tools.py

from langchain_core.tools import tool
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langchain_community.vectorstores import FAISS
from langchain_community.document_loaders import PyPDFLoader
from langchain_openai import OpenAIEmbeddings
from langchain_community.tools import WikipediaQueryRun, ArxivQueryRun
from langchain_community.utilities import WikipediaAPIWrapper, ArxivAPIWrapper
from langchain_community.tools.tavily_search import TavilySearchResults
from dotenv import load_dotenv
import os
import requests

load_dotenv()

# ==============================
# CONFIG
# ==============================
VECTORSTORE_DIR = "data/vectorstore"
os.makedirs(VECTORSTORE_DIR, exist_ok=True)

# ==============================
# GLOBAL RETRIEVER
# ==============================
retriever = None


def load_retriever():
    global retriever
    try:
        embeddings = OpenAIEmbeddings(model="text-embedding-3-small")
        if os.path.exists(os.path.join(VECTORSTORE_DIR, "index.faiss")):
            vectorstore = FAISS.load_local(
                VECTORSTORE_DIR,
                embeddings,
                allow_dangerous_deserialization=True
            )
            retriever = vectorstore.as_retriever(search_kwargs={"k": 3})
            print("✅ Vectorstore loaded from disk")
    except Exception as e:
        print("❌ Failed to load vectorstore:", e)


def build_vectorstore(path: str):
    loader = PyPDFLoader(path)
    docs = loader.load()

    splitter = RecursiveCharacterTextSplitter(
        chunk_size=500,
        chunk_overlap=100
    )

    split_docs = splitter.split_documents(docs)

    embeddings = OpenAIEmbeddings(model="text-embedding-3-small")
    vectorstore = FAISS.from_documents(split_docs, embeddings)

    vectorstore.save_local(VECTORSTORE_DIR)
    return vectorstore


def update_retriever(pdf_path: str):
    global retriever
    vectorstore = build_vectorstore(pdf_path)
    retriever = vectorstore.as_retriever(search_kwargs={"k": 3})


# ==============================
# RAG TOOL (FIXED)
# ==============================
def create_rag_tool():

    @tool
    def rag_search(query: str) -> str:
        """
        Retrieve relevant information from uploaded documents.
        """
        global retriever

        if retriever is None:
            load_retriever()

        if retriever is None:
            return "No document has been uploaded yet."

        docs = retriever.invoke(query)

        if not docs:
            return "No relevant information found in the uploaded document."

        return "\n\n".join(d.page_content for d in docs)

    return rag_search


# ---------------- OTHER TOOLS ---------------- #

@tool
def arxiv_search(query: str) -> dict:
    try:
        arxiv = ArxivQueryRun(api_wrapper=ArxivAPIWrapper())
        return {"results": arxiv.run(query)}
    except Exception as e:
        return {"error": str(e)}


@tool
def wikipedia_search(query: str) -> dict:
    try:
        wiki = WikipediaQueryRun(api_wrapper=WikipediaAPIWrapper())
        return {"results": wiki.run(query)}
    except Exception as e:
        return {"error": str(e)}


@tool
def tavily_search(query: str) -> dict:
    try:
        search = TavilySearchResults(max_results=5)
        return {"results": search.run(query)}
    except Exception as e:
        return {"error": str(e)}