Update app.py

app.py

@@ -3,6 +3,8 @@ import time
 import streamlit as st
 from smolagents import CodeAgent, HfApiModel, tool
 from huggingface_hub import InferenceClient
+import requests
+from bs4 import BeautifulSoup
 
 # Retrieve Hugging Face token
 hf_token = os.getenv("HF_TOKEN")
@@ -15,20 +17,56 @@ client = InferenceClient(token=hf_token)
 # Custom tools for SmolAgents
 @tool
 def search_harry_potter_lore(query: str) -> str:
-    """Search for Harry Potter-related lore or facts from external sources.
+    """Search for Harry Potter-related lore or facts across the entire Harry Potter Fandom site.
+    
     Args:
         query: A specific question or topic about Harry Potter lore.
+    
     Returns:
         A concise and informative response based on the query.
     """
-    # Placeholder for a real implementation using an API or web scraping
-    # In production, you could integrate with fandom APIs or use libraries like BeautifulSoup.
-    if "horcrux" in query.lower():
-        return "A Horcrux is a dark magical object created by splitting one's soul. Tom Riddle created seven of them."
-    elif "chamber of secrets" in query.lower():
-        return "The Chamber of Secrets was built by Salazar Slytherin to purge the school of those he deemed unworthy to study magic."
-    else:
-        return "The information you're looking for is not available in my current resources."
+    headers = {"User-Agent": "Mozilla/5.0"}
+    # Construct the search URL for the Harry Potter Fandom site.
+    search_url = f"https://harrypotter.fandom.com/wiki/Special:Search?query={query}"
+    
+    try:
+        # Fetch the search results page.
+        search_response = requests.get(search_url, headers=headers)
+        if search_response.status_code != 200:
+            return f"Error: Received status code {search_response.status_code} from search."
+        
+        search_soup = BeautifulSoup(search_response.text, 'html.parser')
+        
+        # Look for the first link that appears to be an article.
+        article_url = None
+        for a in search_soup.find_all("a", href=True):
+            href = a["href"]
+            # We want links that start with /wiki/ but skip those that contain "Special:"
+            if href.startswith("/wiki/") and "Special:" not in href:
+                article_url = "https://harrypotter.fandom.com" + href
+                break
+        
+        if not article_url:
+            return "No results found for your query."
+        
+        # Fetch the article page.
+        article_response = requests.get(article_url, headers=headers)
+        if article_response.status_code != 200:
+            return f"Error: Received status code {article_response.status_code} from the article page."
+        
+        article_soup = BeautifulSoup(article_response.text, 'html.parser')
+        
+        # Extract the first meaningful paragraph.
+        paragraphs = article_soup.find_all("p")
+        for p in paragraphs:
+            text = p.get_text().strip()
+            if len(text) > 50:  # A simple threshold to ensure the paragraph is informative.
+                return text
+        
+        return "Couldn't extract detailed lore from the article."
+    
+    except Exception as e:
+        return f"An error occurred: {str(e)}"
 
 # Initialize the SmolAgent model
 model = HfApiModel(model_id="meta-llama/Llama-3.2-3B-Instruct", token=hf_token)