ---
 
 class BasicAgent:
@@ -82,23 +137,27 @@ class BasicAgent:
         self.agent = ToolCallingAgent(
             model=model,
             tools=[
-                DuckDuckGoSearchTool(),
-                PythonInterpreterTool(),
+                WebSearchTool(),
                 WikipediaTool(),
                 YouTubeTranscriptTool(),
                 FileDownloadTool(),
+                VisitWebpageTool(),
+                PythonInterpreterTool(),
             ],
-            max_steps=4,
+            max_steps=5,
         )
 
     def __call__(self, question: str, task_id: str = "") -> str:
         try:
             prompt = f"""Answer the following question accurately.
-Return ONLY the final answer — no explanation, no punctuation, no extra words.
-If the answer is a number, return just the number.
-If the answer is a name, return just the name.
-If the answer is a list, return comma separated values.
-Task ID (use this with download_file tool if the question references a file): {task_id}
+Return ONLY the final answer with no explanation, no punctuation, no extra words.
+- If the answer is a number, return just the number.
+- If the answer is a name, return just the name.
+- If the answer is a list, return comma separated values in alphabetical order.
+- If the question asks about a YouTube video, use the youtube_transcript tool.
+- If the question mentions Wikipedia, use the wikipedia_search tool.
+- If the question references an attached file, use download_file with the task_id below.
+Task ID: {task_id}
 
 Question: {question}"""
             result = self.agent.run(prompt)
@@ -117,24 +176,20 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         username = f"{profile.username}"
         print(f"User logged in: {username}")
     else:
-        print("User not logged in.")
         return "Please Login to Hugging Face with the button.", None
 
     api_url = DEFAULT_API_URL
     questions_url = f"{api_url}/questions"
     submit_url = f"{api_url}/submit"
 
-    # 1. Instantiate Agent
     try:
         agent = BasicAgent()
     except Exception as e:
-        print(f"Error instantiating agent: {e}")
         return f"Error initializing agent: {e}", None
 
     agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
     print(agent_code)
 
-    # 2. Fetch Questions
     print(f"Fetching questions from: {questions_url}")
     try:
         response = requests.get(questions_url, timeout=15)
@@ -146,7 +201,6 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     except Exception as e:
         return f"Error fetching questions: {e}", None
 
-    # 3. Run Agent on each question
     results_log = []
     answers_payload = []
     print(f"Running agent on {len(questions_data)} questions...")
@@ -160,7 +214,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
             continue
 
         print(f"\n[{i+1}/{len(questions_data)}] Task: {task_id}")
-        print(f"Question: {question_text[:100]}...")
+        print(f"Question: {question_text[:120]}...")
 
         try:
             submitted_answer = agent(question_text, task_id)
@@ -171,15 +225,13 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
             print(f"Error on task {task_id}: {e}")
             results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})
 
-        # Rate limit protection — Groq free tier is 6000 tokens/min
         if i < len(questions_data) - 1:
-            print("Waiting 12s to respect rate limits...")
-            time.sleep(12)
+            print("Waiting 15s for rate limits...")
+            time.sleep(15)
 
     if not answers_payload:
         return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
 
-    # 4. Submit
     submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
     print(f"\nSubmitting {len(answers_payload)} answers...")
 
@@ -217,7 +269,7 @@ with gr.Blocks() as demo:
         **Instructions:**
         1. Log in to your Hugging Face account using the button below.
         2. Click 'Run Evaluation & Submit All Answers' to start.
-        3. The agent will answer all 20 questions and submit. Takes ~5 minutes due to rate limits.
+        3. Takes ~6 minutes for all 20 questions due to rate limits.
         """
     )
 
@@ -233,20 +285,16 @@ 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")
-
     if space_host_startup:
         print(f"✅ SPACE_HOST found: {space_host_startup}")
     else:
         print("ℹ️  SPACE_HOST not found (running locally).")
-
     if space_id_startup:
         print(f"✅ SPACE_ID found: {space_id_startup}")
     else:
         print("ℹ️  SPACE_ID not found (running locally).")
-
     print("-"*(60 + len(" App Starting ")) + "\n")
     print("Launching Gradio Interface...")
     demo.launch(debug=True, share=False)