Additional websearch tools.

basic_agent.py

@@ -5,7 +5,11 @@ from ac_tools import DuckDuckGoSearchToolWH
 import requests
 import os
 from PIL import Image
+import wikipedia
+
 from transformers import pipeline
+import requests
+from bs4 import BeautifulSoup
 
 
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
@@ -52,6 +56,59 @@ def download_file(task_id: str, filename: str) -> str:
         raise RuntimeError(error_msg)
 
 
+class WikipediaSearchTool(Tool):
+    name = "wikipedia_search"
+    description = "Searches Wikipedia and returns a short summary of the most relevant article."
+    inputs = {
+        "query": {"type": "string", "description": "The search term or topic to look up on Wikipedia."}
+    }
+    output_type = "string"
+
+    def __init__(self, summary_sentences=3):
+        super().__init__()
+        self.summary_sentences = summary_sentences
+
+    def forward(self, query: str) -> str:
+        try:
+            page_title = wikipedia.search(query)[0]
+            page = wikipedia.page(page_title)
+            return f"**{page.title}**\n\n{page.content}"
+        except IndexError:
+            return "No Wikipedia results found for that query."
+        except Exception as e:
+            return f"Error during Wikipedia search: {e}"
+
+
+class WebpageReaderTool(Tool):
+    name = "read_webpage"
+    description = "Fetches the text content from a given URL and returns the main body text."
+    inputs = {
+        "url": {"type": "string", "description": "The URL of the webpage to read."}
+    }
+    output_type = "string"
+
+    def forward(self, url: str) -> str:
+        try:
+            headers = {
+                "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
+                              "(KHTML, like Gecko) Chrome/114.0.0.0 Safari/537.36"
+            }
+            response = requests.get(url, headers=headers, timeout=10)
+            response.raise_for_status()
+
+            soup = BeautifulSoup(response.text, "html.parser")
+
+            # Extract visible text (ignore scripts/styles)
+            for tag in soup(["script", "style", "noscript"]):
+                tag.extract()
+            text = soup.get_text(separator="\n")
+            cleaned = "\n".join(line.strip() for line in text.splitlines() if line.strip())
+
+            return cleaned[:5000]  # Optionally limit to 5,000 chars
+        except Exception as e:
+            return f"Error reading webpage: {e}"
+
+
 class BasicAgent:
     def __init__(self):
         print("BasicAgent initialized.")
@@ -71,7 +128,9 @@ class BasicSmolAgent:
         if not model:
             model = HfApiModel()
         search_tool = DuckDuckGoSearchToolWH()
-        self.agent = CodeAgent(tools=[search_tool], model=model, max_steps=10, additional_authorized_imports=['pandas'])
+        wiki_tool = WikipediaSearchTool()
+        webpage_tool = WebpageReaderTool()
+        self.agent = CodeAgent(tools=[search_tool, wiki_tool, webpage_tool], model=model, max_steps=10, additional_authorized_imports=['pandas'])
         self.prompt = ("The question is the following:\n ```{}```"
                        " YOUR FINAL ANSWER should be a number OR as few words as possible OR a comma separated list of numbers and/or strings."
                        " If you are asked for a number, don't use comma to write your number neither use units"

requirements.txt

@@ -1,6 +1,8 @@
 --extra-index-url https://download.pytorch.org/whl/cpu
 
 gradio[oauth]
+wikipedia
+bs4
 requests
 openai
 python-dotenv