Upload app.py

app.py

@@ -0,0 +1,63 @@
+# 🧰 Required libraries
+from smolagents import CodeAgent, DuckDuckGoSearchTool, HfApiModel, tool
+from huggingface_hub import InferenceClient
+import requests
+from bs4 import BeautifulSoup
+import json
+import os
+import gradio as gr
+
+HF_TOKEN = os.getenv("HF_API_KEY")
+client = InferenceClient(token=HF_TOKEN)
+
+# 🛠️ Custom tool to get Hugging Face top daily paper
+@tool
+def get_hugging_face_top_daily_paper() -> str:
+    """
+    Returns the title of the most upvoted paper on Hugging Face daily papers.
+    """
+    try:
+        url = "https://huggingface.co/papers"
+        response = requests.get(url)
+        response.raise_for_status()
+        soup = BeautifulSoup(response.content, "html.parser")
+
+        # Use more general selector
+        containers = soup.find_all('div', attrs={'data-props': True})
+        top_paper = ""
+
+        for container in containers:
+            data_props = container.get('data-props', '')
+            if data_props:
+                try:
+                    json_data = json.loads(data_props.replace('"', '"'))
+                    if 'dailyPapers' in json_data:
+                        top_paper = json_data['dailyPapers'][0]['title']
+                        break
+                except json.JSONDecodeError:
+                    continue
+
+        return top_paper or "No top paper found."
+    except requests.exceptions.RequestException as e:
+        return f"Error occurred while fetching the paper: {e}"
+
+# 🚀 Run the agent with a sample query
+if __name__ == "__main__":
+    # 🤖 Load a model from Hugging Face Hub
+    model = HfApiModel(model_id="Qwen/Qwen2.5-Coder-32B-Instruct", api_key=HF_TOKEN)
+
+    # 🔧 Initialize tools
+    search_tool = DuckDuckGoSearchTool()
+
+    # 🧠 Build the agent
+    agent = CodeAgent(
+        tools=[search_tool, get_hugging_face_top_daily_paper],
+        model=model,
+        additional_authorized_imports=["requests", "bs4", "json"]
+    )
+
+    # 🖥️ Web interface
+    def run_agent_interface(query):
+        return agent.run(query)
+
+    gr.Interface(fn=run_agent_interface, inputs="text", outputs="text").launch()