"""Tool definitions for the RAG agent.

The tools are simple wrappers around the vector store functions defined in
`vector_store.py`. They are decorated with `@tool` from LangChain so that the
agent can call them.
"""

from typing import Any

from langchain.tools import tool

from .vector_store import add_document, search

@tool("search_knowledge_base")
def search_knowledge_base(query: str, max_results: int = 5) -> Any:
    """Semantic search in the knowledge base.

    Parameters
    ----------
    query: str
        The user query.
    max_results: int, optional
        Number of top results to return. Defaults to 5.
    """
    results = search(query, max_results)
    # Convert results to a readable string
    formatted = "\n".join(
        f"{i+1}. [{res['metadata'].get('title', 'Unknown')}] {res['content'][:200]}"
        for i, res in enumerate(results)
    )
    return formatted if formatted else "No relevant documents found."

@tool("add_to_knowledge_base")
def add_to_knowledge_base(content: str, title: str) -> Any:
    """Add a new document to the knowledge base.

    Parameters
    ----------
    content: str
        The full text of the document.
    title: str
        A short title that will be stored as metadata.
    """
    add_document(content, title)
    return f"Document '{title}' added to the knowledge base."

__all__ = ["search_knowledge_base", "add_to_knowledge_base"]