NoteWise

NoteWise is an intelligent agent designed to answer questions about specific subjects using a combination of tools including Wikipedia, Arxiv, and a custom vector store. This project leverages the power of various APIs and machine learning models to provide accurate and comprehensive information retrieval.

Features

• Wikipedia Tool: Utilizes the Wikipedia API to fetch and summarize information from Wikipedia articles.
• Arxiv Tool: Accesses scholarly articles from Arxiv to provide detailed academic information.
• Vector Store Tool: Uses a custom-built vector store to retrieve information from a pre-loaded dataset.

Installation

1. Clone the repository:

   git clone https://github.com/francoismavunila/NoteWise.git
   cd notewise

2. Install the required dependencies:

   pip install -r requirements.txt

Usage

1. Load the necessary tools:

   from langchain_community.tools import WikipediaQueryRun
   from langchain_community.utilities import WikipediaAPIWrapper
   from langchain_community.document_loaders import WebBaseLoader
   from langchain_community.vectorstores import FAISS
   from langchain_openai import OpenAIEmbeddings, ChatOpenAI
   from langchain_text_splitters import RecursiveCharacterTextSplitter
   from langchain.tools.retriever import create_retriever_tool
   from langchain_community.utilities import ArxivAPIWrapper
   from langchain_community.tools import ArxivQueryRun

2. Initialize the tools:

   # Wikipedia Tool
   wikipedia = WikipediaQueryRun(api_wrapper=WikipediaAPIWrapper(top_k_results=1, doc_content_chars_max=200))
   
   # Arxiv Tool
   arxiv_wrapper = ArxivAPIWrapper(top_k_results=1, doc_content_chars_max=200)
   arxiv = ArxivQueryRun(api_wrapper=arxiv_wrapper)
   
   # Vector Store Tool
   loader = WebBaseLoader("link_to_any_site_here")
   docs = loader.load()
   documents = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200).split_documents(docs)
   vectordb = FAISS.from_documents(documents, OpenAIEmbeddings())
   retriever = vectordb.as_retriever()
   retriever_tool = create_retriever_tool(retriever, "too_name", "search for information about ......, for any information about ...... use this tool")
   
   # Combine tools
   tools = [wikipedia, arxiv, retriever_tool]

3. Use the tools to answer questions:

   # Example usage
   question = "What is the latest research on AI?"
   response = agent.answer(question, tools=tools)
   print(response)

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

License

This project is licensed under the MIT License.