agent/agent.py

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, HumanMessage

from langchain_core.tools import tool
from langchain.tools.retriever import create_retriever_tool
from supabase.client import Client, create_client

from tools.basic_calculator import add, count_substring, divide, modulus, multiply, power, square_root, subtract
from tools.code_interpreter import execute_code_multilang
from tools.document_processing import save_and_read_file,download_file_from_url, extract_text_from_image, analyze_csv_file, analyze_excel_file
from tools.image_processing import analyze_image, transform_image, draw_on_image, generate_simple_image, combine_images
from tools.web_search import arxiv_search, similar_question_search, wiki_search, web_search


load_dotenv() # load environment variables

# load the system prompt from the file
with open("prompts/system_prompt.txt", "r", encoding="utf-8") as f:
    system_prompt = f.read()
print(system_prompt)

# System message
sys_msg = SystemMessage(content=system_prompt)

# build a retriever
embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2") # set the model to generate embeddings;  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 Retriever", description="A tool to retrieve similar questions from a vector store.")

tools = [web_search, wiki_search, similar_question_search, arxiv_search, multiply, add, subtract, divide, modulus, power, square_root, count_substring, save_and_read_file, download_file_from_url, extract_text_from_image, analyze_csv_file, analyze_excel_file, execute_code_multilang, analyze_image, transform_image, draw_on_image, generate_simple_image, combine_images]


# Build the agent graph
def build_graph(provider: str = "huggingface-qwen"):
    """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-qwen":
        llm = ChatHuggingFace(llm=HuggingFaceEndpoint(repo_id = "Qwen/Qwen2.5-Coder-32B-Instruct"))
    elif provider == "huggingface-llama":
        llm = ChatHuggingFace(llm=HuggingFaceEndpoint(repo_id="TinyLlama/TinyLlama-1.1B-Chat-v1.0", task="text-generation", max_new_tokens=1024, do_sample=False, repetition_penalty=1.03, temperature=0), verbose=True)
    else:
        raise ValueError("Invalid provider. Choose 'google', 'groq', 'huggingface-qwen' or 'huggingface-llama'.")
        
    llm_with_tools = llm.bind_tools(tools) # Bind tools to LLM

    # Node
    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]}

    # create nodes - decision points
    builder = StateGraph(MessagesState)
    builder.add_node("retriever", retriever)
    builder.add_node("assistant", assistant)
    builder.add_node("tools", ToolNode(tools)) # equip the agents with the list of tools

    # connect nodes - control flow
    builder.add_edge(START, "retriever")
    builder.add_edge("retriever", "assistant")
    builder.add_conditional_edges("assistant", tools_condition)
    builder.add_edge("tools", "assistant")

    # Compile graph
    return builder.compile()