docs/docs_ru/cookbook/gigachat_qa.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3ea857b1",
   "metadata": {},
   "source": [
    "# Retrieval Augmented Generation(RAG) с использованием GigaChat на примере задачи \"разговор с книгой\"\n",
    "\n",
    "Подход RAG позволяет большим языковым моделям (LLM) отвечать на вопросы по документам, которы не помещаются в промпт.\n",
    "Ниже приведен пример того, как можно научить модель отвечать на вопросы, используя текст из книги \"Мастер и Маргарита\".\n",
    "\n",
    "Подробнее про RAG вы можете прочитать в [документации LangChain](https://python.langchain.com/docs/use_cases/question_answering/) и в курсе по промпт-инженирингу от Сбера (ссылка будет позже).\n",
    "\n",
    "В качестве примера мы рассмотрим текст романа Булгакова \"Мастер и Маргарита\" (главы 1 и 2).\n",
    "\n",
    "Вопрос будет - `Какой плащ был у Понтия Пилата?`. Ответ содержится во второй главе книги:\n",
    "`В белом плаще с кровавым подбоем, шаркающей кавалерийской походкой, ранним утром четырнадцатого числа весеннего месяца нисана в крытую колоннаду между двумя крыльями дворца ирода великого вышел прокуратор Иудеи Понтий Пилат.`"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9ea31150",
   "metadata": {},
   "source": [
    "## Установка\n",
    "\n",
    "Для работы нам понадобится векторая база данных. Мы будем использовать Chroma."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "a7dc1ec5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[33mWARNING: There was an error checking the latest version of pip.\u001b[0m\u001b[33m\n",
      "\u001b[0mNote: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "%pip install langchain-chroma --quiet"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5e7543fa",
   "metadata": {},
   "source": [
    "## Инициализация модели\n",
    "Теперь инициализируем модель GigaChat."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "921afeaa",
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain_gigachat.chat_models import GigaChat\n",
    "\n",
    "llm = GigaChat(credentials=\"ключ_авторизации\", verify_ssl_certs=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "84443c6f",
   "metadata": {},
   "source": [
    "Для проверки спросим у модели вопрос про цвет плаща без какого-либо контекста. Возможно, она и так будет давать ожидаемый ответ..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "af0398d5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Понтий Пилат, римский прокуратор Иудеи, носил плащ, который назывался \"тога\". Тога была традиционной одеждой римлян и представляла собой длинное прямоугольное полотно ткани, которое оборачивалось вокр'"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from langchain.schema import HumanMessage\n",
    "\n",
    "question = \"Какой плащ был у Понтия Пилата?\"\n",
    "llm([HumanMessage(content=question)]).content[0:200]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b2e4e31",
   "metadata": {},
   "source": [
    "Видим, что модель не отвечает так, как нам хотелось бы, поэтому применим RAG-подход.\n",
    "\n",
    "## Подготовка документа\n",
    "\n",
    "Для работы с документом нам нужно разделить его на части. Для этого используем `TextLoader` для загрузки книги и `RecursiveCharacterTextSplitter`, чтобы разделить текст на приблизительно равные куски в ~1000 символов с перекрытием в ~200 символов. Этот тип сплиттера сам выбирает каким способом следует оптимально разделять документ (по абзацам, по предложениям и т.д.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "f8cf5765",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total documents: 91\n"
     ]
    }
   ],
   "source": [
    "from langchain.document_loaders import TextLoader\n",
    "from langchain.text_splitter import (\n",
    "    RecursiveCharacterTextSplitter,\n",
    ")\n",
    "\n",
    "loader = TextLoader(\"../мастер_и_маргарита.txt\")\n",
    "documents = loader.load()\n",
    "text_splitter = RecursiveCharacterTextSplitter(\n",
    "    chunk_size=1000,\n",
    "    chunk_overlap=200,\n",
    ")\n",
    "documents = text_splitter.split_documents(documents)\n",
    "print(f\"Total documents: {len(documents)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "131d5059",
   "metadata": {},
   "source": [
    "После нарезки мы получили 91 документ частями книги.\n",
    "\n",
    "## Создание базы данных эмбеддингов\n",
    "\n",
    "Эмбеддинг это векторное представление текста, которое может быть использовано для определения смысловой близости текстов. Векторная база данных хранит тексты и соответствующие им эмбеддинги, а также умеет выполнять поиск по ним. Для работы с базой данных мы создаем объект GigaChatEmbeddings и передаем его в базу данных Chroma.\n",
    "\n",
    "> Обратите внимание, что сервис для вычисления эмбеддингов может тарифицироваться отдельно от стоимости модели GigaChat."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "fdce8923",
   "metadata": {},
   "outputs": [],
   "source": [
    "from chromadb.config import Settings\n",
    "from langchain_gigachat.embeddings.gigachat import GigaChatEmbeddings\n",
    "from langchain_chroma import Chroma\n",
    "\n",
    "embeddings = GigaChatEmbeddings(\n",
    "    credentials=\"ключ_авторизации\", verify_ssl_certs=False\n",
    ")\n",
    "\n",
    "db = Chroma.from_documents(\n",
    "    documents,\n",
    "    embeddings,\n",
    "    client_settings=Settings(anonymized_telemetry=False),\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "29137915",
   "metadata": {},
   "source": [
    "## Поиск по базе данных\n",
    "\n",
    "Теперь можно обратиться к базе данных и попросить найти документы, которые с наибольшей вероятностью содержат ответ на наш вопрос.\n",
    "\n",
    "По-умолчанию база данных возвращает 4 наиболее релевантных документа. Этот параметр можно изменить в зависимости от решаемой задачи и типа документов.\n",
    "\n",
    "Видно, что первый же документ содержит внутри себя часть книги с ответом на наш вопрос."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "b0c55e98",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "docs = db.similarity_search(question, k=4)\n",
    "len(docs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "32b43339",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "... акцент почему-то пропал: – Все просто: в белом плаще...\\n\\n\\n\\nГлава 2\\n\\nПонтий Пилат\\n\\nВ белом плаще с кровавым подбоем, шаркающей кавалерийской походкой, ранним утром четырнадц ...\n"
     ]
    }
   ],
   "source": [
    "print(f\"... {str(docs[0])[620:800]} ...\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "96a063c7",
   "metadata": {},
   "source": [
    "## QnA цепочка\n",
    "\n",
    "Теперь мы создадим цепочку QnA, которая специально предназначена для ответов на вопросы по документам. В качестве аргументов есть передается языковая модель и ретривер (база данных)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "24671e9a",
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain.chains import RetrievalQA\n",
    "\n",
    "qa_chain = RetrievalQA.from_chain_type(llm, retriever=db.as_retriever())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "592151a2",
   "metadata": {},
   "source": [
    "Наконец можно задать вопрос нашей цепочке и получить правильный ответ!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "f3b2f453",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/19563044/Documents/Giga/gigachain/venv/lib/python3.11/site-packages/langchain_core/_api/deprecation.py:117: LangChainDeprecationWarning: The function `__call__` was deprecated in LangChain 0.1.0 and will be removed in 0.2.0. Use invoke instead.\n",
      "  warn_deprecated(\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'query': 'Какой плащ был у Понтия Пилата?',\n",
       " 'result': 'У Понтия Пилата был белый плащ с кровавым подбоем.'}"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "qa_chain({\"query\": question})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ffa10ec4",
   "metadata": {},
   "source": [
    "Несколько дополнительных вопросов для проверки работоспособности:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "8f619e09",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'query': 'Какая трость была у Воланда?',\n",
       " 'result': 'У Воланда была трость с черным набалдашником в виде головы пуделя.'}"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "qa_chain({\"query\": \"Какая трость была у Воланда?\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "a4a54b82",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'query': 'что не смогли купить герои романа на Патриарших?',\n",
       " 'result': 'Герои романа \"Мастер и Маргарита\" не смогли купить нарзан в будочке на Патриарших.'}"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "qa_chain({\"query\": \"что не смогли купить герои романа на Патриарших?\"})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e4cf02ad",
   "metadata": {},
   "source": [
    "# Бонус - работа с научной литературой\n",
    "Рассмотрим автономный пример работы с научной литературой. В качестве образца возьмем книгу \"Дискретная математика\" В. Е. Алексеева.\n",
    "\n",
    "Из примеров видно, что применение метода RAG к научным книгам оправдано и позволяет модели отвечать на вопросы, учитывая знания, написанные в книге."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "438e1017",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m24.0\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m24.2\u001b[0m\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpip install --upgrade pip\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "!pip install pypdf langchain-chroma --quiet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "d8581014",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "139\n"
     ]
    }
   ],
   "source": [
    "from langchain_community.document_loaders import PyPDFLoader\n",
    "\n",
    "# Если ссылка на книгу перестанет работать, можно найти и использовать локальную версию.\n",
    "# Для работы этого демо используется ссылка на библиотеку сайта ННГУ (http://www.unn.ru/)\n",
    "loader = PyPDFLoader(\"http://www.unn.ru/books/met_files/Alekseev.pdf\")\n",
    "\n",
    "docs = loader.load()\n",
    "print(len(docs))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "269b2921",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " \n",
      " \n",
      "МИНИСТЕРСТВО ОБРАЗОВАНИЯ И НАУКИ  \n",
      "РОССИЙСКОЙ ФЕДЕРАЦИИ  \n",
      " \n",
      "Нижегородский государственный универ\n",
      "{'source': 'http://www.unn.ru/books/met_files/Alekseev.pdf', 'page': 0}\n"
     ]
    }
   ],
   "source": [
    "print(docs[0].page_content[0:100])\n",
    "print(docs[0].metadata)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5a912c35",
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain_gigachat.chat_models import GigaChat\n",
    "from langchain.schema import HumanMessage\n",
    "\n",
    "giga = GigaChat(\n",
    "    credentials=\"ключ_автоирзации\",\n",
    "    model=\"GigaChat-Pro\",\n",
    "    verify_ssl_certs=False,\n",
    "    timeout=1200,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "7b901e8e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain_chroma import Chroma\n",
    "from langchain_gigachat.embeddings.gigachat import GigaChatEmbeddings\n",
    "from langchain_text_splitters import RecursiveCharacterTextSplitter\n",
    "\n",
    "embeddings = GigaChatEmbeddings(\n",
    "    credentials=\"ключ_авторизации\", verify_ssl_certs=False\n",
    ")\n",
    "\n",
    "\n",
    "text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)\n",
    "splits = text_splitter.split_documents(docs)\n",
    "vectorstore = Chroma.from_documents(documents=splits, embedding=embeddings)\n",
    "\n",
    "retriever = vectorstore.as_retriever()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "1b91b729",
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain.chains import create_retrieval_chain\n",
    "from langchain.chains.combine_documents import create_stuff_documents_chain\n",
    "from langchain_core.prompts import ChatPromptTemplate\n",
    "\n",
    "system_prompt = (\n",
    "    \"Ты должен ответить на вопрос пользователя с использованием данных из книги.\\n\"\n",
    "    \"Отвечаай коротко, не более 2-3 предложений.\\n\"\n",
    "    \"Вот части книги контекст для ответа:\"\n",
    "    \"\\n\\n\"\n",
    "    \"{context}\"\n",
    ")\n",
    "\n",
    "prompt = ChatPromptTemplate.from_messages(\n",
    "    [\n",
    "        (\"system\", system_prompt),\n",
    "        (\"human\", \"{input}\"),\n",
    "    ]\n",
    ")\n",
    "\n",
    "\n",
    "question_answer_chain = create_stuff_documents_chain(giga, prompt)\n",
    "rag_chain = create_retrieval_chain(retriever, question_answer_chain)"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "b7c94605",
   "metadata": {},
   "source": [
    "## Проверка работы модели\n",
    "Для сравнения мы будем задвать вопрос сначала модели напрямую (pure), затем через rag-цепочку\n",
    "\n",
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "aae231c4",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/19563044/Documents/Giga/gigachain/.venv/lib/python3.11/site-packages/langchain_core/_api/deprecation.py:139: LangChainDeprecationWarning: The method `BaseChatModel.__call__` was deprecated in langchain-core 0.1.7 and will be removed in 0.3.0. Use invoke instead.\n",
      "  warn_deprecated(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pure: Теорема об эйлеровом цикле гласит, что в любом графе без мостов и висячих ребер сумма степеней всех вершин равна удвоенному числу рёбер.\n",
      "RAG: Теорема об эйлеровом цикле гласит, что в связном графе эйлеров цикл существует тогда и только тогда, когда степени всех вершин графа четны.\n"
     ]
    }
   ],
   "source": [
    "\"\"\" Оригинал: Теорема 5.7 (об эйлеровом цикле). Эйлеров цикл в связном графе\n",
    "существует тогда и только тогда, когда степени всех вершин этого графа\n",
    "четны. \"\"\"\n",
    "q = \"О чем говорит теорема об эйлеровом цикле?\"\n",
    "pure = giga([HumanMessage(content=q)]).content\n",
    "print(f\"Pure: {pure}\")\n",
    "\n",
    "rag = rag_chain.invoke({\"input\": q})[\"answer\"]\n",
    "print(f\"RAG: {rag}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "8a69ef6e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pure: Граф не будет двудольным, если он содержит цикл нечётной длины или хотя бы одну пару несмежных рёбер.\n",
      "\n",
      "RAG: Граф не будет двудольным, если в нем содержатся нечетные циклы.\n"
     ]
    }
   ],
   "source": [
    "\"\"\" Оригинал:  Оказывается, присутствие в графе нечетных циклов (т.е. циклов\n",
    "нечетной длины) – единственная причина, по которой граф не является\n",
    "двудольным. \"\"\"\n",
    "q = \"Короткий ответ: По каким причинам граф не будет двудольным? \"\n",
    "pure = giga([HumanMessage(content=q)]).content\n",
    "print(f\"Pure: {pure}\\n\")\n",
    "\n",
    "rag = rag_chain.invoke({\"input\": q})[\"answer\"]\n",
    "print(f\"RAG: {rag}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "33658e37",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pure: Слово — это лексическая единица языка, которая имеет определённое значение. Набор — это совокупность предметов, явлений, понятий, объединённых по какому-либо признаку.\n",
      "\n",
      "Пример:\n",
      "\n",
      "Слово «стол» — это лексическая единица русского языка, обозначающая предмет мебели для работы или приёма пищи.\n",
      "\n",
      "Набор столовых приборов — это совокупность различных предметов (ложек, вилок, ножей), которые используются во время еды.\n",
      "\n",
      "RAG: Разница между словом и набором заключается в том, что слово - это упорядоченная последовательность букв, в которой каждая буква может повторяться, тогда как набор - это просто последовательность элементов без учета их повторений или порядка.\n",
      "\n",
      "Пример: \"мама\" - это слово, потому что оно состоит из букв, которые могут повторяться (\"м\" и \"а\"), и порядок букв имеет значение. \"ма\", \"ам\", \"ама\" - это другие слова, но \"мама\" - единственное слово из этих трех.\n",
      "\n",
      "С другой стороны, набор \"мама\" будет содержать только одну запись, потому что набор не учитывает повторения элементов или их порядок.\n"
     ]
    }
   ],
   "source": [
    "\"\"\" Оригинал: Термин набор означает конечную последовательность элементов.\n",
    "Понятие набора отличается от понятия множества тем, что\n",
    "* в наборе могут быть одинаковые элементы,\n",
    "* порядок расположения элементов в наборе важен (наборы, состоящие\n",
    "из одинаковых элементов, расположенных в разном порядке,\n",
    "считаются различными). \"\"\"\n",
    "q = \"Разница между словом и набором. Приведи пример.\"\n",
    "pure = giga([HumanMessage(content=q)]).content\n",
    "print(f\"Pure: {pure}\\n\")\n",
    "\n",
    "rag = rag_chain.invoke({\"input\": q})[\"answer\"]\n",
    "print(f\"RAG: {rag}\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}