added tools

app.py

@@ -7,16 +7,144 @@ from tools.final_answer import FinalAnswerTool
 
 from Gradio_UI import GradioUI
 
-# Below is an example of a tool that does nothing. Amaze us with your creativity !
+
+@tool
+def get_top_daily_paper() -> str:
+    """
+    Retrieves the current top upvoted paper from Hugging Face Daily Papers.
+    Returns:
+        str: The title and URL of the paper.
+    """
+    try:
+        import requests
+        from bs4 import BeautifulSoup
+        url = "https://huggingface.co/papers"
+        resp = requests.get(url, timeout=10)
+        resp.raise_for_status()
+        soup = BeautifulSoup(resp.content, "html.parser")
+        
+        # Find the first article (top paper by upvotes)
+        paper = soup.select_one("article")
+        if not paper:
+            return "🔍 Error: No papers found on the page"
+            
+        # Find the title and link within the h3 > a structure
+        title_element = paper.select_one("h3 a")
+        if not title_element:
+            return "🔍 Error: Could not find paper title"
+        
+        title = title_element.get_text(strip=True)
+        link = title_element.get("href")
+        
+        if not link:
+            return "🔍 Error: Could not find paper link"
+            
+        full_url = f"https://huggingface.co{link}"
+        return f"Top Daily Paper: {title} — {full_url}"
+    except Exception as e:
+        return f"🔍 Error fetching top paper: {str(e)}"
+
+@tool
+def get_paper_abstract(paper_url: str) -> str:
+    """
+    Retrieves the abstract from a Hugging Face paper page.
+    Args:
+        paper_url: The URL of the paper page
+    Returns:
+        str: The paper abstract including AI summary if available
+    """
+    try:
+        import requests
+        from bs4 import BeautifulSoup
+        
+        resp = requests.get(paper_url, timeout=10)
+        resp.raise_for_status()
+        soup = BeautifulSoup(resp.content, "html.parser")
+        
+        # Find the abstract section
+        abstract_section = soup.find("h2", string="Abstract")
+        if not abstract_section:
+            return "🔍 Error: Abstract section not found"
+        
+        # Get the parent container of the abstract
+        abstract_container = abstract_section.find_next_sibling("div")
+        if not abstract_container:
+            return "🔍 Error: Abstract content not found"
+        
+        result_parts = []
+        
+        # Look for AI-generated summary (blue box)
+        ai_summary = abstract_container.select_one(".bg-blue-500\\/6 p")
+        if ai_summary:
+            summary_text = ai_summary.get_text(strip=True)
+            result_parts.append(f"🤖 AI Summary: {summary_text}")
+        
+        # Get the main abstract text
+        main_abstract = abstract_container.select_one("p.text-gray-600")
+        if main_abstract:
+            # Clean up the text by removing link artifacts and extra spaces
+            abstract_text = ""
+            for element in main_abstract.descendants:
+                if element.name is None:  # Text node
+                    abstract_text += element.strip() + " "
+            
+            abstract_text = " ".join(abstract_text.split())  # Normalize whitespace
+            result_parts.append(f"📄 Abstract: {abstract_text}")
+        
+        if not result_parts:
+            return "🔍 Error: No abstract content found"
+        
+        return "\n\n".join(result_parts)
+        
+    except Exception as e:
+        return f"🔍 Error fetching abstract: {str(e)}"
+
 @tool
-def my_custom_tool(arg1:str, arg2:int)-> str: #it's import to specify the return type
-    #Keep this format for the description / args / args description but feel free to modify the tool
-    """A tool that does nothing yet 
+def summarize_text(text: str, max_sentences: int = 3, model_name: str = "google/pegasus-cnn_dailymail") -> str:
+    """
+    Summarize a body of text using a Hugging Face Transformers pipeline.
+
     Args:
-        arg1: the first argument
-        arg2: the second argument
+        text: The text to be summarized.
+        max_sentences: Approximate upper limit for the number of sentences in the output.
+        model_name: The Hugging Face model to use for summarization (default is a CNN/DailyMail–fine‑tuned Pegasus).
+
+    Returns:
+        A concise summary string, or an error message.
     """
-    return "What magic will you build ?"
+    try:
+        from transformers import pipeline
+
+        # Load summarization pipeline once (could be optimized by caching)
+        summarizer = pipeline("summarization", model=model_name)
+
+        # Heuristically chunk long text into manageable parts for the model
+        max_chunk = 1024  # tokens; varies by model
+        # Naive chunking, splitting on sentences or whitespace:
+        chunks = [text[i:i + max_chunk] for i in range(0, len(text), max_chunk)]
+
+        # Summarize each chunk
+        summaries = []
+        for chunk in chunks:
+            out = summarizer(chunk,
+                             max_length=max_sentences * 20,
+                             min_length=max_sentences * 10,
+                             do_sample=False)
+            summaries.append(out[0]['summary_text'])
+
+        # Combine chunk-level summaries, optionally resummarize
+        combined = " ".join(summaries)
+        if len(chunks) > 1:
+            final = summarizer(combined,
+                               max_length=max_sentences * 20,
+                               min_length=max_sentences * 10,
+                               do_sample=False)[0]['summary_text']
+            return final
+        else:
+            return summaries[0]
+
+    except Exception as e:
+        return f"Error during summarization: {e}"
 
 @tool
 def get_current_time_in_timezone(timezone: str) -> str:
@@ -55,7 +183,7 @@ with open("prompts.yaml", 'r') as stream:
     
 agent = CodeAgent(
     model=model,
-    tools=[final_answer], ## add your tools here (don't remove final answer)
+    tools=[final_answer,get_current_time_in_timezone,get_top_daily_paper,get_paper_abstract,summarize_text], ## add your tools here (don't remove final answer)
     max_steps=6,
     verbosity_level=1,
     grammar=None,