Update agent.py

agent.py

@@ -1,213 +1,266 @@
-"""LangGraph Agent"""
-import os
 from dotenv import load_dotenv
-from langgraph.graph import START, StateGraph, MessagesState
-from langgraph.prebuilt import tools_condition
-from langgraph.prebuilt import ToolNode
-from langchain_google_genai import ChatGoogleGenerativeAI
-from langchain_groq import ChatGroq
-from langchain_huggingface import ChatHuggingFace, HuggingFaceEndpoint, HuggingFaceEmbeddings
-from langchain_community.tools.tavily_search import TavilySearchResults
+
+from langchain_openai import ChatOpenAI
+from langchain_core.tools import tool
 from langchain_community.document_loaders import WikipediaLoader
 from langchain_community.document_loaders import ArxivLoader
-from langchain_community.vectorstores import SupabaseVectorStore
+from langchain_community.tools.tavily_search import TavilySearchResults
+from langchain_tavily import TavilyExtract
+from youtube_transcript_api import YouTubeTranscriptApi
+
 from langchain_core.messages import SystemMessage, HumanMessage
-from langchain_core.tools import tool
-from langchain.tools.retriever import create_retriever_tool
-from supabase.client import Client, create_client
+from langgraph.graph import START, StateGraph, MessagesState
+from langgraph.prebuilt import ToolNode
+from langgraph.prebuilt import tools_condition
+import base64
+import httpx
+
 
 load_dotenv()
 
 @tool
-def multiply(a: int, b: int) -> int:
-    """Multiply two numbers.
+def add(a: int, b: int) -> int:
+    """
+    Add b to a.
+    
     Args:
-        a: first int
-        b: second int
+        a: first int number
+        b: second int number
     """
-    return a * b
+    return a + b
 
 @tool
-def add(a: int, b: int) -> int:
-    """Add two numbers.
+def substract(a: int, b: int) -> int:
+    """
+    Subtract b from a.
     
     Args:
-        a: first int
-        b: second int
+        a: first int number
+        b: second int number
     """
-    return a + b
+    return a - b
 
 @tool
-def subtract(a: int, b: int) -> int:
-    """Subtract two numbers.
+def multiply(a: int, b: int) -> int:
+    """
+    Multiply a by b.
     
     Args:
-        a: first int
-        b: second int
+        a: first int number
+        b: second int number
     """
-    return a - b
+    return a * b
 
 @tool
 def divide(a: int, b: int) -> int:
-    """Divide two numbers.
+    """
+    Divide a by b.
     
     Args:
-        a: first int
-        b: second int
+        a: first int number
+        b: second int number
     """
     if b == 0:
-        raise ValueError("Cannot divide by zero.")
+        raise ValueError("Can't divide by zero.")
     return a / b
 
 @tool
-def modulus(a: int, b: int) -> int:
-    """Get the modulus of two numbers.
+def mod(a: int, b: int) -> int:
+    """
+    Remainder of a devided by b.
     
     Args:
-        a: first int
-        b: second int
+        a: first int number
+        b: second int number
     """
     return a % b
 
 @tool
 def wiki_search(query: str) -> str:
-    """Search Wikipedia for a query and return maximum 2 results.
+    """
+    Search Wikipedia.
     
     Args:
-        query: The search query."""
-    search_docs = WikipediaLoader(query=query, load_max_docs=2).load()
-    formatted_search_docs = "\n\n---\n\n".join(
+        query: what to search for
+    """
+    search_docs = WikipediaLoader(query=query, load_max_docs=3).load()
+    formatted_search_docs = "".join(
         [
-            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content}\n</Document>'
+            f'<START source="{doc.metadata["source"]}">{doc.page_content[:1000]}<END>'
             for doc in search_docs
         ])
     return {"wiki_results": formatted_search_docs}
 
 @tool
-def web_search(query: str) -> str:
-    """Search Tavily for a query and return maximum 3 results.
+def arvix_search(query: str) -> str:
+    """
+    Search arXiv which is online archive of preprint and postprint manuscripts 
+    for different fields of science.
     
     Args:
-        query: The search query."""
-    search_docs = TavilySearchResults(max_results=3).invoke(query=query)
-    formatted_search_docs = "\n\n---\n\n".join(
+        query: what to search for
+    """
+    search_docs = ArxivLoader(query=query, load_max_docs=3).load()
+    formatted_search_docs = "".join(
         [
-            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content}\n</Document>'
+            f'<START source="{doc.metadata["source"]}">{doc.page_content[:1000]}<END>'
             for doc in search_docs
         ])
-    return {"web_results": formatted_search_docs}
+    return {"arvix_results": formatted_search_docs}
 
 @tool
-def arvix_search(query: str) -> str:
-    """Search Arxiv for a query and return maximum 3 result.
+def web_search(query: str) -> str:
+    """
+    Search WEB.
     
     Args:
-        query: The search query."""
-    search_docs = ArxivLoader(query=query, load_max_docs=3).load()
-    formatted_search_docs = "\n\n---\n\n".join(
+        query: what to search for
+    """
+    search_docs = TavilySearchResults(max_results=3, include_answer=True).invoke({"query": query})
+    formatted_search_docs = "".join(
         [
-            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content[:1000]}\n</Document>'
+            f'<START source="{doc["url"]}">{doc["content"][:1000]}<END>'
             for doc in search_docs
         ])
-    return {"arvix_results": formatted_search_docs}
-
-
-
-# load the system prompt from the file
-with open("system_prompt.txt", "r", encoding="utf-8") as f:
-    system_prompt = f.read()
-
-# System message
-sys_msg = SystemMessage(content=system_prompt)
+    return {"web_results": formatted_search_docs}
 
-# build a retriever
-embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2") #  dim=768
-supabase: Client = create_client(
-    os.environ.get("SUPABASE_URL"), 
-    os.environ.get("SUPABASE_SERVICE_KEY"))
-vector_store = SupabaseVectorStore(
-    client=supabase,
-    embedding= embeddings,
-    table_name="documents",
-    query_name="match_documents_langchain",
-)
-create_retriever_tool = create_retriever_tool(
-    retriever=vector_store.as_retriever(),
-    name="Question Search",
-    description="A tool to retrieve similar questions from a vector store.",
-)
+@tool
+def open_web_page(url: str) -> str:
+    """
+    Open web page and get its content.
+    
+    Args:
+        url: web page url in ""
+    """
+    search_docs = TavilyExtract().invoke({"urls": [url]})
+    formatted_search_docs = f'<START source="{search_docs["results"][0]["url"]}">{search_docs["results"][0]["raw_content"][:1000]}<END>'
+    return {"web_page_content": formatted_search_docs}
 
+@tool
+def youtube_transcript(url: str) -> str:
+    """
+    Get transcript of YouTube video.
+    Args:
+        url: YouTube video url in ""
+    """    
+    video_id = url.partition("https://www.youtube.com/watch?v=")[2]
+    transcript = YouTubeTranscriptApi.get_transcript(video_id)
+    transcript_text = " ".join([item["text"] for item in transcript])
+    return {"youtube_transcript": transcript_text}
 
 
 tools = [
-    multiply,
     add,
-    subtract,
+    substract,
+    multiply,
     divide,
-    modulus,
+    mod,
     wiki_search,
-    web_search,
     arvix_search,
+    web_search,
+    open_web_page,
+    youtube_transcript,
 ]
 
-# Build graph function
-def build_graph(provider: str = "groq"):
-    """Build the graph"""
-    # Load environment variables from .env file
-    if provider == "google":
-        # Google Gemini
-        llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash", temperature=0)
-    elif provider == "groq":
-        # Groq https://console.groq.com/docs/models
-        llm = ChatGroq(model="qwen-qwq-32b", temperature=0) # optional : qwen-qwq-32b gemma2-9b-it
-    elif provider == "huggingface":
-        # TODO: Add huggingface endpoint
-        llm = ChatHuggingFace(
-            llm=HuggingFaceEndpoint(
-                url="https://api-inference.huggingface.co/models/Meta-DeepLearning/llama-2-7b-chat-hf",
-                temperature=0,
-            ),
-        )
-    else:
-        raise ValueError("Invalid provider. Choose 'google', 'groq' or 'huggingface'.")
-    # Bind tools to LLM
-    llm_with_tools = llm.bind_tools(tools)
+# System prompt
+system_prompt = f"""
+You are a general AI assistant. I will ask you a question.
+First, provide a step-by-step explanation of your reasoning to arrive at the answer.
+Then, respond with your final answer in a single line, formatted as follows: "FINAL ANSWER: [YOUR FINAL ANSWER]".
+[YOUR FINAL ANSWER] should be a number, a string, or a comma-separated list of numbers and/or strings, depending on the question.
+If the answer is a number, do not use commas or units (e.g., $, %) unless specified.
+If the answer is a string, do not use articles or abbreviations (e.g., for cities), and write digits in plain text unless specified.
+If the answer is a comma-separated list, apply the above rules for each element based on whether it is a number or a string.
+"""
+system_message = SystemMessage(content=system_prompt)
+
+# Build graph
+def build_graph():
+    """Build LangGrapth graph of agent."""
 
-    # Node
+    # Language model and tools
+    llm = ChatOpenAI(
+        model="gpt-4.1",
+        temperature=0,
+        max_retries=2
+    )
+    llm_with_tools = llm.bind_tools(tools, strict=True)
+
+    # Nodes
     def assistant(state: MessagesState):
-        """Assistant node"""
-        return {"messages": [llm_with_tools.invoke(state["messages"])]}
-    
-    def retriever(state: MessagesState):
-        """Retriever node"""
-        similar_question = vector_store.similarity_search(state["messages"][0].content)
-        example_msg = HumanMessage(
-            content=f"Here I provide a similar question and answer for reference: \n\n{similar_question[0].page_content}",
-        )
-        return {"messages": [sys_msg] + state["messages"] + [example_msg]}
+        """Assistant node."""
+        return {"messages": [llm_with_tools.invoke([system_message] + state["messages"])]}
 
+    # Graph
     builder = StateGraph(MessagesState)
-    builder.add_node("retriever", retriever)
     builder.add_node("assistant", assistant)
     builder.add_node("tools", ToolNode(tools))
-    builder.add_edge(START, "retriever")
-    builder.add_edge("retriever", "assistant")
-    builder.add_conditional_edges(
-        "assistant",
-        tools_condition,
-    )
+    builder.add_edge(START, "assistant")
+    builder.add_conditional_edges("assistant", tools_condition)
     builder.add_edge("tools", "assistant")
 
     # Compile graph
     return builder.compile()
 
-# test
+
+# Testing and solving particular tasks
 if __name__ == "__main__":
-    question = "When was a picture of St. Thomas Aquinas first added to the Wikipedia page on the Principle of double effect?"
-    # Build the graph
-    graph = build_graph(provider="groq")
-    # Run the graph
-    messages = [HumanMessage(content=question)]
-    messages = graph.invoke({"messages": messages})
-    for m in messages["messages"]:
-        m.pretty_print()
+
+    agent = build_graph()
+    
+    question = """
+Review the chess position provided in the image. It is black's turn. 
+Provide the correct next move for black which guarantees a win. 
+Please provide your response in algebraic notation.
+"""
+    content_urls = {
+        "image": "https://agents-course-unit4-scoring.hf.space/files/cca530fc-4052-43b2-b130-b30968d8aa44",
+        "audio": None
+    }
+    
+    # Define user message and add all the content
+    content = [
+        {
+            "type": "text",
+            "text": question
+        }
+    ]
+    if content_urls["image"]:
+        image_data = base64.b64encode(httpx.get(content_urls["image"]).content).decode("utf-8")
+        content.append(
+            {
+                "type": "image",
+                "source_type": "base64",
+                "data": image_data,
+                "mime_type": "image/jpeg"
+            }
+        )
+    if content_urls["audio"]:
+        audio_data = base64.b64encode(httpx.get(content_urls["audio"]).content).decode("utf-8")
+        content.append(
+            {
+                "type": "audio",
+                "source_type": "base64",
+                "data": audio_data,
+                "mime_type": "audio/wav"
+            }
+        )
+    messages = {
+        "role": "user",
+        "content": content
+    }
+
+    # Run agent on the question
+    messages = agent.invoke({"messages": messages})
+    for message in messages["messages"]:
+        message.pretty_print()
+
+    answer = messages["messages"][-1].content
+    index = answer.find("FINAL ANSWER: ")
+    
+    print("\n")
+    print("="*30)
+    if index == -1:
+        print(answer)
+    print(answer[index+14:])
+    print("="*30)