Rename new_agent.py to agent.py

new_agent.py → agent.py

@@ -1,25 +1,42 @@
 import base64
+from pathlib import Path
 import os
 import gradio as gr
 import requests
 import pandas as pd
+from PIL import Image
+from io import BytesIO
 from dotenv import load_dotenv
 from smolagents import (
     InferenceClientModel,
     CodeAgent,
+    ToolCallingAgent,
     VisitWebpageTool,
     DuckDuckGoSearchTool,
     WikipediaSearchTool,
-    tool,
 )
-from langchain_community.document_loaders import ArxivLoader
+from smolagents.utils import encode_image_base64, make_image_url
 
 load_dotenv()
 
 token = os.environ["HF_KEY"]
-# model_id = "deepseek-ai/DeepSeek-R1-0528"
-model_id = "meta-llama/Llama-3.3-70B-Instruct"
-
+# model_id = "meta-llama/Llama-3.3-70B-Instruct"
+
+manager_model = "deepseek-ai/DeepSeek-R1-0528"
+web_model = "meta-llama/Llama-3.3-70B-Instruct"
+
+web_agent = ToolCallingAgent(
+    model=InferenceClientModel(model_id=web_model, provider="nebius", token=token),
+    tools=[
+        VisitWebpageTool(),
+        DuckDuckGoSearchTool(),
+        WikipediaSearchTool(),
+    ],
+    name="web_agent",
+    description="Browses the web to find information",
+    verbosity_level=0,
+    max_steps=10,
+)
 
 # (Keep Constants as is)
 # --- Constants ---
@@ -30,7 +47,9 @@ DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 # ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
 class BasicAgent:
     def __init__(self):
-        model = InferenceClientModel(model_id=model_id, provider="nebius", token=token)
+        model = InferenceClientModel(
+            model_id=manager_model, provider="nebius", token=token
+        )
         self.agent = CodeAgent(
             tools=[
                 VisitWebpageTool(),
@@ -38,6 +57,7 @@ class BasicAgent:
                 WikipediaSearchTool(),
             ],
             model=model,
+            managed_agents=[web_agent],
             additional_authorized_imports=[
                 "pandas",
                 "numpy",
@@ -71,27 +91,22 @@ class BasicAgent:
             try:
                 if file_path.endswith(".xlsx"):
                     df = pd.read_excel(file_path)
-                    file_info = f"\n[Attached Excel file preview]:\n{df.head(500).to_string(index=False)}\n"
+                    file = df.to_string()
                 elif file_path.endswith(".py"):
-                    with open(file_path, "r", encoding="utf-8") as f:
-                        code = "".join([line for _, line in zip(range(500), f)])
-                    file_info = f"\n[Attached Python file]:\n{code}\n"
-                elif file_path.endswith(".png"):
-                    with open(file_path, "rb") as img_file:
-                        b64_string = base64.b64encode(img_file.read()).decode("utf-8")
-                    file_info = f"\n[Attached PNG image as base64]:\n{b64_string}\n"
+                    file = Path(file_path).read_text()
+                # elif file_path.endswith(".png"):
+                #     with open(file_path, "rb") as img_file:
+                #         image = Image.open(BytesIO(img_file))
+                #     file = image
                 else:
                     with open(file_path, "r", encoding="utf-8") as f:
-                        content = f.read(2000)
-                    file_info = (
-                        f"\n[Attached file content (first 2000 chars)]:\n{content}\n"
-                    )
+                        file = f.read(2000)
             except Exception as e:
-                file_info = f"\n[Could not read attached file: {e}]\n"
+                file = f"\n[Could not read attached file: {e}]\n, Skip the question and move to next one"
         answer = self.agent.run(
             f"""You are a general AI assistant. You can use the provided tools and websearch for finding answers. Some questions may include attached files like excel, python code, or images. Include them while evaluating for answer. I will ask you a question. Report your thoughts, and finish your answer. 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 such as $ or percent sign unless specified otherwise. If you are asked for a string, don't use articles, neither abbreviations (e.g. for cities), and write the digits in plain text unless specified otherwise. If you are asked for a comma separated list, apply the above rules depending of whether the element to be put in the list is a number or a string.
             {question}
-            {file_info}
+            additional_args={"file":{file}}
             """,
         )
         print(f"Agent returning answer: {answer}")