update

level_3_agent.py

@@ -0,0 +1,209 @@
+"""LangGraph Agent for Level 1, 2, 3 Reasoning Tasks"""
+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_community.document_loaders import WikipediaLoader
+from langchain_community.document_loaders import ArxivLoader
+from langchain_community.vectorstores import SupabaseVectorStore
+from langchain_core.messages import SystemMessage
+from langchain_core.tools import tool
+from langchain.tools.retriever import create_retriever_tool
+from supabase.client import Client, create_client
+
+load_dotenv()
+
+@tool
+def multiply(a: int, b: int) -> int:
+    """Multiply two numbers.
+    Args:
+        a: first int
+        b: second int
+    """
+    return a * b
+
+@tool
+def add(a: int, b: int) -> int:
+    """Add two numbers.
+    
+    Args:
+        a: first int
+        b: second int
+    """
+    return a + b
+
+@tool
+def subtract(a: int, b: int) -> int:
+    """Subtract two numbers.
+    
+    Args:
+        a: first int
+        b: second int
+    """
+    return a - b
+
+@tool
+def divide(a: int, b: int) -> int:
+    """Divide two numbers.
+    
+    Args:
+        a: first int
+        b: second int
+    """
+    if b == 0:
+        raise ValueError("Cannot divide by zero.")
+    return a / b
+
+@tool
+def modulus(a: int, b: int) -> int:
+    """Get the modulus of two numbers.
+    
+    Args:
+        a: first int
+        b: second int
+    """
+    return a % b
+
+@tool
+def wiki_search(query: str) -> str:
+    """Search Wikipedia for a query and return maximum 2 results.
+    
+    Args:
+        query: The search query."""
+    search_docs = WikipediaLoader(query=query, load_max_docs=2).load()
+    formatted_search_docs = "\n\n---\n\n".join(
+        [
+            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content}\n</Document>'
+            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.
+    
+    Args:
+        query: The search query."""
+    search_docs = TavilySearchResults(max_results=3).invoke(query=query)
+    formatted_search_docs = "\n\n---\n\n".join(
+        [
+            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content}\n</Document>'
+            for doc in search_docs
+        ])
+    return {"web_results": formatted_search_docs}
+
+@tool
+def arvix_search(query: str) -> str:
+    """Search Arxiv for a query and return maximum 3 result.
+    
+    Args:
+        query: The search query."""
+    search_docs = ArxivLoader(query=query, load_max_docs=3).load()
+    formatted_search_docs = "\n\n---\n\n".join(
+        [
+            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content[:1000]}\n</Document>'
+            for doc in search_docs
+        ])
+    return {"arvix_results": formatted_search_docs}
+
+
+# Load the system prompt
+# Assuming we are in model09, the path to system_prompt is RobotPai/system_prompt.txt
+prompt_path = os.path.join(os.path.dirname(__file__), "RobotPai", "system_prompt.txt")
+try:
+    with open(prompt_path, "r", encoding="utf-8") as f:
+        system_prompt = f.read()
+except FileNotFoundError:
+    system_prompt = (
+        "You are a helpful assistant tasked with answering questions using a set of tools.\n\n"
+        "Your final answer must strictly follow this format:\n"
+        "FINAL ANSWER: [ANSWER]\n\n"
+        "Only write the answer in that exact format. Do not explain anything. Do not include any other text.\n\n"
+        "If you are provided with a similar question and its final answer, and the current question is **exactly the same**, then simply return the same final answer without using any tools.\n\n"
+        "Only use tools if the current question is different from the similar one.\n\n"
+        "Examples:\n"
+        "- FINAL ANSWER: FunkMonk\n"
+        "- FINAL ANSWER: Paris\n"
+        "- FINAL ANSWER: 128\n\n"
+        "If you do not follow this format exactly, your response will be considered incorrect."
+    )
+
+sys_msg = SystemMessage(content=system_prompt)
+
+# 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",
+)
+
+question_search_tool = create_retriever_tool(
+    retriever=vector_store.as_retriever(),
+    name="question_search",
+    description="A tool to retrieve similar questions from a vector store.",
+)
+
+tools = [
+    multiply,
+    add,
+    subtract,
+    divide,
+    modulus,
+    wiki_search,
+    web_search,
+    arvix_search,
+    question_search_tool,
+]
+
+def build_graph(provider: str = "google"):
+    """Build the ReAct graph for Level 1, 2, 3 tasks"""
+    if provider == "google":
+        llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash", temperature=0)
+    elif provider == "groq":
+        llm = ChatGroq(model="qwen-qwq-32b", temperature=0)
+    elif provider == "huggingface":
+        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'.")
+        
+    llm_with_tools = llm.bind_tools(tools)
+
+    def assistant(state: MessagesState):
+        """Assistant node that dynamically thinks and uses tools"""
+        # We prepend the system message so the LLM respects the strict output constraints
+        return {"messages": [llm_with_tools.invoke([sys_msg] + state["messages"])]}
+
+    builder = StateGraph(MessagesState)
+    builder.add_node("assistant", assistant)
+    builder.add_node("tools", ToolNode(tools))
+    
+    # Setup the ReAct loop
+    builder.add_edge(START, "assistant")
+    builder.add_conditional_edges(
+        "assistant",
+        tools_condition, # Routes to "tools" if there are tool calls, otherwise "END"
+    )
+    builder.add_edge("tools", "assistant")
+
+    return builder.compile()
+
+if __name__ == "__main__":
+    # Little test to show it compiles and works
+    graph = build_graph()
+    print("Level 1-3 Agent successfully built!")