02_LCEL_ChatWithHistory.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain.schema import Document\n",
    "from langchain_openai import OpenAIEmbeddings\n",
    "from langchain_community.vectorstores import Chroma\n",
    "import os\n",
    "from dotenv import load_dotenv\n",
    "\n",
    "app_dir = os.path.join(os.getcwd(), \"app\")\n",
    "load_dotenv(os.path.join(app_dir, \".env\"))\n",
    "\n",
    "embedding_function = OpenAIEmbeddings()\n",
    "\n",
    "docs = [\n",
    "    Document(\n",
    "        page_content=\"the dog loves to eat pizza\", metadata={\"source\": \"animal.txt\"}\n",
    "    ),\n",
    "    Document(\n",
    "        page_content=\"the cat loves to eat lasagna\", metadata={\"source\": \"animal.txt\"}\n",
    "    ),\n",
    "]\n",
    "\n",
    "\n",
    "db = Chroma.from_documents(docs, embedding_function)\n",
    "retriever = db.as_retriever()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "retriever.invoke(\"What exactly?\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain.prompts.prompt import PromptTemplate\n",
    "\n",
    "rephrase_template = \"\"\"Given the following conversation and a follow up question, rephrase the follow up question to be a standalone question, in its original language.\n",
    "\n",
    "Chat History:\n",
    "{chat_history}\n",
    "Follow Up Input: {question}\n",
    "Standalone question:\"\"\"\n",
    "REPHRASE_TEMPLATE = PromptTemplate.from_template(rephrase_template)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain_core.messages import AIMessage, HumanMessage\n",
    "from langchain_openai import ChatOpenAI\n",
    "from langchain_core.output_parsers import StrOutputParser\n",
    "\n",
    "rephrase_chain = REPHRASE_TEMPLATE | ChatOpenAI(temperature=0) | StrOutputParser()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "rephrase_chain.invoke(\n",
    "    {\n",
    "        \"question\": \"No, really?\",\n",
    "        \"chat_history\": [\n",
    "            HumanMessage(content=\"What does the dog like to eat?\"),\n",
    "            AIMessage(content=\"Thuna!\"),\n",
    "        ],\n",
    "    }\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain_core.prompts import ChatPromptTemplate\n",
    "\n",
    "template = \"\"\"Answer the question based only on the following context:\n",
    "{context}\n",
    "\n",
    "Question: {question}\n",
    "\"\"\"\n",
    "ANSWER_PROMPT = ChatPromptTemplate.from_template(template)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain_core.runnables import RunnablePassthrough\n",
    "\n",
    "retrieval_chain = (\n",
    "    {\"context\": retriever, \"question\": RunnablePassthrough()}\n",
    "    | ANSWER_PROMPT\n",
    "    | ChatOpenAI(temperature=0)\n",
    "    | StrOutputParser()\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_chain = rephrase_chain | retrieval_chain"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_chain.invoke(\n",
    "    {\n",
    "        \"question\": \"No, really?\",\n",
    "        \"chat_history\": [\n",
    "            HumanMessage(content=\"What does the dog like to eat?\"),\n",
    "            AIMessage(content=\"Thuna!\"),\n",
    "        ],\n",
    "    }\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Chat with returning documents"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "retrieved_documents = {\"docs\": retriever, \"question\": RunnablePassthrough()}\n",
    "final_inputs = {\n",
    "    \"context\": lambda x: \"\\n\".join(doc.page_content for doc in x[\"docs\"]),\n",
    "    \"question\": lambda x: x[\"question\"],\n",
    "}\n",
    "answer = {\n",
    "    \"answer\": final_inputs | ANSWER_PROMPT | ChatOpenAI() | StrOutputParser(),\n",
    "    \"docs\": lambda x: x[\"docs\"],\n",
    "}\n",
    "\n",
    "final_chain = rephrase_chain | retrieved_documents | answer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "result = final_chain.invoke(\n",
    "    {\n",
    "        \"question\": \"No, really?\",\n",
    "        \"chat_history\": [\n",
    "            HumanMessage(content=\"What does the dog like to eat?\"),\n",
    "            AIMessage(content=\"Thuna!\"),\n",
    "        ],\n",
    "    }\n",
    ")\n",
    "print(result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "result[\"answer\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "result[\"docs\"]"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.11.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}