feat: initail working agent

.gitignore

@@ -0,0 +1 @@
+.env

.python-version

@@ -0,0 +1 @@
+3.11

app.py

@@ -3,20 +3,43 @@ import gradio as gr
 import requests
 import inspect
 import pandas as pd
+from typing import Any
 
+from src.final_assignment_template.agent import manager_agent
 # (Keep Constants as is)
 # --- Constants ---
+
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
 # --- Basic Agent Definition ---
 # ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
+# class BasicAgent:
+#     def __init__(self):
+#         print("BasicAgent initialized.")
+#     def __call__(self, question: str) -> str:
+#         print(f"Agent received question (first 50 chars): {question[:50]}...")
+#         fixed_answer = "This is a default answer."
+#         print(f"Agent returning fixed answer: {fixed_answer}")
+#         return fixed_answer
+
 class BasicAgent:
     def __init__(self):
         print("BasicAgent initialized.")
-    def __call__(self, question: str) -> str:
-        print(f"Agent received question (first 50 chars): {question[:50]}...")
-        fixed_answer = "This is a default answer."
-        print(f"Agent returning fixed answer: {fixed_answer}")
+    def __call__(self, item:dict[str,Any]) -> str:
+        print(item)
+        question_text = item.get("question") or ''
+        task_id = item.get("task_id") or ''
+        file_name = item.get("file_name") or None
+
+        fixed_answer = ''
+        if task_id and file_name:
+          print('With task_id')
+          print(task_id)
+          fixed_answer = manager_agent.run(f"""<Task>{question_text}</Task>\n<TaskID>{task_id}</TaskID>""")
+        else:
+          fixed_answer = manager_agent.run(f'<Task>{question_text}</Task>')
+          print(f'---------------------fixed_answer----------------\n{fixed_answer}')
+
         return fixed_answer
 
 def run_and_submit_all( profile: gr.OAuthProfile | None):
@@ -73,14 +96,22 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     results_log = []
     answers_payload = []
     print(f"Running agent on {len(questions_data)} questions...")
+
     for item in questions_data:
         task_id = item.get("task_id")
         question_text = item.get("question")
+        file_name = item.get("file_name")
+        file_data = None
+        #  or file_name != ''
         if not task_id or question_text is None:
             print(f"Skipping item with missing task_id or question: {item}")
             continue
         try:
-            submitted_answer = agent(question_text)
+            # if file_name != '':
+            #   file_data = get_file(task_id)
+            #   item['file_data'] = file_data
+
+            submitted_answer = agent(item)
             answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
             results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
         except Exception as e:

pyproject.toml

@@ -0,0 +1,24 @@
+[project]
+name = "final-assignment-template"
+version = "0.1.0"
+description = "Add your description here"
+readme = "README.md"
+authors = [
+    { name = "Muhammad Maaz Uddin", email = "[email protected]" }
+]
+requires-python = ">=3.11"
+dependencies = [
+    "gradio",
+    "itsdangerous>=2.2.0",
+    "langchain-community>=0.3.22",
+    "litellm>=1.67.2",
+    "requests",
+    "smolagents>=1.14.0",
+]
+
+[project.scripts]
+final-assignment-template = "final_assignment_template:main"
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"

src/final_assignment_template/__init__.py

@@ -0,0 +1,2 @@
+def main() -> None:
+    print("Hello from final-assignment-template!")

src/final_assignment_template/agent.py

@@ -0,0 +1,55 @@
+from smolagents import load_tool, Tool, tool, ToolCallingAgent, CodeAgent, GoogleSearchTool,FinalAnswerTool,PythonInterpreterTool , LiteLLMModel,  VisitWebpageTool,  DuckDuckGoSearchTool
+from litellm import completion
+
+from langchain.agents import load_tools
+from langchain_community.tools.tavily_search import TavilySearchResults
+
+import os
+from src.final_assignment_template.models import openrouter_qwenCoder_model, modelLiteLLm
+from src.final_assignment_template.tools import travily_tool, Video_understanding_tool, image_understanding_tool, get_task_file
+# (Keep Constants as is)
+# --- Constants ---
+
+
+
+
+
+
+web_agent = CodeAgent(
+    model=openrouter_qwenCoder_model,
+    tools=[
+        # GoogleSearchTool(provider="serper"),
+        # DuckDuckGoSearchTool(max_results=10),
+        travily_tool,
+        VisitWebpageTool(),
+    ],
+    name="web_agent",
+    description="""Browses the web to find information""",
+    verbosity_level=1,
+    max_steps=5,
+)
+
+manager_agent = CodeAgent(
+    name="Task_Agent",
+    description="""You will be provided a task and you need to verify before giving final answer
+    You can perform tasks which are text and image based, skip all other
+    """,
+    model=modelLiteLLm,
+    tools=[PythonInterpreterTool(),Video_understanding_tool,image_understanding_tool,get_task_file],
+    managed_agents=[web_agent],
+    additional_authorized_imports=[
+        "json",
+        "pandas",
+        "numpy",
+        "markdown"
+        'math', 'statistics', 're', 'unicodedata', 'random', 
+        'datetime', 'queue', 'time', 'collections', 'stat', 'itertools',
+        'PIL','requests'
+    ],
+    planning_interval=3,
+    verbosity_level=1,
+    # final_answer_checks=[check_reasoning_and_plot],
+    max_steps=5,
+)
+
+

src/final_assignment_template/models.py

@@ -0,0 +1,29 @@
+from smolagents import LiteLLMModel
+import os
+
+
+openrouter_qwenCoder_model = LiteLLMModel(
+    model_id="openrouter/qwen/qwen-2.5-coder-32b-instruct:free",
+    api_base="https://openrouter.ai/api/v1",
+    api_key=os.getenv("OPENROUTER_API_KEY")
+)
+
+modelLiteLLm = LiteLLMModel(
+    model_id="openrouter/deepseek/deepseek-r1:free",
+     api_base="https://openrouter.ai/api/v1",
+    api_key=os.getenv("OPENROUTER_API_KEY")
+)
+
+
+videoLiteLLm = LiteLLMModel(
+    model_id="openrouter/google/gemini-2.0-flash-exp:free",
+     api_base="https://openrouter.ai/api/v1",
+    api_key=os.getenv("OPENROUTER_API_KEY")
+)
+
+
+imageLiteLLm = LiteLLMModel(
+    model_id="openrouter/meta-llama/llama-4-maverick:free",
+     api_base="https://openrouter.ai/api/v1",
+    api_key=os.getenv("OPENROUTER_API_KEY")
+)

src/final_assignment_template/tools.py

@@ -0,0 +1,115 @@
+from smolagents import  Tool, tool
+
+from langchain_community.tools.tavily_search import TavilySearchResults
+
+import requests
+import inspect
+import pandas as pd
+from PIL import Image
+from io import BytesIO
+import base64
+
+
+from src.final_assignment_template.models import videoLiteLLm, imageLiteLLm
+
+DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
+
+travily_tool = Tool.from_langchain(TavilySearchResults(max_results=25,))
+
+from smolagents import Tool
+
+# class SearchTool(Tool):
+#     name = "SearchTool"
+#     description = """
+#     This is a tool that returns the most downloaded model of a given task on the Hugging Face Hub.
+#     It returns the name of the checkpoint."""
+#     inputs = {
+#         "task": {
+#             "type": "string",
+#             "description": "the task category (such as text-classification, depth-estimation, etc)",
+#         }
+#     }
+#     output_type = "string"
+
+#     def forward(self, task: str):
+#         from huggingface_hub import list_models
+
+#         model = next(iter(list_models(filter=task, sort="downloads", direction=-1)))
+#         return model.id
+
+# model_downloads_tool = HFModelDownloadsTool()
+
+
+@tool
+def Video_understanding_tool(query:str)->str:
+  """
+    This tool for understanding or finding something in the video link.
+
+    Args:
+        query: link with your query.
+  """
+  print("processcing vidoe ",query)
+  messages =[ {"role": "user", "content": [{"type": "text", "text": query}]} ]
+  resp = videoLiteLLm(messages)
+  return resp.content or 'No data'
+
+
+
+@tool
+def get_task_file(task_id:str)->requests.models.Response:
+    """
+    This tool is for get the task file using task_id.
+    it will return the request response and then this response will be used for other tools.
+    
+    Args:
+        task_id: Task ID
+    """
+
+    url = f"{DEFAULT_API_URL}/files/{task_id}"
+    print(url)
+    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)
+    return response
+
+@tool
+def image_understanding_tool(query:str,response:requests.models.Response)->str:
+  """
+    This tool for understanding or perform any query on the image.
+    Provide the image base64 image data
+    
+
+    Args:
+        query: Query for the image.
+        response : The return value from the get_task_file which returns the response.
+  """
+  print("processcing image ")
+  
+  
+  image = Image.open(BytesIO(response.content)).convert("RGB")
+
+  buffered = BytesIO()
+  image.save(buffered, format="PNG")  # change format if necessary
+  img_bytes = buffered.getvalue()
+  img_b64 = base64.b64encode(img_bytes).decode('utf-8')
+
+  print(img_b64)
+  messages =[ {
+        "role": "user",
+        "content": [
+            {"type": "text", "text": query},
+            {
+                "type": "image_url",
+                "image_url": {
+                    "url": img_b64,
+                    "format": "image/png"  # Adjust MIME type if necessary
+                }
+            }
+        ]
+    } ]
+  resp = imageLiteLLm(messages)
+  print(resp.content)
+  return resp.content or 'No data'
+
+