{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7015d967",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from dotenv import load_dotenv\n",
    "from langchain.chat_models import ChatOpenAI\n",
    "from langchain.chains import ConversationalRetrievalChain\n",
    "from langchain.memory import ConversationBufferMemory\n",
    "from langchain.vectorstores import Chroma\n",
    "from langchain.embeddings import OpenAIEmbeddings\n",
    "from langchain.document_loaders import DirectoryLoader, TextLoader\n",
    "from langchain.text_splitter import RecursiveCharacterTextSplitter\n",
    "from langchain_experimental.text_splitter import SemanticChunker\n",
    "from langchain.schema import Document\n",
    "import gradio as gr\n",
    "import glob"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "87646db6",
   "metadata": {},
   "outputs": [],
   "source": [
    "load_dotenv(override=True)\n",
    "api_key = os.getenv('OPENAI_API_KEY')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1019e3b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "file_paths = glob.glob(\"knowledge_base/**/*.md\", recursive=True)\n",
    "\n",
    "documents = []\n",
    "for path in file_paths:\n",
    "    with open(path, \"r\", encoding=\"utf-8\") as f:\n",
    "        text = f.read()\n",
    "        doc_type = os.path.basename(os.path.dirname(path)) \n",
    "\n",
    "        documents.append(\n",
    "            Document(\n",
    "                page_content=text,\n",
    "                metadata={\n",
    "                    \"doc_type\": doc_type,\n",
    "                },\n",
    "            )\n",
    "        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "54527a21",
   "metadata": {},
   "outputs": [],
   "source": [
    "embeddings = OpenAIEmbeddings(model=\"text-embedding-3-small\", openai_api_key=api_key)\n",
    "\n",
    "# text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)\n",
    "text_splitter = SemanticChunker(embeddings)\n",
    "chunks = text_splitter.split_documents(documents)\n",
    "\n",
    "print(f\"Total number of chunks: {len(chunks)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "15579dda",
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorstore = Chroma.from_documents(\n",
    "    documents=chunks,\n",
    "    embedding=embeddings,\n",
    "    persist_directory=\"chroma_db\"\n",
    ")\n",
    "vectorstore.persist()\n",
    "print(\"Chroma vector store built.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ca3b4d55",
   "metadata": {},
   "outputs": [],
   "source": [
    "llm = ChatOpenAI(model=\"gpt-4o-mini\", temperature=0, openai_api_key=api_key)\n",
    "memory = ConversationBufferMemory(memory_key='chat_history', return_messages=True)\n",
    "retriever = vectorstore.as_retriever()\n",
    "conversation_chain = ConversationalRetrievalChain.from_llm(\n",
    "    llm=llm,\n",
    "    retriever=retriever,\n",
    "    memory=memory,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "94b3a75a",
   "metadata": {},
   "outputs": [],
   "source": [
    "query = \"Tell me about Langchain.\"\n",
    "result = conversation_chain({\"question\": query})\n",
    "\n",
    "print(\"Answer:\")\n",
    "print(result[\"answer\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e814f910",
   "metadata": {},
   "outputs": [],
   "source": [
    "def rag_chat(query, history):\n",
    "    response = conversation_chain({\"question\": query})\n",
    "    answer = response[\"answer\"]\n",
    "    return answer\n",
    "\n",
    "with gr.Blocks(theme=gr.themes.Soft()) as rag_ui:\n",
    "    gr.Markdown(\"# RAG Chat Assistant\")\n",
    "    gr.Markdown(\"Ask questions about your Markdown knowledge base.\")\n",
    "    chat_box = gr.ChatInterface(\n",
    "        fn=rag_chat,\n",
    "        title=\"RAG Knowledge Base Assistant\",\n",
    "        description=\"Chat with your Markdown-based knowledge base using RAG.\"\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "eef8d2ee",
   "metadata": {},
   "outputs": [],
   "source": [
    "rag_ui.launch(debug=True, share=True)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "llm-engineering",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}