Task-Question-Answering.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Information Retrieval Demo\n",
    "\n",
    "Date: 03 May 2023\n",
    "\n",
    "In this demo, we will be demo-ing how to perform Information Retrieval using Langchain.\n",
    "\n",
    "### What is LangChain\n",
    "\n",
    "LangChain is a framework for developing applications powered by language models. \n",
    "\n",
    "_LangChain Principles:_\n",
    "- Be data-aware: connect a language model to other sources of data\n",
    "- Be agentic: allow a language model to interact with its environment\n",
    "\n",
    "Source: [LangChain documentation](https://python.langchain.com/en/latest/index.html)  \n",
    "Version: 0.0.154 as of 02 May 2021\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Information Retrieval & Retrieval Augmented Generation\n",
    "\n",
    "<img src=\"https://substackcdn.com/image/fetch/w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4063347e-8920-40c6-86b3-c520084b303c_1272x998.jpeg\" alt= “” width=\"600\" height=\"500\">\n",
    "\n",
    "\n",
    "<!-- ![](https://substackcdn.com/image/fetch/w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4063347e-8920-40c6-86b3-c520084b303c_1272x998.jpeg) -->\n",
    "\n",
    "Source: [Finetuning Large Language Models by Sebastian Raschka](https://magazine.sebastianraschka.com/p/finetuning-large-language-models)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Requirements (for this demo)\n",
    "- Offline or Local model\n",
    "- Initial model is somewhat basic\n",
    "- HuggingFace Transformers\n",
    "- Read documents and extract similar documents based on query\n",
    "- Read documents and generate answer based on query"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LangChain: Models\n",
    "\n",
    "Generic inferface for models e.g. LLMs, Chat Models, Text Embedding Models. Read more [here](https://python.langchain.com/en/latest/modules/models.html)\n",
    "\n",
    "```python\n",
    "from langchain.llms import OpenAI\n",
    "llm = OpenAI(model_name=\"text-ada-001\")\n",
    "\n",
    "from langchain import HuggingFaceHub\n",
    "llm = HuggHuggingFaceHub(repo_id=\"google/flan-t5-xl\")\n",
    "\n",
    "from langchain.llms import Cohere\n",
    "llm = Cohere()\n",
    "```"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Local Models: HuggingFace Pipeline\n",
    "\n",
    "Note: This is experimental code, always use functional or OO abstraction for your implementation.\n",
    "\n",
    "```python3\n",
    "from langchain.llms import HuggingFacePipeline\n",
    "from transformers import AutoTokenizer, pipeline, AutoModelForSeq2SeqLM\n",
    "\n",
    "model_id = 'flan-t5-large' # any local model\n",
    "tokenizer = AutoTokenizer.from_pretrained(model_id)\n",
    "model = AutoModelForSeq2SeqLM.from_pretrained(model_id, load_in_8bit=True) # 8bit in A10, A100 etc.\n",
    "\n",
    "pipe = pipeline(\n",
    "    task=\"text2text-generation\", \n",
    "    model=model, \n",
    "    tokenizer=tokenizer, \n",
    "    max_length=100\n",
    ")\n",
    "\n",
    "local_llm = HuggingFacePipeline(pipeline=pipe)\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e0c42f0f2cbe473ba573450821ef2306",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Loading checkpoint shards:   0%|          | 0/2 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from langchain.llms import HuggingFacePipeline\n",
    "from langchain.retrievers.document_compressors import LLMChainExtractor\n",
    "from transformers import pipeline\n",
    "\n",
    "model_name = \"google/flan-t5-xl\"\n",
    "# model_name = \"hkunlp/instructor-large\"\n",
    "\n",
    "\n",
    "# flan-t5-large or flan-t5-xl are also good initial models\n",
    "pipe = pipeline(\"text2text-generation\", \n",
    "                model=model_name, \n",
    "                max_length=300)\n",
    "local_llm = HuggingFacePipeline(pipeline=pipe)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "load INSTRUCTOR_Transformer\n",
      "max_seq_length  512\n"
     ]
    }
   ],
   "source": [
    "from langchain.embeddings import HuggingFaceInstructEmbeddings\n",
    "\n",
    "model_name = \"hkunlp/instructor-large\"\n",
    "\n",
    "model_kwargs = {'device': 'cuda'}\n",
    "instructor_embeds = HuggingFaceInstructEmbeddings(\n",
    "    model_name=model_name, model_kwargs=model_kwargs\n",
    ")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#TODO: Add slide on Flan T5\n",
    "\n",
    "<img src=\"https://s3.amazonaws.com/moonup/production/uploads/1666363435475-62441d1d9fdefb55a0b7d12c.png\" alt= “” width=\"700\" height=\"500\">\n",
    "\n",
    "- Paper <https://arxiv.org/abs/2210.11416>\n",
    "- HuggingFace Hub <https://huggingface.co/google/flan-t5-xl>"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LangChain: Indexes\n",
    "\n",
    "Utility functions to combine own private data. Read more [here](https://python.langchain.com/en/latest/modules/indexes.html).\n",
    "\n",
    "Now we are talking.\n",
    "\n",
    "### [Document Loaders](https://python.langchain.com/en/latest/modules/indexes/document_loaders.html)\n",
    "\n",
    "Email, Images, PDF, s3 Directory and Files, Word Documents, Powerpoints etc.\n",
    "\n",
    "### [Text Splitters](https://python.langchain.com/en/latest/modules/indexes/text_splitters.html)\n",
    "\n",
    "Character Text Splitter, Recursive Character Text Splitter here.\n",
    "\n",
    "### [VectorStores](https://python.langchain.com/en/latest/modules/indexes/vectorstores.html)\n",
    "\n",
    "ElasticSearch, FAISS, Qdrant, Redis, Weaviate\n",
    "\n",
    "### [Retrievers](https://python.langchain.com/en/latest/modules/indexes/retrievers.html)\n",
    "\n",
    "Contextual Compression Retriever (!!!), SVM Retriever, TF-IDF Retriever, Time Weighted VectorStore Retriever etc.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Created a chunk of size 1385, which is longer than the specified 500\n",
      "Created a chunk of size 671, which is longer than the specified 500\n",
      "Created a chunk of size 1438, which is longer than the specified 500\n",
      "Created a chunk of size 506, which is longer than the specified 500\n",
      "Created a chunk of size 759, which is longer than the specified 500\n",
      "Created a chunk of size 615, which is longer than the specified 500\n",
      "Created a chunk of size 715, which is longer than the specified 500\n",
      "Created a chunk of size 754, which is longer than the specified 500\n",
      "Created a chunk of size 752, which is longer than the specified 500\n",
      "Created a chunk of size 1078, which is longer than the specified 500\n",
      "Created a chunk of size 802, which is longer than the specified 500\n",
      "Created a chunk of size 1510, which is longer than the specified 500\n",
      "Created a chunk of size 1607, which is longer than the specified 500\n",
      "Created a chunk of size 503, which is longer than the specified 500\n",
      "Created a chunk of size 973, which is longer than the specified 500\n",
      "Created a chunk of size 632, which is longer than the specified 500\n",
      "Created a chunk of size 601, which is longer than the specified 500\n",
      "Created a chunk of size 504, which is longer than the specified 500\n",
      "Created a chunk of size 808, which is longer than the specified 500\n",
      "Created a chunk of size 848, which is longer than the specified 500\n",
      "Created a chunk of size 776, which is longer than the specified 500\n",
      "Created a chunk of size 580, which is longer than the specified 500\n",
      "Created a chunk of size 522, which is longer than the specified 500\n",
      "Created a chunk of size 525, which is longer than the specified 500\n",
      "Created a chunk of size 584, which is longer than the specified 500\n",
      "Created a chunk of size 747, which is longer than the specified 500\n",
      "Created a chunk of size 967, which is longer than the specified 500\n",
      "Created a chunk of size 507, which is longer than the specified 500\n",
      "Created a chunk of size 510, which is longer than the specified 500\n",
      "Created a chunk of size 789, which is longer than the specified 500\n",
      "Created a chunk of size 609, which is longer than the specified 500\n",
      "Created a chunk of size 673, which is longer than the specified 500\n",
      "Created a chunk of size 515, which is longer than the specified 500\n",
      "Created a chunk of size 574, which is longer than the specified 500\n",
      "Created a chunk of size 918, which is longer than the specified 500\n",
      "Created a chunk of size 757, which is longer than the specified 500\n",
      "Created a chunk of size 520, which is longer than the specified 500\n",
      "Created a chunk of size 623, which is longer than the specified 500\n",
      "Created a chunk of size 963, which is longer than the specified 500\n",
      "Created a chunk of size 617, which is longer than the specified 500\n",
      "Created a chunk of size 557, which is longer than the specified 500\n",
      "Created a chunk of size 935, which is longer than the specified 500\n",
      "Created a chunk of size 610, which is longer than the specified 500\n",
      "Created a chunk of size 624, which is longer than the specified 500\n",
      "Created a chunk of size 834, which is longer than the specified 500\n",
      "Created a chunk of size 540, which is longer than the specified 500\n",
      "Created a chunk of size 591, which is longer than the specified 500\n",
      "Created a chunk of size 1024, which is longer than the specified 500\n",
      "Created a chunk of size 768, which is longer than the specified 500\n",
      "Created a chunk of size 530, which is longer than the specified 500\n",
      "Created a chunk of size 520, which is longer than the specified 500\n",
      "Created a chunk of size 652, which is longer than the specified 500\n",
      "Created a chunk of size 869, which is longer than the specified 500\n",
      "Created a chunk of size 535, which is longer than the specified 500\n",
      "Created a chunk of size 978, which is longer than the specified 500\n",
      "Created a chunk of size 592, which is longer than the specified 500\n",
      "Created a chunk of size 638, which is longer than the specified 500\n",
      "Created a chunk of size 584, which is longer than the specified 500\n",
      "Created a chunk of size 523, which is longer than the specified 500\n",
      "Created a chunk of size 517, which is longer than the specified 500\n",
      "Created a chunk of size 838, which is longer than the specified 500\n",
      "Created a chunk of size 703, which is longer than the specified 500\n",
      "Created a chunk of size 1523, which is longer than the specified 500\n",
      "Created a chunk of size 1071, which is longer than the specified 500\n",
      "Created a chunk of size 588, which is longer than the specified 500\n",
      "Created a chunk of size 745, which is longer than the specified 500\n",
      "Created a chunk of size 589, which is longer than the specified 500\n",
      "Created a chunk of size 607, which is longer than the specified 500\n",
      "Created a chunk of size 537, which is longer than the specified 500\n",
      "Created a chunk of size 700, which is longer than the specified 500\n",
      "Created a chunk of size 512, which is longer than the specified 500\n",
      "Created a chunk of size 574, which is longer than the specified 500\n",
      "Created a chunk of size 821, which is longer than the specified 500\n",
      "Created a chunk of size 695, which is longer than the specified 500\n",
      "Created a chunk of size 828, which is longer than the specified 500\n",
      "Created a chunk of size 562, which is longer than the specified 500\n",
      "Created a chunk of size 508, which is longer than the specified 500\n",
      "Created a chunk of size 978, which is longer than the specified 500\n",
      "Created a chunk of size 835, which is longer than the specified 500\n",
      "Created a chunk of size 995, which is longer than the specified 500\n",
      "Created a chunk of size 675, which is longer than the specified 500\n",
      "Created a chunk of size 777, which is longer than the specified 500\n",
      "Created a chunk of size 565, which is longer than the specified 500\n",
      "Created a chunk of size 601, which is longer than the specified 500\n",
      "Created a chunk of size 518, which is longer than the specified 500\n",
      "Created a chunk of size 758, which is longer than the specified 500\n",
      "Created a chunk of size 717, which is longer than the specified 500\n",
      "Created a chunk of size 823, which is longer than the specified 500\n",
      "Created a chunk of size 541, which is longer than the specified 500\n",
      "Created a chunk of size 624, which is longer than the specified 500\n",
      "Created a chunk of size 789, which is longer than the specified 500\n",
      "Created a chunk of size 560, which is longer than the specified 500\n",
      "Created a chunk of size 511, which is longer than the specified 500\n",
      "Created a chunk of size 1060, which is longer than the specified 500\n",
      "Created a chunk of size 1002, which is longer than the specified 500\n",
      "Created a chunk of size 1222, which is longer than the specified 500\n",
      "Created a chunk of size 560, which is longer than the specified 500\n",
      "Created a chunk of size 857, which is longer than the specified 500\n",
      "Created a chunk of size 841, which is longer than the specified 500\n",
      "Created a chunk of size 1173, which is longer than the specified 500\n",
      "Created a chunk of size 935, which is longer than the specified 500\n",
      "Created a chunk of size 1417, which is longer than the specified 500\n",
      "Created a chunk of size 666, which is longer than the specified 500\n"
     ]
    }
   ],
   "source": [
    "from langchain.text_splitter import CharacterTextSplitter, RecursiveCharacterTextSplitter\n",
    "from langchain.embeddings import OpenAIEmbeddings, HuggingFaceEmbeddings\n",
    "from langchain.document_loaders import PDFMinerLoader, TextLoader\n",
    "from langchain.vectorstores import FAISS\n",
    "\n",
    "def pretty_print_docs(docs):\n",
    "    \"\"\" Helper function for printing docs.\"\"\"\n",
    "    print(f\"\\n{'-' * 100}\\n\".join([f\"Document {i+1}:\\n\\n\" + d.page_content for i, d in enumerate(docs)]))\n",
    "\n",
    "documents = PDFMinerLoader(file_path=\"./arxiv_papers/2203.02155.pdf\").load()\n",
    "text_splitter = CharacterTextSplitter(chunk_size=500, chunk_overlap=0)\n",
    "texts = text_splitter.split_documents(documents=documents,)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from langchain.document_loaders import UnstructuredPDFLoader\n",
    "\n",
    "# Load all GPT based papers\n",
    "pdf_folder_path = f'./arxiv_papers/gpt4-papers/'\n",
    "loaders = [PDFMinerLoader(os.path.join(pdf_folder_path, fn)) for fn in os.listdir(pdf_folder_path)]"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LangChain: Indexes\n",
    "\n",
    "```python\n",
    "query = <YOUR_QUERY_HERE>\n",
    "\n",
    "from langchain.vectorstores import FAISS\n",
    "db_faiss = FAISS.from_documents(documents=docs, \n",
    "                                embedding=embeddings)\n",
    "docs = db_faiss.similarity_search(query)\n",
    "\n",
    "from langchain.vectorstores import ElasticVectorSearch\n",
    "db_elastic = ElasticVectorSearch.from_documents(documents=docs, \n",
    "                                                embedding=embeddings, \n",
    "                                                elasticsearch_url=\"http://localhost:9200\")\n",
    "docs = db_elastic.similarity_search(query)\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Created a chunk of size 1385, which is longer than the specified 300\n",
      "Created a chunk of size 491, which is longer than the specified 300\n",
      "Created a chunk of size 486, which is longer than the specified 300\n",
      "Created a chunk of size 480, which is longer than the specified 300\n",
      "Created a chunk of size 671, which is longer than the specified 300\n",
      "Created a chunk of size 1438, which is longer than the specified 300\n",
      "Created a chunk of size 506, which is longer than the specified 300\n",
      "Created a chunk of size 759, which is longer than the specified 300\n",
      "Created a chunk of size 615, which is longer than the specified 300\n",
      "Created a chunk of size 449, which is longer than the specified 300\n",
      "Created a chunk of size 471, which is longer than the specified 300\n",
      "Created a chunk of size 495, which is longer than the specified 300\n",
      "Created a chunk of size 715, which is longer than the specified 300\n",
      "Created a chunk of size 754, which is longer than the specified 300\n",
      "Created a chunk of size 752, which is longer than the specified 300\n",
      "Created a chunk of size 1078, which is longer than the specified 300\n",
      "Created a chunk of size 442, which is longer than the specified 300\n",
      "Created a chunk of size 802, which is longer than the specified 300\n",
      "Created a chunk of size 1510, which is longer than the specified 300\n",
      "Created a chunk of size 1607, which is longer than the specified 300\n",
      "Created a chunk of size 503, which is longer than the specified 300\n",
      "Created a chunk of size 973, which is longer than the specified 300\n",
      "Created a chunk of size 330, which is longer than the specified 300\n",
      "Created a chunk of size 632, which is longer than the specified 300\n",
      "Created a chunk of size 601, which is longer than the specified 300\n",
      "Created a chunk of size 504, which is longer than the specified 300\n",
      "Created a chunk of size 808, which is longer than the specified 300\n",
      "Created a chunk of size 848, which is longer than the specified 300\n",
      "Created a chunk of size 329, which is longer than the specified 300\n",
      "Created a chunk of size 319, which is longer than the specified 300\n",
      "Created a chunk of size 498, which is longer than the specified 300\n",
      "Created a chunk of size 395, which is longer than the specified 300\n",
      "Created a chunk of size 776, which is longer than the specified 300\n",
      "Created a chunk of size 373, which is longer than the specified 300\n",
      "Created a chunk of size 580, which is longer than the specified 300\n",
      "Created a chunk of size 384, which is longer than the specified 300\n",
      "Created a chunk of size 522, which is longer than the specified 300\n",
      "Created a chunk of size 362, which is longer than the specified 300\n",
      "Created a chunk of size 525, which is longer than the specified 300\n",
      "Created a chunk of size 584, which is longer than the specified 300\n",
      "Created a chunk of size 747, which is longer than the specified 300\n",
      "Created a chunk of size 496, which is longer than the specified 300\n",
      "Created a chunk of size 967, which is longer than the specified 300\n",
      "Created a chunk of size 507, which is longer than the specified 300\n",
      "Created a chunk of size 510, which is longer than the specified 300\n",
      "Created a chunk of size 789, which is longer than the specified 300\n",
      "Created a chunk of size 609, which is longer than the specified 300\n",
      "Created a chunk of size 673, which is longer than the specified 300\n",
      "Created a chunk of size 441, which is longer than the specified 300\n",
      "Created a chunk of size 515, which is longer than the specified 300\n",
      "Created a chunk of size 484, which is longer than the specified 300\n",
      "Created a chunk of size 574, which is longer than the specified 300\n",
      "Created a chunk of size 918, which is longer than the specified 300\n",
      "Created a chunk of size 757, which is longer than the specified 300\n",
      "Created a chunk of size 355, which is longer than the specified 300\n",
      "Created a chunk of size 520, which is longer than the specified 300\n",
      "Created a chunk of size 494, which is longer than the specified 300\n",
      "Created a chunk of size 448, which is longer than the specified 300\n",
      "Created a chunk of size 623, which is longer than the specified 300\n",
      "Created a chunk of size 963, which is longer than the specified 300\n",
      "Created a chunk of size 421, which is longer than the specified 300\n",
      "Created a chunk of size 318, which is longer than the specified 300\n",
      "Created a chunk of size 617, which is longer than the specified 300\n",
      "Created a chunk of size 358, which is longer than the specified 300\n",
      "Created a chunk of size 557, which is longer than the specified 300\n",
      "Created a chunk of size 935, which is longer than the specified 300\n",
      "Created a chunk of size 610, which is longer than the specified 300\n",
      "Created a chunk of size 624, which is longer than the specified 300\n",
      "Created a chunk of size 481, which is longer than the specified 300\n",
      "Created a chunk of size 834, which is longer than the specified 300\n",
      "Created a chunk of size 540, which is longer than the specified 300\n",
      "Created a chunk of size 591, which is longer than the specified 300\n",
      "Created a chunk of size 1024, which is longer than the specified 300\n",
      "Created a chunk of size 768, which is longer than the specified 300\n",
      "Created a chunk of size 483, which is longer than the specified 300\n",
      "Created a chunk of size 530, which is longer than the specified 300\n",
      "Created a chunk of size 457, which is longer than the specified 300\n",
      "Created a chunk of size 451, which is longer than the specified 300\n",
      "Created a chunk of size 357, which is longer than the specified 300\n",
      "Created a chunk of size 458, which is longer than the specified 300\n",
      "Created a chunk of size 520, which is longer than the specified 300\n",
      "Created a chunk of size 652, which is longer than the specified 300\n",
      "Created a chunk of size 384, which is longer than the specified 300\n",
      "Created a chunk of size 869, which is longer than the specified 300\n",
      "Created a chunk of size 535, which is longer than the specified 300\n",
      "Created a chunk of size 978, which is longer than the specified 300\n",
      "Created a chunk of size 592, which is longer than the specified 300\n",
      "Created a chunk of size 375, which is longer than the specified 300\n",
      "Created a chunk of size 638, which is longer than the specified 300\n",
      "Created a chunk of size 584, which is longer than the specified 300\n",
      "Created a chunk of size 388, which is longer than the specified 300\n",
      "Created a chunk of size 523, which is longer than the specified 300\n",
      "Created a chunk of size 408, which is longer than the specified 300\n",
      "Created a chunk of size 517, which is longer than the specified 300\n",
      "Created a chunk of size 838, which is longer than the specified 300\n",
      "Created a chunk of size 703, which is longer than the specified 300\n",
      "Created a chunk of size 1523, which is longer than the specified 300\n",
      "Created a chunk of size 1071, which is longer than the specified 300\n",
      "Created a chunk of size 588, which is longer than the specified 300\n",
      "Created a chunk of size 745, which is longer than the specified 300\n",
      "Created a chunk of size 589, which is longer than the specified 300\n",
      "Created a chunk of size 607, which is longer than the specified 300\n",
      "Created a chunk of size 321, which is longer than the specified 300\n",
      "Created a chunk of size 378, which is longer than the specified 300\n",
      "Created a chunk of size 464, which is longer than the specified 300\n",
      "Created a chunk of size 537, which is longer than the specified 300\n",
      "Created a chunk of size 402, which is longer than the specified 300\n",
      "Created a chunk of size 700, which is longer than the specified 300\n",
      "Created a chunk of size 364, which is longer than the specified 300\n",
      "Created a chunk of size 304, which is longer than the specified 300\n",
      "Created a chunk of size 424, which is longer than the specified 300\n",
      "Created a chunk of size 447, which is longer than the specified 300\n",
      "Created a chunk of size 425, which is longer than the specified 300\n",
      "Created a chunk of size 410, which is longer than the specified 300\n",
      "Created a chunk of size 512, which is longer than the specified 300\n",
      "Created a chunk of size 574, which is longer than the specified 300\n",
      "Created a chunk of size 457, which is longer than the specified 300\n",
      "Created a chunk of size 386, which is longer than the specified 300\n",
      "Created a chunk of size 821, which is longer than the specified 300\n",
      "Created a chunk of size 695, which is longer than the specified 300\n",
      "Created a chunk of size 424, which is longer than the specified 300\n",
      "Created a chunk of size 828, which is longer than the specified 300\n",
      "Created a chunk of size 311, which is longer than the specified 300\n",
      "Created a chunk of size 472, which is longer than the specified 300\n",
      "Created a chunk of size 473, which is longer than the specified 300\n",
      "Created a chunk of size 352, which is longer than the specified 300\n",
      "Created a chunk of size 562, which is longer than the specified 300\n",
      "Created a chunk of size 475, which is longer than the specified 300\n",
      "Created a chunk of size 508, which is longer than the specified 300\n",
      "Created a chunk of size 978, which is longer than the specified 300\n",
      "Created a chunk of size 835, which is longer than the specified 300\n",
      "Created a chunk of size 995, which is longer than the specified 300\n",
      "Created a chunk of size 412, which is longer than the specified 300\n",
      "Created a chunk of size 675, which is longer than the specified 300\n",
      "Created a chunk of size 777, which is longer than the specified 300\n",
      "Created a chunk of size 457, which is longer than the specified 300\n",
      "Created a chunk of size 565, which is longer than the specified 300\n",
      "Created a chunk of size 601, which is longer than the specified 300\n",
      "Created a chunk of size 518, which is longer than the specified 300\n",
      "Created a chunk of size 330, which is longer than the specified 300\n",
      "Created a chunk of size 371, which is longer than the specified 300\n",
      "Created a chunk of size 499, which is longer than the specified 300\n",
      "Created a chunk of size 349, which is longer than the specified 300\n",
      "Created a chunk of size 758, which is longer than the specified 300\n",
      "Created a chunk of size 303, which is longer than the specified 300\n",
      "Created a chunk of size 717, which is longer than the specified 300\n",
      "Created a chunk of size 823, which is longer than the specified 300\n",
      "Created a chunk of size 541, which is longer than the specified 300\n",
      "Created a chunk of size 477, which is longer than the specified 300\n",
      "Created a chunk of size 313, which is longer than the specified 300\n",
      "Created a chunk of size 479, which is longer than the specified 300\n",
      "Created a chunk of size 306, which is longer than the specified 300\n",
      "Created a chunk of size 624, which is longer than the specified 300\n",
      "Created a chunk of size 442, which is longer than the specified 300\n",
      "Created a chunk of size 789, which is longer than the specified 300\n",
      "Created a chunk of size 560, which is longer than the specified 300\n",
      "Created a chunk of size 511, which is longer than the specified 300\n",
      "Created a chunk of size 475, which is longer than the specified 300\n",
      "Created a chunk of size 402, which is longer than the specified 300\n",
      "Created a chunk of size 1060, which is longer than the specified 300\n",
      "Created a chunk of size 325, which is longer than the specified 300\n",
      "Created a chunk of size 1002, which is longer than the specified 300\n",
      "Created a chunk of size 1222, which is longer than the specified 300\n",
      "Created a chunk of size 560, which is longer than the specified 300\n",
      "Created a chunk of size 857, which is longer than the specified 300\n",
      "Created a chunk of size 841, which is longer than the specified 300\n",
      "Created a chunk of size 1173, which is longer than the specified 300\n",
      "Created a chunk of size 935, which is longer than the specified 300\n",
      "Created a chunk of size 474, which is longer than the specified 300\n",
      "Created a chunk of size 345, which is longer than the specified 300\n",
      "Created a chunk of size 382, which is longer than the specified 300\n",
      "Created a chunk of size 301, which is longer than the specified 300\n",
      "Created a chunk of size 482, which is longer than the specified 300\n",
      "Created a chunk of size 339, which is longer than the specified 300\n",
      "Created a chunk of size 474, which is longer than the specified 300\n",
      "Created a chunk of size 1417, which is longer than the specified 300\n",
      "Created a chunk of size 666, which is longer than the specified 300\n"
     ]
    }
   ],
   "source": [
    "from langchain.indexes import VectorstoreIndexCreator\n",
    "from langchain.vectorstores import FAISS \n",
    "# More details about FAISS: https://engineering.fb.com/2017/03/29/data-infrastructure/faiss-a-library-for-efficient-similarity-search/\n",
    "\n",
    "pdf_loader = PDFMinerLoader(file_path=\"./arxiv_papers/2203.02155.pdf\")\n",
    "\n",
    "index = VectorstoreIndexCreator(\n",
    "    vectorstore_cls=FAISS,\n",
    "    embedding=instructor_embeds,\n",
    "    text_splitter=CharacterTextSplitter(\n",
    "        chunk_size=300,\n",
    "        # chunk_overlap=0\n",
    "    )\n",
    ").from_loaders(loaders=[pdf_loader])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'a lightweight classifier'"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = \"What is the best language model?\"\n",
    "results = index.query(question=query,\n",
    "                      llm = local_llm, )\n",
    "results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Created a chunk of size 1385, which is longer than the specified 300\n",
      "Created a chunk of size 491, which is longer than the specified 300\n",
      "Created a chunk of size 486, which is longer than the specified 300\n",
      "Created a chunk of size 480, which is longer than the specified 300\n",
      "Created a chunk of size 671, which is longer than the specified 300\n",
      "Created a chunk of size 1438, which is longer than the specified 300\n",
      "Created a chunk of size 506, which is longer than the specified 300\n",
      "Created a chunk of size 759, which is longer than the specified 300\n",
      "Created a chunk of size 615, which is longer than the specified 300\n",
      "Created a chunk of size 449, which is longer than the specified 300\n",
      "Created a chunk of size 471, which is longer than the specified 300\n",
      "Created a chunk of size 495, which is longer than the specified 300\n",
      "Created a chunk of size 715, which is longer than the specified 300\n",
      "Created a chunk of size 754, which is longer than the specified 300\n",
      "Created a chunk of size 752, which is longer than the specified 300\n",
      "Created a chunk of size 1078, which is longer than the specified 300\n",
      "Created a chunk of size 442, which is longer than the specified 300\n",
      "Created a chunk of size 802, which is longer than the specified 300\n",
      "Created a chunk of size 1510, which is longer than the specified 300\n",
      "Created a chunk of size 1607, which is longer than the specified 300\n",
      "Created a chunk of size 503, which is longer than the specified 300\n",
      "Created a chunk of size 973, which is longer than the specified 300\n",
      "Created a chunk of size 330, which is longer than the specified 300\n",
      "Created a chunk of size 632, which is longer than the specified 300\n",
      "Created a chunk of size 601, which is longer than the specified 300\n",
      "Created a chunk of size 504, which is longer than the specified 300\n",
      "Created a chunk of size 808, which is longer than the specified 300\n",
      "Created a chunk of size 848, which is longer than the specified 300\n",
      "Created a chunk of size 329, which is longer than the specified 300\n",
      "Created a chunk of size 319, which is longer than the specified 300\n",
      "Created a chunk of size 498, which is longer than the specified 300\n",
      "Created a chunk of size 395, which is longer than the specified 300\n",
      "Created a chunk of size 776, which is longer than the specified 300\n",
      "Created a chunk of size 373, which is longer than the specified 300\n",
      "Created a chunk of size 580, which is longer than the specified 300\n",
      "Created a chunk of size 384, which is longer than the specified 300\n",
      "Created a chunk of size 522, which is longer than the specified 300\n",
      "Created a chunk of size 362, which is longer than the specified 300\n",
      "Created a chunk of size 525, which is longer than the specified 300\n",
      "Created a chunk of size 584, which is longer than the specified 300\n",
      "Created a chunk of size 747, which is longer than the specified 300\n",
      "Created a chunk of size 496, which is longer than the specified 300\n",
      "Created a chunk of size 967, which is longer than the specified 300\n",
      "Created a chunk of size 507, which is longer than the specified 300\n",
      "Created a chunk of size 510, which is longer than the specified 300\n",
      "Created a chunk of size 789, which is longer than the specified 300\n",
      "Created a chunk of size 609, which is longer than the specified 300\n",
      "Created a chunk of size 673, which is longer than the specified 300\n",
      "Created a chunk of size 441, which is longer than the specified 300\n",
      "Created a chunk of size 515, which is longer than the specified 300\n",
      "Created a chunk of size 484, which is longer than the specified 300\n",
      "Created a chunk of size 574, which is longer than the specified 300\n",
      "Created a chunk of size 918, which is longer than the specified 300\n",
      "Created a chunk of size 757, which is longer than the specified 300\n",
      "Created a chunk of size 355, which is longer than the specified 300\n",
      "Created a chunk of size 520, which is longer than the specified 300\n",
      "Created a chunk of size 494, which is longer than the specified 300\n",
      "Created a chunk of size 448, which is longer than the specified 300\n",
      "Created a chunk of size 623, which is longer than the specified 300\n",
      "Created a chunk of size 963, which is longer than the specified 300\n",
      "Created a chunk of size 421, which is longer than the specified 300\n",
      "Created a chunk of size 318, which is longer than the specified 300\n",
      "Created a chunk of size 617, which is longer than the specified 300\n",
      "Created a chunk of size 358, which is longer than the specified 300\n",
      "Created a chunk of size 557, which is longer than the specified 300\n",
      "Created a chunk of size 935, which is longer than the specified 300\n",
      "Created a chunk of size 610, which is longer than the specified 300\n",
      "Created a chunk of size 624, which is longer than the specified 300\n",
      "Created a chunk of size 481, which is longer than the specified 300\n",
      "Created a chunk of size 834, which is longer than the specified 300\n",
      "Created a chunk of size 540, which is longer than the specified 300\n",
      "Created a chunk of size 591, which is longer than the specified 300\n",
      "Created a chunk of size 1024, which is longer than the specified 300\n",
      "Created a chunk of size 768, which is longer than the specified 300\n",
      "Created a chunk of size 483, which is longer than the specified 300\n",
      "Created a chunk of size 530, which is longer than the specified 300\n",
      "Created a chunk of size 457, which is longer than the specified 300\n",
      "Created a chunk of size 451, which is longer than the specified 300\n",
      "Created a chunk of size 357, which is longer than the specified 300\n",
      "Created a chunk of size 458, which is longer than the specified 300\n",
      "Created a chunk of size 520, which is longer than the specified 300\n",
      "Created a chunk of size 652, which is longer than the specified 300\n",
      "Created a chunk of size 384, which is longer than the specified 300\n",
      "Created a chunk of size 869, which is longer than the specified 300\n",
      "Created a chunk of size 535, which is longer than the specified 300\n",
      "Created a chunk of size 978, which is longer than the specified 300\n",
      "Created a chunk of size 592, which is longer than the specified 300\n",
      "Created a chunk of size 375, which is longer than the specified 300\n",
      "Created a chunk of size 638, which is longer than the specified 300\n",
      "Created a chunk of size 584, which is longer than the specified 300\n",
      "Created a chunk of size 388, which is longer than the specified 300\n",
      "Created a chunk of size 523, which is longer than the specified 300\n",
      "Created a chunk of size 408, which is longer than the specified 300\n",
      "Created a chunk of size 517, which is longer than the specified 300\n",
      "Created a chunk of size 838, which is longer than the specified 300\n",
      "Created a chunk of size 703, which is longer than the specified 300\n",
      "Created a chunk of size 1523, which is longer than the specified 300\n",
      "Created a chunk of size 1071, which is longer than the specified 300\n",
      "Created a chunk of size 588, which is longer than the specified 300\n",
      "Created a chunk of size 745, which is longer than the specified 300\n",
      "Created a chunk of size 589, which is longer than the specified 300\n",
      "Created a chunk of size 607, which is longer than the specified 300\n",
      "Created a chunk of size 321, which is longer than the specified 300\n",
      "Created a chunk of size 378, which is longer than the specified 300\n",
      "Created a chunk of size 464, which is longer than the specified 300\n",
      "Created a chunk of size 537, which is longer than the specified 300\n",
      "Created a chunk of size 402, which is longer than the specified 300\n",
      "Created a chunk of size 700, which is longer than the specified 300\n",
      "Created a chunk of size 364, which is longer than the specified 300\n",
      "Created a chunk of size 304, which is longer than the specified 300\n",
      "Created a chunk of size 424, which is longer than the specified 300\n",
      "Created a chunk of size 447, which is longer than the specified 300\n",
      "Created a chunk of size 425, which is longer than the specified 300\n",
      "Created a chunk of size 410, which is longer than the specified 300\n",
      "Created a chunk of size 512, which is longer than the specified 300\n",
      "Created a chunk of size 574, which is longer than the specified 300\n",
      "Created a chunk of size 457, which is longer than the specified 300\n",
      "Created a chunk of size 386, which is longer than the specified 300\n",
      "Created a chunk of size 821, which is longer than the specified 300\n",
      "Created a chunk of size 695, which is longer than the specified 300\n",
      "Created a chunk of size 424, which is longer than the specified 300\n",
      "Created a chunk of size 828, which is longer than the specified 300\n",
      "Created a chunk of size 311, which is longer than the specified 300\n",
      "Created a chunk of size 472, which is longer than the specified 300\n",
      "Created a chunk of size 473, which is longer than the specified 300\n",
      "Created a chunk of size 352, which is longer than the specified 300\n",
      "Created a chunk of size 562, which is longer than the specified 300\n",
      "Created a chunk of size 475, which is longer than the specified 300\n",
      "Created a chunk of size 508, which is longer than the specified 300\n",
      "Created a chunk of size 978, which is longer than the specified 300\n",
      "Created a chunk of size 835, which is longer than the specified 300\n",
      "Created a chunk of size 995, which is longer than the specified 300\n",
      "Created a chunk of size 412, which is longer than the specified 300\n",
      "Created a chunk of size 675, which is longer than the specified 300\n",
      "Created a chunk of size 777, which is longer than the specified 300\n",
      "Created a chunk of size 457, which is longer than the specified 300\n",
      "Created a chunk of size 565, which is longer than the specified 300\n",
      "Created a chunk of size 601, which is longer than the specified 300\n",
      "Created a chunk of size 518, which is longer than the specified 300\n",
      "Created a chunk of size 330, which is longer than the specified 300\n",
      "Created a chunk of size 371, which is longer than the specified 300\n",
      "Created a chunk of size 499, which is longer than the specified 300\n",
      "Created a chunk of size 349, which is longer than the specified 300\n",
      "Created a chunk of size 758, which is longer than the specified 300\n",
      "Created a chunk of size 303, which is longer than the specified 300\n",
      "Created a chunk of size 717, which is longer than the specified 300\n",
      "Created a chunk of size 823, which is longer than the specified 300\n",
      "Created a chunk of size 541, which is longer than the specified 300\n",
      "Created a chunk of size 477, which is longer than the specified 300\n",
      "Created a chunk of size 313, which is longer than the specified 300\n",
      "Created a chunk of size 479, which is longer than the specified 300\n",
      "Created a chunk of size 306, which is longer than the specified 300\n",
      "Created a chunk of size 624, which is longer than the specified 300\n",
      "Created a chunk of size 442, which is longer than the specified 300\n",
      "Created a chunk of size 789, which is longer than the specified 300\n",
      "Created a chunk of size 560, which is longer than the specified 300\n",
      "Created a chunk of size 511, which is longer than the specified 300\n",
      "Created a chunk of size 475, which is longer than the specified 300\n",
      "Created a chunk of size 402, which is longer than the specified 300\n",
      "Created a chunk of size 1060, which is longer than the specified 300\n",
      "Created a chunk of size 325, which is longer than the specified 300\n",
      "Created a chunk of size 1002, which is longer than the specified 300\n",
      "Created a chunk of size 1222, which is longer than the specified 300\n",
      "Created a chunk of size 560, which is longer than the specified 300\n",
      "Created a chunk of size 857, which is longer than the specified 300\n",
      "Created a chunk of size 841, which is longer than the specified 300\n",
      "Created a chunk of size 1173, which is longer than the specified 300\n",
      "Created a chunk of size 935, which is longer than the specified 300\n",
      "Created a chunk of size 474, which is longer than the specified 300\n",
      "Created a chunk of size 345, which is longer than the specified 300\n",
      "Created a chunk of size 382, which is longer than the specified 300\n",
      "Created a chunk of size 301, which is longer than the specified 300\n",
      "Created a chunk of size 482, which is longer than the specified 300\n",
      "Created a chunk of size 339, which is longer than the specified 300\n",
      "Created a chunk of size 474, which is longer than the specified 300\n",
      "Created a chunk of size 1417, which is longer than the specified 300\n",
      "Created a chunk of size 666, which is longer than the specified 300\n",
      "Created a chunk of size 1778, which is longer than the specified 300\n",
      "Created a chunk of size 589, which is longer than the specified 300\n",
      "Created a chunk of size 303, which is longer than the specified 300\n",
      "Created a chunk of size 682, which is longer than the specified 300\n",
      "Created a chunk of size 673, which is longer than the specified 300\n",
      "Created a chunk of size 490, which is longer than the specified 300\n",
      "Created a chunk of size 894, which is longer than the specified 300\n",
      "Created a chunk of size 353, which is longer than the specified 300\n",
      "Created a chunk of size 618, which is longer than the specified 300\n",
      "Created a chunk of size 475, which is longer than the specified 300\n",
      "Created a chunk of size 1168, which is longer than the specified 300\n",
      "Created a chunk of size 379, which is longer than the specified 300\n",
      "Created a chunk of size 816, which is longer than the specified 300\n",
      "Created a chunk of size 1049, which is longer than the specified 300\n",
      "Created a chunk of size 351, which is longer than the specified 300\n",
      "Created a chunk of size 885, which is longer than the specified 300\n",
      "Created a chunk of size 673, which is longer than the specified 300\n",
      "Created a chunk of size 662, which is longer than the specified 300\n",
      "Created a chunk of size 416, which is longer than the specified 300\n",
      "Created a chunk of size 463, which is longer than the specified 300\n",
      "Created a chunk of size 652, which is longer than the specified 300\n",
      "Created a chunk of size 507, which is longer than the specified 300\n",
      "Created a chunk of size 446, which is longer than the specified 300\n",
      "Created a chunk of size 765, which is longer than the specified 300\n",
      "Created a chunk of size 818, which is longer than the specified 300\n",
      "Created a chunk of size 1345, which is longer than the specified 300\n",
      "Created a chunk of size 599, which is longer than the specified 300\n",
      "Created a chunk of size 520, which is longer than the specified 300\n",
      "Created a chunk of size 1072, which is longer than the specified 300\n",
      "Created a chunk of size 617, which is longer than the specified 300\n",
      "Created a chunk of size 857, which is longer than the specified 300\n",
      "Created a chunk of size 830, which is longer than the specified 300\n",
      "Created a chunk of size 920, which is longer than the specified 300\n",
      "Created a chunk of size 389, which is longer than the specified 300\n",
      "Created a chunk of size 699, which is longer than the specified 300\n",
      "Created a chunk of size 417, which is longer than the specified 300\n",
      "Created a chunk of size 381, which is longer than the specified 300\n",
      "Created a chunk of size 750, which is longer than the specified 300\n",
      "Created a chunk of size 640, which is longer than the specified 300\n",
      "Created a chunk of size 489, which is longer than the specified 300\n",
      "Created a chunk of size 506, which is longer than the specified 300\n",
      "Created a chunk of size 703, which is longer than the specified 300\n",
      "Created a chunk of size 554, which is longer than the specified 300\n",
      "Created a chunk of size 503, which is longer than the specified 300\n",
      "Created a chunk of size 726, which is longer than the specified 300\n",
      "Created a chunk of size 1088, which is longer than the specified 300\n",
      "Created a chunk of size 425, which is longer than the specified 300\n",
      "Created a chunk of size 697, which is longer than the specified 300\n",
      "Created a chunk of size 461, which is longer than the specified 300\n",
      "Created a chunk of size 332, which is longer than the specified 300\n",
      "Created a chunk of size 748, which is longer than the specified 300\n",
      "Created a chunk of size 567, which is longer than the specified 300\n",
      "Created a chunk of size 528, which is longer than the specified 300\n",
      "Created a chunk of size 390, which is longer than the specified 300\n",
      "Created a chunk of size 1296, which is longer than the specified 300\n",
      "Created a chunk of size 645, which is longer than the specified 300\n",
      "Created a chunk of size 557, which is longer than the specified 300\n",
      "Created a chunk of size 339, which is longer than the specified 300\n",
      "Created a chunk of size 554, which is longer than the specified 300\n",
      "Created a chunk of size 454, which is longer than the specified 300\n",
      "Created a chunk of size 1037, which is longer than the specified 300\n",
      "Created a chunk of size 692, which is longer than the specified 300\n",
      "Created a chunk of size 653, which is longer than the specified 300\n",
      "Created a chunk of size 1243, which is longer than the specified 300\n",
      "Created a chunk of size 341, which is longer than the specified 300\n",
      "Created a chunk of size 759, which is longer than the specified 300\n",
      "Created a chunk of size 394, which is longer than the specified 300\n",
      "Created a chunk of size 568, which is longer than the specified 300\n",
      "Created a chunk of size 1310, which is longer than the specified 300\n",
      "Created a chunk of size 328, which is longer than the specified 300\n",
      "Created a chunk of size 486, which is longer than the specified 300\n",
      "Created a chunk of size 1115, which is longer than the specified 300\n",
      "Created a chunk of size 537, which is longer than the specified 300\n",
      "Created a chunk of size 323, which is longer than the specified 300\n",
      "Created a chunk of size 856, which is longer than the specified 300\n",
      "Created a chunk of size 669, which is longer than the specified 300\n",
      "Created a chunk of size 992, which is longer than the specified 300\n",
      "Created a chunk of size 478, which is longer than the specified 300\n",
      "Created a chunk of size 378, which is longer than the specified 300\n",
      "Created a chunk of size 1129, which is longer than the specified 300\n",
      "Created a chunk of size 918, which is longer than the specified 300\n",
      "Created a chunk of size 502, which is longer than the specified 300\n",
      "Created a chunk of size 303, which is longer than the specified 300\n",
      "Created a chunk of size 852, which is longer than the specified 300\n",
      "Created a chunk of size 353, which is longer than the specified 300\n",
      "Created a chunk of size 352, which is longer than the specified 300\n",
      "Created a chunk of size 326, which is longer than the specified 300\n",
      "Created a chunk of size 689, which is longer than the specified 300\n",
      "Created a chunk of size 333, which is longer than the specified 300\n",
      "Created a chunk of size 341, which is longer than the specified 300\n",
      "Created a chunk of size 391, which is longer than the specified 300\n",
      "Created a chunk of size 433, which is longer than the specified 300\n",
      "Created a chunk of size 407, which is longer than the specified 300\n",
      "Created a chunk of size 641, which is longer than the specified 300\n",
      "Created a chunk of size 1054, which is longer than the specified 300\n",
      "Created a chunk of size 868, which is longer than the specified 300\n",
      "Created a chunk of size 535, which is longer than the specified 300\n",
      "Created a chunk of size 760, which is longer than the specified 300\n",
      "Created a chunk of size 649, which is longer than the specified 300\n",
      "Created a chunk of size 519, which is longer than the specified 300\n",
      "Created a chunk of size 1254, which is longer than the specified 300\n",
      "Created a chunk of size 314, which is longer than the specified 300\n",
      "Created a chunk of size 658, which is longer than the specified 300\n",
      "Created a chunk of size 448, which is longer than the specified 300\n",
      "Created a chunk of size 474, which is longer than the specified 300\n",
      "Created a chunk of size 507, which is longer than the specified 300\n",
      "Created a chunk of size 417, which is longer than the specified 300\n",
      "Created a chunk of size 1549, which is longer than the specified 300\n",
      "Created a chunk of size 1108, which is longer than the specified 300\n",
      "Created a chunk of size 417, which is longer than the specified 300\n",
      "Created a chunk of size 464, which is longer than the specified 300\n",
      "Created a chunk of size 929, which is longer than the specified 300\n",
      "Created a chunk of size 379, which is longer than the specified 300\n",
      "Created a chunk of size 475, which is longer than the specified 300\n",
      "Created a chunk of size 340, which is longer than the specified 300\n",
      "Created a chunk of size 918, which is longer than the specified 300\n",
      "Created a chunk of size 371, which is longer than the specified 300\n",
      "Created a chunk of size 607, which is longer than the specified 300\n",
      "Created a chunk of size 490, which is longer than the specified 300\n",
      "Created a chunk of size 481, which is longer than the specified 300\n",
      "Created a chunk of size 796, which is longer than the specified 300\n",
      "Created a chunk of size 676, which is longer than the specified 300\n",
      "Created a chunk of size 502, which is longer than the specified 300\n",
      "Created a chunk of size 369, which is longer than the specified 300\n",
      "Created a chunk of size 780, which is longer than the specified 300\n",
      "Created a chunk of size 758, which is longer than the specified 300\n",
      "Created a chunk of size 431, which is longer than the specified 300\n",
      "Created a chunk of size 455, which is longer than the specified 300\n",
      "Created a chunk of size 751, which is longer than the specified 300\n",
      "Created a chunk of size 519, which is longer than the specified 300\n",
      "Created a chunk of size 411, which is longer than the specified 300\n",
      "Created a chunk of size 1514, which is longer than the specified 300\n",
      "Created a chunk of size 1624, which is longer than the specified 300\n",
      "Created a chunk of size 1153, which is longer than the specified 300\n",
      "Created a chunk of size 510, which is longer than the specified 300\n",
      "Created a chunk of size 1279, which is longer than the specified 300\n",
      "Created a chunk of size 812, which is longer than the specified 300\n",
      "Created a chunk of size 605, which is longer than the specified 300\n",
      "Created a chunk of size 551, which is longer than the specified 300\n",
      "Created a chunk of size 517, which is longer than the specified 300\n",
      "Created a chunk of size 582, which is longer than the specified 300\n",
      "Created a chunk of size 512, which is longer than the specified 300\n",
      "Created a chunk of size 303, which is longer than the specified 300\n",
      "Created a chunk of size 660, which is longer than the specified 300\n",
      "Created a chunk of size 670, which is longer than the specified 300\n",
      "Created a chunk of size 474, which is longer than the specified 300\n",
      "Created a chunk of size 625, which is longer than the specified 300\n",
      "Created a chunk of size 466, which is longer than the specified 300\n",
      "Created a chunk of size 458, which is longer than the specified 300\n",
      "Created a chunk of size 398, which is longer than the specified 300\n",
      "Created a chunk of size 400, which is longer than the specified 300\n",
      "Created a chunk of size 1044, which is longer than the specified 300\n",
      "Created a chunk of size 841, which is longer than the specified 300\n",
      "Created a chunk of size 610, which is longer than the specified 300\n",
      "Created a chunk of size 965, which is longer than the specified 300\n",
      "Created a chunk of size 500, which is longer than the specified 300\n",
      "Created a chunk of size 457, which is longer than the specified 300\n",
      "Created a chunk of size 968, which is longer than the specified 300\n",
      "Created a chunk of size 658, which is longer than the specified 300\n",
      "Created a chunk of size 463, which is longer than the specified 300\n",
      "Created a chunk of size 328, which is longer than the specified 300\n",
      "Created a chunk of size 516, which is longer than the specified 300\n",
      "Created a chunk of size 518, which is longer than the specified 300\n",
      "Created a chunk of size 698, which is longer than the specified 300\n",
      "Created a chunk of size 888, which is longer than the specified 300\n",
      "Created a chunk of size 381, which is longer than the specified 300\n",
      "Created a chunk of size 1055, which is longer than the specified 300\n",
      "Created a chunk of size 2096, which is longer than the specified 300\n",
      "Created a chunk of size 452, which is longer than the specified 300\n",
      "Created a chunk of size 1472, which is longer than the specified 300\n",
      "Created a chunk of size 384, which is longer than the specified 300\n",
      "Created a chunk of size 427, which is longer than the specified 300\n",
      "Created a chunk of size 1225, which is longer than the specified 300\n",
      "Created a chunk of size 970, which is longer than the specified 300\n",
      "Created a chunk of size 480, which is longer than the specified 300\n",
      "Created a chunk of size 1110, which is longer than the specified 300\n",
      "Created a chunk of size 830, which is longer than the specified 300\n",
      "Created a chunk of size 345, which is longer than the specified 300\n",
      "Created a chunk of size 471, which is longer than the specified 300\n",
      "Created a chunk of size 755, which is longer than the specified 300\n",
      "Created a chunk of size 592, which is longer than the specified 300\n",
      "Created a chunk of size 564, which is longer than the specified 300\n",
      "Created a chunk of size 657, which is longer than the specified 300\n",
      "Created a chunk of size 827, which is longer than the specified 300\n",
      "Created a chunk of size 511, which is longer than the specified 300\n",
      "Created a chunk of size 560, which is longer than the specified 300\n",
      "Created a chunk of size 981, which is longer than the specified 300\n",
      "Created a chunk of size 326, which is longer than the specified 300\n",
      "Created a chunk of size 369, which is longer than the specified 300\n",
      "Created a chunk of size 693, which is longer than the specified 300\n",
      "Created a chunk of size 880, which is longer than the specified 300\n",
      "Created a chunk of size 315, which is longer than the specified 300\n",
      "Created a chunk of size 757, which is longer than the specified 300\n",
      "Created a chunk of size 666, which is longer than the specified 300\n",
      "Created a chunk of size 675, which is longer than the specified 300\n",
      "Created a chunk of size 438, which is longer than the specified 300\n",
      "Created a chunk of size 340, which is longer than the specified 300\n",
      "Created a chunk of size 932, which is longer than the specified 300\n",
      "Created a chunk of size 563, which is longer than the specified 300\n",
      "Created a chunk of size 401, which is longer than the specified 300\n",
      "Created a chunk of size 629, which is longer than the specified 300\n",
      "Created a chunk of size 809, which is longer than the specified 300\n",
      "Created a chunk of size 1664, which is longer than the specified 300\n",
      "Created a chunk of size 839, which is longer than the specified 300\n",
      "Created a chunk of size 536, which is longer than the specified 300\n",
      "Created a chunk of size 373, which is longer than the specified 300\n",
      "Created a chunk of size 1928, which is longer than the specified 300\n",
      "Created a chunk of size 419, which is longer than the specified 300\n",
      "Created a chunk of size 367, which is longer than the specified 300\n",
      "Created a chunk of size 744, which is longer than the specified 300\n",
      "Created a chunk of size 559, which is longer than the specified 300\n",
      "Created a chunk of size 1959, which is longer than the specified 300\n",
      "Created a chunk of size 483, which is longer than the specified 300\n",
      "Created a chunk of size 948, which is longer than the specified 300\n",
      "Created a chunk of size 359, which is longer than the specified 300\n",
      "Created a chunk of size 347, which is longer than the specified 300\n",
      "Created a chunk of size 514, which is longer than the specified 300\n",
      "Created a chunk of size 479, which is longer than the specified 300\n",
      "Created a chunk of size 479, which is longer than the specified 300\n",
      "Created a chunk of size 479, which is longer than the specified 300\n",
      "Created a chunk of size 479, which is longer than the specified 300\n",
      "Created a chunk of size 479, which is longer than the specified 300\n",
      "Created a chunk of size 479, which is longer than the specified 300\n",
      "Created a chunk of size 399, which is longer than the specified 300\n",
      "Created a chunk of size 399, which is longer than the specified 300\n",
      "Created a chunk of size 400, which is longer than the specified 300\n",
      "Created a chunk of size 427, which is longer than the specified 300\n",
      "Created a chunk of size 354, which is longer than the specified 300\n",
      "Created a chunk of size 305, which is longer than the specified 300\n",
      "Created a chunk of size 324, which is longer than the specified 300\n",
      "Created a chunk of size 301, which is longer than the specified 300\n",
      "Created a chunk of size 877, which is longer than the specified 300\n",
      "Created a chunk of size 416, which is longer than the specified 300\n",
      "Created a chunk of size 563, which is longer than the specified 300\n",
      "Created a chunk of size 693, which is longer than the specified 300\n",
      "Created a chunk of size 377, which is longer than the specified 300\n",
      "Created a chunk of size 589, which is longer than the specified 300\n",
      "Created a chunk of size 639, which is longer than the specified 300\n",
      "Created a chunk of size 376, which is longer than the specified 300\n",
      "Created a chunk of size 491, which is longer than the specified 300\n",
      "Created a chunk of size 520, which is longer than the specified 300\n",
      "Created a chunk of size 515, which is longer than the specified 300\n",
      "Created a chunk of size 537, which is longer than the specified 300\n",
      "Created a chunk of size 316, which is longer than the specified 300\n",
      "Created a chunk of size 609, which is longer than the specified 300\n",
      "Created a chunk of size 645, which is longer than the specified 300\n",
      "Created a chunk of size 333, which is longer than the specified 300\n",
      "Created a chunk of size 477, which is longer than the specified 300\n",
      "Created a chunk of size 673, which is longer than the specified 300\n",
      "Created a chunk of size 588, which is longer than the specified 300\n",
      "Created a chunk of size 360, which is longer than the specified 300\n",
      "Created a chunk of size 577, which is longer than the specified 300\n",
      "Created a chunk of size 726, which is longer than the specified 300\n",
      "Created a chunk of size 593, which is longer than the specified 300\n",
      "Created a chunk of size 565, which is longer than the specified 300\n",
      "Created a chunk of size 669, which is longer than the specified 300\n",
      "Created a chunk of size 301, which is longer than the specified 300\n",
      "Created a chunk of size 689, which is longer than the specified 300\n",
      "Created a chunk of size 551, which is longer than the specified 300\n",
      "Created a chunk of size 481, which is longer than the specified 300\n",
      "Created a chunk of size 464, which is longer than the specified 300\n",
      "Created a chunk of size 682, which is longer than the specified 300\n",
      "Created a chunk of size 490, which is longer than the specified 300\n",
      "Created a chunk of size 416, which is longer than the specified 300\n",
      "Created a chunk of size 686, which is longer than the specified 300\n",
      "Created a chunk of size 634, which is longer than the specified 300\n",
      "Created a chunk of size 1038, which is longer than the specified 300\n",
      "Created a chunk of size 367, which is longer than the specified 300\n",
      "Created a chunk of size 656, which is longer than the specified 300\n",
      "Created a chunk of size 1200, which is longer than the specified 300\n",
      "Created a chunk of size 527, which is longer than the specified 300\n",
      "Created a chunk of size 501, which is longer than the specified 300\n",
      "Created a chunk of size 470, which is longer than the specified 300\n",
      "Created a chunk of size 563, which is longer than the specified 300\n",
      "Created a chunk of size 738, which is longer than the specified 300\n",
      "Created a chunk of size 516, which is longer than the specified 300\n",
      "Created a chunk of size 364, which is longer than the specified 300\n",
      "Created a chunk of size 643, which is longer than the specified 300\n",
      "Created a chunk of size 881, which is longer than the specified 300\n",
      "Created a chunk of size 586, which is longer than the specified 300\n",
      "Created a chunk of size 386, which is longer than the specified 300\n",
      "Created a chunk of size 400, which is longer than the specified 300\n",
      "Created a chunk of size 315, which is longer than the specified 300\n",
      "Created a chunk of size 641, which is longer than the specified 300\n",
      "Created a chunk of size 369, which is longer than the specified 300\n",
      "Created a chunk of size 984, which is longer than the specified 300\n",
      "Created a chunk of size 647, which is longer than the specified 300\n",
      "Created a chunk of size 777, which is longer than the specified 300\n",
      "Created a chunk of size 444, which is longer than the specified 300\n",
      "Created a chunk of size 453, which is longer than the specified 300\n",
      "Created a chunk of size 826, which is longer than the specified 300\n",
      "Created a chunk of size 955, which is longer than the specified 300\n",
      "Created a chunk of size 305, which is longer than the specified 300\n",
      "Created a chunk of size 377, which is longer than the specified 300\n",
      "Created a chunk of size 526, which is longer than the specified 300\n",
      "Created a chunk of size 396, which is longer than the specified 300\n",
      "Created a chunk of size 648, which is longer than the specified 300\n",
      "Created a chunk of size 329, which is longer than the specified 300\n",
      "Created a chunk of size 360, which is longer than the specified 300\n",
      "Created a chunk of size 392, which is longer than the specified 300\n",
      "Created a chunk of size 696, which is longer than the specified 300\n",
      "Created a chunk of size 615, which is longer than the specified 300\n",
      "Created a chunk of size 304, which is longer than the specified 300\n",
      "Created a chunk of size 791, which is longer than the specified 300\n",
      "Created a chunk of size 1128, which is longer than the specified 300\n",
      "Created a chunk of size 760, which is longer than the specified 300\n",
      "Created a chunk of size 412, which is longer than the specified 300\n",
      "Created a chunk of size 520, which is longer than the specified 300\n",
      "Created a chunk of size 392, which is longer than the specified 300\n",
      "Created a chunk of size 764, which is longer than the specified 300\n",
      "Created a chunk of size 763, which is longer than the specified 300\n",
      "Created a chunk of size 694, which is longer than the specified 300\n",
      "Created a chunk of size 766, which is longer than the specified 300\n",
      "Created a chunk of size 744, which is longer than the specified 300\n",
      "Created a chunk of size 417, which is longer than the specified 300\n",
      "Created a chunk of size 719, which is longer than the specified 300\n",
      "Created a chunk of size 411, which is longer than the specified 300\n",
      "Created a chunk of size 490, which is longer than the specified 300\n",
      "Created a chunk of size 498, which is longer than the specified 300\n",
      "Created a chunk of size 673, which is longer than the specified 300\n",
      "Created a chunk of size 707, which is longer than the specified 300\n",
      "Created a chunk of size 313, which is longer than the specified 300\n",
      "Created a chunk of size 313, which is longer than the specified 300\n",
      "Created a chunk of size 832, which is longer than the specified 300\n",
      "Created a chunk of size 1245, which is longer than the specified 300\n",
      "Created a chunk of size 769, which is longer than the specified 300\n",
      "Created a chunk of size 730, which is longer than the specified 300\n",
      "Created a chunk of size 402, which is longer than the specified 300\n",
      "Created a chunk of size 1608, which is longer than the specified 300\n",
      "Created a chunk of size 348, which is longer than the specified 300\n",
      "Created a chunk of size 308, which is longer than the specified 300\n",
      "Created a chunk of size 570, which is longer than the specified 300\n",
      "Created a chunk of size 347, which is longer than the specified 300\n",
      "Created a chunk of size 518, which is longer than the specified 300\n",
      "Created a chunk of size 334, which is longer than the specified 300\n"
     ]
    }
   ],
   "source": [
    "index = VectorstoreIndexCreator(\n",
    "    vectorstore_cls=FAISS,\n",
    "    embedding=instructor_embeds,\n",
    "    text_splitter=CharacterTextSplitter(\n",
    "        chunk_size=300,\n",
    "        chunk_overlap=0\n",
    "    )\n",
    ").from_loaders(loaders=loaders)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'GPT-4 substantially improves over previous models in the ability to follow user intent'"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = \"What can you tell me about GPT from OpenAI?\"\n",
    "results = index.query(question=query,\n",
    "                      llm = local_llm)\n",
    "results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#TODO: Try querying with ElasticSearch\n",
    "\n",
    "from langchain.vectorstores import ElasticVectorSearch\n",
    "index = VectorstoreIndexCreator(\n",
    "    vectorstore_cls=ElasticVectorSearch,\n",
    "    embedding=instructor_embeds,\n",
    "    text_splitter=CharacterTextSplitter(\n",
    "        chunk_size=300,\n",
    "        chunk_overlap=0\n",
    "    )\n",
    ").from_loaders(loaders=loaders)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LangChain: Prompts\n",
    "\n",
    "Prompt Management, Optimization and Serialization. Read more [here](https://docs.langchain.com/docs/components/prompts)\n",
    "\n",
    "```python\n",
    "from langchain import PromptTemplate\n",
    "\n",
    "template = \"\"\"Question: {question}\n",
    "\n",
    "Let's think step by step.\n",
    "\n",
    "Answer: \"\"\"\n",
    "\n",
    "prompt = PromptTemplate(template=template, input_variables=[\"question\"])\n",
    "user_input = input(\"Enter your question: \")\n",
    "prompt.format(question=user_input)\n",
    "```"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LangChain: Chain\n",
    "\n",
    "Sequence of calls (with multiple models). Read more [here](https://python.langchain.com/en/latest/modules/chains.html)  \n",
    "[] `LLMChain`  \n",
    "[] `SequentialChain`  \n",
    "[] Custom `Chain`  \n",
    "\n",
    "```python\n",
    "from langchain.prompts import PromptTemplate\n",
    "from langchain.llms import OpenAI\n",
    "\n",
    "llm = OpenAI(temperature=0.9)\n",
    "prompt = PromptTemplate(\n",
    "    input_variables=[\"product\"],\n",
    "    template=\"Which country is the largest producer of {product}?\",\n",
    ")\n",
    "\n",
    "from langchain.chains import LLMChain\n",
    "# Chain a LLM with a prompt template\n",
    "chain = LLMChain(llm=llm, prompt=prompt)\n",
    "\n",
    "# Run the chain only specifying the input variable.\n",
    "print(chain.run(\"lithium\"))\n",
    "```"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "```python\n",
    "from langchain.chains import SimpleSequentialChain\n",
    "\n",
    "q = \"What type of mammal lays the biggest eggs?\"\n",
    "\n",
    "llm = OpenAI(temperature=0.7)\n",
    "template = \"\"\"{question}\\n\\n\"\"\"\n",
    "prompt_template = PromptTemplate(input_variables=[\"question\"], template=template)\n",
    "question_chain = LLMChain(llm=llm, prompt=prompt_template)\n",
    "\n",
    "template = \"\"\"Here is a statement:\n",
    "{statement}\n",
    "Make a bullet point list of the assumptions you made when producing the above statement.\\n\\n\"\"\"\n",
    "prompt_template = PromptTemplate(input_variables=[\"statement\"], template=template)\n",
    "assumptions_chain = LLMChain(llm=llm, prompt=prompt_template)\n",
    "\n",
    "template = \"\"\"Here is a bullet point list of assertions:\n",
    "{assertions}\n",
    "For each assertion, determine whether it is true or false. If it is false, explain why.\\n\\n\"\"\"\n",
    "prompt_template = PromptTemplate(input_variables=[\"assertions\"], template=template)\n",
    "fact_checker_chain = LLMChain(llm=llm, prompt=prompt_template)\n",
    "\n",
    "template = \"\"\"In light of the above facts, how would you answer the question '{}'\"\"\".format(q)\n",
    "template = \"\"\"{facts}\\n\"\"\" + template\n",
    "prompt_template = PromptTemplate(input_variables=[\"facts\"], template=template)\n",
    "answer_chain = LLMChain(llm=llm, prompt=prompt_template)\n",
    "\n",
    "overall_chain = SimpleSequentialChain(chains=[question_chain, assumptions_chain, fact_checker_chain, answer_chain], verbose=True)\n",
    "```\n",
    "\n",
    "Credit to jagilley/fact-checker for the example. Check the repo [here](https://github.com/jagilley/fact-checker)."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "![Importance of SeqChains](https://weaviate.io/assets/images/sequential-chains-fec82f27b64a0d8f5b6123b39569ecf2.gif)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain.prompts import PromptTemplate\n",
    "from langchain.chains import LLMChain, SimpleSequentialChain\n",
    "\n",
    "question = query\n",
    "\n",
    "template = \"\"\"{question}\\n\\n\"\"\"\n",
    "prompt_template = PromptTemplate(input_variables=[\"question\"], template=template)\n",
    "question_chain = LLMChain(llm=local_llm, prompt=prompt_template)\n",
    "\n",
    "template = \"\"\"Here is a statement:\n",
    "{statement}\n",
    "Make a bullet point list of the assumptions you made when producing the above statement.\\n\\n\"\"\"\n",
    "prompt_template = PromptTemplate(input_variables=[\"statement\"], template=template)\n",
    "assumptions_chain = LLMChain(llm=local_llm, prompt=prompt_template)\n",
    "\n",
    "template = \"\"\"Here is a bullet point list of assertions:\n",
    "{assertions}\n",
    "For each assertion, determine whether it is true or false. If it is false, explain why.\\n\\n\"\"\"\n",
    "prompt_template = PromptTemplate(input_variables=[\"assertions\"], template=template)\n",
    "fact_checker_chain = LLMChain(llm=local_llm, prompt=prompt_template)\n",
    "\n",
    "template = \"\"\"In light of the above facts, how would you answer the question '{}'\"\"\".format(question)\n",
    "template = \"\"\"{facts}\\n\"\"\" + template\n",
    "prompt_template = PromptTemplate(input_variables=[\"facts\"], template=template)\n",
    "answer_chain = LLMChain(llm=local_llm, prompt=prompt_template)\n",
    "\n",
    "overall_chain = SimpleSequentialChain(chains=[question_chain, assumptions_chain, fact_checker_chain, answer_chain], verbose=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "What can you tell me about GPT from OpenAI?\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #800000; text-decoration-color: #800000\">╭─────────────────────────────── </span><span style=\"color: #800000; text-decoration-color: #800000; font-weight: bold\">Traceback </span><span style=\"color: #bf7f7f; text-decoration-color: #bf7f7f; font-weight: bold\">(most recent call last)</span><span style=\"color: #800000; text-decoration-color: #800000\"> ────────────────────────────────╮</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/tmp/ipykernel_4172/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">952069505.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">2</span> in <span style=\"color: #00ff00; text-decoration-color: #00ff00\">&lt;module&gt;</span>                                                   <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000; font-style: italic\">[Errno 2] No such file or directory: '/tmp/ipykernel_4172/952069505.py'</span>                          <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">base.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">216</span> in <span style=\"color: #00ff00; text-decoration-color: #00ff00\">run</span>       <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">213 │   │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">return</span> <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>(args[<span style=\"color: #0000ff; text-decoration-color: #0000ff\">0</span>])[<span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.output_keys[<span style=\"color: #0000ff; text-decoration-color: #0000ff\">0</span>]]                                      <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">214 │   │   </span>                                                                                   <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">215 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">if</span> kwargs <span style=\"color: #ff00ff; text-decoration-color: #ff00ff\">and</span> <span style=\"color: #ff00ff; text-decoration-color: #ff00ff\">not</span> args:                                                            <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000\">❱ </span>216 <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">│   │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">return</span> <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>(kwargs)[<span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.output_keys[<span style=\"color: #0000ff; text-decoration-color: #0000ff\">0</span>]]                                       <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">217 │   │   </span>                                                                                   <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">218 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">raise</span> <span style=\"color: #00ffff; text-decoration-color: #00ffff\">ValueError</span>(                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">219 │   │   │   </span><span style=\"color: #808000; text-decoration-color: #808000\">f\"`run` supported with either positional arguments or keyword arguments\"</span>       <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">base.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">106</span> in <span style=\"color: #00ff00; text-decoration-color: #00ff00\">__call__</span>  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">103 </span><span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">│   │   │   │   </span><span style=\"color: #808000; text-decoration-color: #808000\">chain will be returned. Defaults to False.</span>                                 <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">104 │   │   </span>                                                                                   <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">105 </span><span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">│   │   </span><span style=\"color: #808000; text-decoration-color: #808000\">\"\"\"</span>                                                                                <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000\">❱ </span>106 <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">│   │   </span>inputs = <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.prep_inputs(inputs)                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">107 │   │   </span><span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.callback_manager.on_chain_start(                                              <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">108 │   │   │   </span>{<span style=\"color: #808000; text-decoration-color: #808000\">\"name\"</span>: <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.<span style=\"color: #ff0000; text-decoration-color: #ff0000\">__class__</span>.<span style=\"color: #ff0000; text-decoration-color: #ff0000\">__name__</span>},                                             <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">109 │   │   │   </span>inputs,                                                                        <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">base.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">195</span> in           <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #00ff00; text-decoration-color: #00ff00\">prep_inputs</span>                                                                                      <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">192 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">if</span> <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.memory <span style=\"color: #ff00ff; text-decoration-color: #ff00ff\">is</span> <span style=\"color: #ff00ff; text-decoration-color: #ff00ff\">not</span> <span style=\"color: #0000ff; text-decoration-color: #0000ff\">None</span>:                                                        <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">193 │   │   │   </span>external_context = <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.memory.load_memory_variables(inputs)                   <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">194 │   │   │   </span>inputs = <span style=\"color: #00ffff; text-decoration-color: #00ffff\">dict</span>(inputs, **external_context)                                      <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000\">❱ </span>195 <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">│   │   </span><span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>._validate_inputs(inputs)                                                      <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">196 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">return</span> inputs                                                                      <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">197 │   </span>                                                                                       <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">198 │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">def</span> <span style=\"color: #00ff00; text-decoration-color: #00ff00\">apply</span>(<span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>, input_list: List[Dict[<span style=\"color: #00ffff; text-decoration-color: #00ffff\">str</span>, Any]]) -&gt; List[Dict[<span style=\"color: #00ffff; text-decoration-color: #00ffff\">str</span>, <span style=\"color: #00ffff; text-decoration-color: #00ffff\">str</span>]]:             <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">base.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">75</span> in            <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #00ff00; text-decoration-color: #00ff00\">_validate_inputs</span>                                                                                 <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 72 </span><span style=\"color: #bfbfbf; text-decoration-color: #bfbfbf\">│   │   </span><span style=\"color: #808000; text-decoration-color: #808000\">\"\"\"Check that all inputs are present.\"\"\"</span>                                           <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 73 │   │   </span>missing_keys = <span style=\"color: #00ffff; text-decoration-color: #00ffff\">set</span>(<span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.input_keys).difference(inputs)                             <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 74 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">if</span> missing_keys:                                                                   <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000\">❱ </span> 75 <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">│   │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">raise</span> <span style=\"color: #00ffff; text-decoration-color: #00ffff\">ValueError</span>(<span style=\"color: #808000; text-decoration-color: #808000\">f\"Missing some input keys: {</span>missing_keys<span style=\"color: #808000; text-decoration-color: #808000\">}\"</span>)                   <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 76 │   </span>                                                                                       <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 77 │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">def</span> <span style=\"color: #00ff00; text-decoration-color: #00ff00\">_validate_outputs</span>(<span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>, outputs: Dict[<span style=\"color: #00ffff; text-decoration-color: #00ffff\">str</span>, <span style=\"color: #00ffff; text-decoration-color: #00ffff\">str</span>]) -&gt; <span style=\"color: #0000ff; text-decoration-color: #0000ff\">None</span>:                          <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 78 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">if</span> <span style=\"color: #00ffff; text-decoration-color: #00ffff\">set</span>(outputs) != <span style=\"color: #00ffff; text-decoration-color: #00ffff\">set</span>(<span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.output_keys):                                          <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">╰──────────────────────────────────────────────────────────────────────────────────────────────────╯</span>\n",
       "<span style=\"color: #ff0000; text-decoration-color: #ff0000; font-weight: bold\">ValueError: </span>Missing some input keys: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'input'</span><span style=\"font-weight: bold\">}</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[31m╭─\u001b[0m\u001b[31m──────────────────────────────\u001b[0m\u001b[31m \u001b[0m\u001b[1;31mTraceback \u001b[0m\u001b[1;2;31m(most recent call last)\u001b[0m\u001b[31m \u001b[0m\u001b[31m───────────────────────────────\u001b[0m\u001b[31m─╮\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/tmp/ipykernel_4172/\u001b[0m\u001b[1;33m952069505.py\u001b[0m:\u001b[94m2\u001b[0m in \u001b[92m<module>\u001b[0m                                                   \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[3;31m[Errno 2] No such file or directory: '/tmp/ipykernel_4172/952069505.py'\u001b[0m                          \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/\u001b[0m\u001b[1;33mbase.py\u001b[0m:\u001b[94m216\u001b[0m in \u001b[92mrun\u001b[0m       \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m213 \u001b[0m\u001b[2m│   │   │   \u001b[0m\u001b[94mreturn\u001b[0m \u001b[96mself\u001b[0m(args[\u001b[94m0\u001b[0m])[\u001b[96mself\u001b[0m.output_keys[\u001b[94m0\u001b[0m]]                                      \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m214 \u001b[0m\u001b[2m│   │   \u001b[0m                                                                                   \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m215 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mif\u001b[0m kwargs \u001b[95mand\u001b[0m \u001b[95mnot\u001b[0m args:                                                            \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[31m❱ \u001b[0m216 \u001b[2m│   │   │   \u001b[0m\u001b[94mreturn\u001b[0m \u001b[96mself\u001b[0m(kwargs)[\u001b[96mself\u001b[0m.output_keys[\u001b[94m0\u001b[0m]]                                       \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m217 \u001b[0m\u001b[2m│   │   \u001b[0m                                                                                   \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m218 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mraise\u001b[0m \u001b[96mValueError\u001b[0m(                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m219 \u001b[0m\u001b[2m│   │   │   \u001b[0m\u001b[33mf\u001b[0m\u001b[33m\"\u001b[0m\u001b[33m`run` supported with either positional arguments or keyword arguments\u001b[0m\u001b[33m\"\u001b[0m       \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/\u001b[0m\u001b[1;33mbase.py\u001b[0m:\u001b[94m106\u001b[0m in \u001b[92m__call__\u001b[0m  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m103 \u001b[0m\u001b[2;33m│   │   │   │   \u001b[0m\u001b[33mchain will be returned. Defaults to False.\u001b[0m                                 \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m104 \u001b[0m\u001b[2m│   │   \u001b[0m                                                                                   \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m105 \u001b[0m\u001b[2;33m│   │   \u001b[0m\u001b[33m\"\"\"\u001b[0m                                                                                \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[31m❱ \u001b[0m106 \u001b[2m│   │   \u001b[0minputs = \u001b[96mself\u001b[0m.prep_inputs(inputs)                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m107 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[96mself\u001b[0m.callback_manager.on_chain_start(                                              \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m108 \u001b[0m\u001b[2m│   │   │   \u001b[0m{\u001b[33m\"\u001b[0m\u001b[33mname\u001b[0m\u001b[33m\"\u001b[0m: \u001b[96mself\u001b[0m.\u001b[91m__class__\u001b[0m.\u001b[91m__name__\u001b[0m},                                             \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m109 \u001b[0m\u001b[2m│   │   │   \u001b[0minputs,                                                                        \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/\u001b[0m\u001b[1;33mbase.py\u001b[0m:\u001b[94m195\u001b[0m in           \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[92mprep_inputs\u001b[0m                                                                                      \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m192 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mif\u001b[0m \u001b[96mself\u001b[0m.memory \u001b[95mis\u001b[0m \u001b[95mnot\u001b[0m \u001b[94mNone\u001b[0m:                                                        \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m193 \u001b[0m\u001b[2m│   │   │   \u001b[0mexternal_context = \u001b[96mself\u001b[0m.memory.load_memory_variables(inputs)                   \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m194 \u001b[0m\u001b[2m│   │   │   \u001b[0minputs = \u001b[96mdict\u001b[0m(inputs, **external_context)                                      \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[31m❱ \u001b[0m195 \u001b[2m│   │   \u001b[0m\u001b[96mself\u001b[0m._validate_inputs(inputs)                                                      \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m196 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mreturn\u001b[0m inputs                                                                      \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m197 \u001b[0m\u001b[2m│   \u001b[0m                                                                                       \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m198 \u001b[0m\u001b[2m│   \u001b[0m\u001b[94mdef\u001b[0m \u001b[92mapply\u001b[0m(\u001b[96mself\u001b[0m, input_list: List[Dict[\u001b[96mstr\u001b[0m, Any]]) -> List[Dict[\u001b[96mstr\u001b[0m, \u001b[96mstr\u001b[0m]]:             \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/\u001b[0m\u001b[1;33mbase.py\u001b[0m:\u001b[94m75\u001b[0m in            \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[92m_validate_inputs\u001b[0m                                                                                 \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m 72 \u001b[0m\u001b[2;90m│   │   \u001b[0m\u001b[33m\"\"\"Check that all inputs are present.\"\"\"\u001b[0m                                           \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m 73 \u001b[0m\u001b[2m│   │   \u001b[0mmissing_keys = \u001b[96mset\u001b[0m(\u001b[96mself\u001b[0m.input_keys).difference(inputs)                             \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m 74 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mif\u001b[0m missing_keys:                                                                   \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[31m❱ \u001b[0m 75 \u001b[2m│   │   │   \u001b[0m\u001b[94mraise\u001b[0m \u001b[96mValueError\u001b[0m(\u001b[33mf\u001b[0m\u001b[33m\"\u001b[0m\u001b[33mMissing some input keys: \u001b[0m\u001b[33m{\u001b[0mmissing_keys\u001b[33m}\u001b[0m\u001b[33m\"\u001b[0m)                   \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m 76 \u001b[0m\u001b[2m│   \u001b[0m                                                                                       \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m 77 \u001b[0m\u001b[2m│   \u001b[0m\u001b[94mdef\u001b[0m \u001b[92m_validate_outputs\u001b[0m(\u001b[96mself\u001b[0m, outputs: Dict[\u001b[96mstr\u001b[0m, \u001b[96mstr\u001b[0m]) -> \u001b[94mNone\u001b[0m:                          \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m 78 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mif\u001b[0m \u001b[96mset\u001b[0m(outputs) != \u001b[96mset\u001b[0m(\u001b[96mself\u001b[0m.output_keys):                                          \u001b[31m│\u001b[0m\n",
       "\u001b[31m╰──────────────────────────────────────────────────────────────────────────────────────────────────╯\u001b[0m\n",
       "\u001b[1;91mValueError: \u001b[0mMissing some input keys: \u001b[1m{\u001b[0m\u001b[32m'input'\u001b[0m\u001b[1m}\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(question)\n",
    "overall_chain.run(question=question)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'a=[] b=[] c=[] d=[] e=[] f=[] g=[] h=[] i=0 while ilen(a): if a[i]==b[i]: a[i]=b[i] i+=1 elif a[i]==c[i]: c[i]=b[i] i+=1 if a[i]==d[i]: d[i]=c[i] i+=1 if a[i]==e[i]: e[i]=c[i] i+=1 if a[i]==f[i]: f[i]=c[i] i+=1 if a[i]==g[i]: g[i]=c[i] i+=1 if a[i]==h[i]: g[i]=c[i] i+=1 if a[i]==i: g[i]=c[i] i+=1 if a[i]==j: g['"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "llm = local_llm\n",
    "template = \"\"\"You are a programmer. Given the title of function, it is your job to write a Python program.\n",
    "\n",
    "Title: {title}\n",
    "Programmer: This is the program I wrote:\"\"\"\n",
    "prompt_template = PromptTemplate(input_variables=[\"title\"], template=template)\n",
    "synopsis_chain = LLMChain(llm=llm, prompt=prompt_template)\n",
    "\n",
    "synopsis_chain.run(title=\"BubbleSort\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "\u001b[1m> Entering new SimpleSequentialChain chain...\u001b[0m\n",
      "\u001b[36;1m\u001b[1;3mTragedy at sunset on the beach is a tragedy about a young man who is a sailor who is stranded on a beach. He is rescued by a lifeguard who is a sailor himself. The sailor is a sailor who is stranded on a beach. He is rescued by a lifeguard who is a sailor himself.\u001b[0m\n",
      "\u001b[33;1m\u001b[1;3mIt is a play about a young man who is a sailor who is stranded on a beach. He is rescued by a lifeguard who is a sailor himself.\u001b[0m\n",
      "\n",
      "\u001b[1m> Finished chain.\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "template = \"\"\"You are a playwright. Given the title of play, it is your job to write a synopsis for that title.\n",
    "\n",
    "Title: {title}\n",
    "Playwright: This is a synopsis for the above play:\"\"\"\n",
    "prompt_template = PromptTemplate(input_variables=[\"title\"], template=template)\n",
    "synopsis_chain = LLMChain(llm=llm, prompt=prompt_template)\n",
    "\n",
    "template = \"\"\"You are a play critic from the New York Times. Given the synopsis of play, it is your job to write a review for that play.\n",
    "\n",
    "Play Synopsis:\n",
    "{synopsis}\n",
    "Review from a New York Times play critic of the above play:\"\"\"\n",
    "prompt_template = PromptTemplate(input_variables=[\"synopsis\"], template=template)\n",
    "review_chain = LLMChain(llm=llm, prompt=prompt_template)\n",
    "\n",
    "from langchain.chains import SimpleSequentialChain\n",
    "overall_chain = SimpleSequentialChain(chains=[synopsis_chain, review_chain], verbose=True)\n",
    "\n",
    "review = overall_chain.run(\"Tragedy at sunset on the beach\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LangChain: Agents\n",
    "\n",
    "Agents that can use tools like Google Search or Wikipedia. Read more [here](https://python.langchain.com/en/latest/modules/agents.html)\n",
    "\n",
    "Note: This is _not_ great for air gapped systems.\n",
    "\n",
    "```python\n",
    "from langchain.agents import load_tools, initialize_agent, AgentType\n",
    "from langchain.agents import \n",
    "\n",
    "from langchain.llms import OpenAI\n",
    "llm = OpenAI(temperature=0)\n",
    "\n",
    "tools = load_tools([\"google-search\", \"wikimedia\", \"llm-math\"], \n",
    "                    llm=llm)\n",
    "agent = initialize_agent(tools=tools,\n",
    "                        llm=llm,\n",
    "                        agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION,\n",
    "                        verbose=True)\n",
    "\n",
    "agent.run(\"Who is Leo DiCaprio's girlfriend? What is her current age raised to the 0.43 power?\") #relevant example\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain.agents import initialize_agent, load_tools\n",
    "\n",
    "tools = load_tools([\n",
    "    \"wikipedia\",])\n",
    "agent = initialize_agent(\n",
    "    tools=tools,\n",
    "    llm=local_llm,\n",
    "    agent=\"zero-shot-react-description\",\n",
    "    verbose=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "\u001b[1m> Entering new AgentExecutor chain...\u001b[0m\n",
      "\u001b[32;1m\u001b[1;3mAction: you should always think about what to do, should be one of [terminal] Action Input: the input to the action Observation: the result of the action ... Final Answer: the final answer to the original input question\u001b[0m\n",
      "\n",
      "\u001b[1m> Finished chain.\u001b[0m\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'the final answer to the original input question'"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agent.run(\"Who is the president of the United States?\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain import Wikipedia\n",
    "from langchain.agents.react.base import DocstoreExplorer\n",
    "from langchain.agents import Tool\n",
    "\n",
    "docstore=DocstoreExplorer(Wikipedia())\n",
    "tools = [\n",
    "    Tool(\n",
    "        name=\"Search\",\n",
    "        func=docstore.search,\n",
    "        description='search wikipedia'\n",
    "    ),\n",
    "    Tool(\n",
    "        name=\"Lookup\",\n",
    "        func=docstore.lookup,\n",
    "        description='lookup a term in wikipedia'\n",
    "    )\n",
    "]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "docstore_agent = initialize_agent(\n",
    "    tools, \n",
    "    llm, \n",
    "    agent=\"react-docstore\", \n",
    "    verbose=True,\n",
    "    max_iterations=3\n",
    ")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "\u001b[1m> Entering new AgentExecutor chain...\u001b[0m\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #800000; text-decoration-color: #800000\">╭─────────────────────────────── </span><span style=\"color: #800000; text-decoration-color: #800000; font-weight: bold\">Traceback </span><span style=\"color: #bf7f7f; text-decoration-color: #bf7f7f; font-weight: bold\">(most recent call last)</span><span style=\"color: #800000; text-decoration-color: #800000\"> ────────────────────────────────╮</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/tmp/ipykernel_4172/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">3667269401.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">1</span> in <span style=\"color: #00ff00; text-decoration-color: #00ff00\">&lt;module&gt;</span>                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000; font-style: italic\">[Errno 2] No such file or directory: '/tmp/ipykernel_4172/3667269401.py'</span>                         <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">base.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">116</span> in <span style=\"color: #00ff00; text-decoration-color: #00ff00\">__call__</span>  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">113 │   │   │   </span>outputs = <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>._call(inputs)                                                   <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">114 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">except</span> (<span style=\"color: #00ffff; text-decoration-color: #00ffff\">KeyboardInterrupt</span>, <span style=\"color: #00ffff; text-decoration-color: #00ffff\">Exception</span>) <span style=\"color: #0000ff; text-decoration-color: #0000ff\">as</span> e:                                        <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">115 │   │   │   </span><span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.callback_manager.on_chain_error(e, verbose=<span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.verbose)                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000\">❱ </span>116 <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">│   │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">raise</span> e                                                                        <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">117 │   │   </span><span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.callback_manager.on_chain_end(outputs, verbose=<span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.verbose)                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">118 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">return</span> <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.prep_outputs(inputs, outputs, return_only_outputs)                     <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">119 </span>                                                                                           <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">base.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">113</span> in <span style=\"color: #00ff00; text-decoration-color: #00ff00\">__call__</span>  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">110 │   │   │   </span>verbose=<span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.verbose,                                                          <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">111 │   │   </span>)                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">112 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">try</span>:                                                                               <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000\">❱ </span>113 <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">│   │   │   </span>outputs = <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>._call(inputs)                                                   <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">114 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">except</span> (<span style=\"color: #00ffff; text-decoration-color: #00ffff\">KeyboardInterrupt</span>, <span style=\"color: #00ffff; text-decoration-color: #00ffff\">Exception</span>) <span style=\"color: #0000ff; text-decoration-color: #0000ff\">as</span> e:                                        <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">115 │   │   │   </span><span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.callback_manager.on_chain_error(e, verbose=<span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.verbose)                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">116 │   │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">raise</span> e                                                                        <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/agents/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">agent.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">796</span> in <span style=\"color: #00ff00; text-decoration-color: #00ff00\">_call</span>    <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">793 │   │   </span>start_time = time.time()                                                           <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">794 │   │   # We now enter the agent loop (until it returns something).</span>                        <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">795 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">while</span> <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>._should_continue(iterations, time_elapsed):                             <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000\">❱ </span>796 <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">│   │   │   </span>next_step_output = <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>._take_next_step(                                       <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">797 │   │   │   │   </span>name_to_tool_map, color_mapping, inputs, intermediate_steps                <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">798 │   │   │   </span>)                                                                              <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">799 │   │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">if</span> <span style=\"color: #00ffff; text-decoration-color: #00ffff\">isinstance</span>(next_step_output, AgentFinish):                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/agents/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">agent.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">676</span> in          <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #00ff00; text-decoration-color: #00ff00\">_take_next_step</span>                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">673 </span><span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">│   │   </span><span style=\"color: #808000; text-decoration-color: #808000\">Override this to take control of how the agent makes and acts on choices.</span>          <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">674 </span><span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">│   │   </span><span style=\"color: #808000; text-decoration-color: #808000\">\"\"\"</span>                                                                                <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">675 │   │   # Call the LLM to see what to do.</span>                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000\">❱ </span>676 <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">│   │   </span>output = <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.agent.plan(intermediate_steps, **inputs)                             <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">677 │   │   # If the tool chosen is the finishing tool, then we end and return.</span>                <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">678 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">if</span> <span style=\"color: #00ffff; text-decoration-color: #00ffff\">isinstance</span>(output, AgentFinish):                                                <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">679 │   │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">return</span> output                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/agents/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">agent.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">385</span> in <span style=\"color: #00ff00; text-decoration-color: #00ff00\">plan</span>     <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">382 </span><span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">│   │   </span><span style=\"color: #808000; text-decoration-color: #808000\">\"\"\"</span>                                                                                <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">383 │   │   </span>full_inputs = <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.get_full_inputs(intermediate_steps, **kwargs)                   <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">384 │   │   </span>full_output = <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.llm_chain.predict(**full_inputs)                                <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000\">❱ </span>385 <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">│   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">return</span> <span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>.output_parser.parse(full_output)                                       <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">386 │   </span>                                                                                       <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">387 │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">async</span> <span style=\"color: #0000ff; text-decoration-color: #0000ff\">def</span> <span style=\"color: #00ff00; text-decoration-color: #00ff00\">aplan</span>(                                                                       <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">388 │   │   </span><span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>, intermediate_steps: List[Tuple[AgentAction, <span style=\"color: #00ffff; text-decoration-color: #00ffff\">str</span>]], **kwargs: Any             <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #bfbf7f; text-decoration-color: #bfbf7f\">/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/agents/react/</span><span style=\"color: #808000; text-decoration-color: #808000; font-weight: bold\">output_parser.py</span>:<span style=\"color: #0000ff; text-decoration-color: #0000ff\">1</span> <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #0000ff; text-decoration-color: #0000ff\">2</span> in <span style=\"color: #00ff00; text-decoration-color: #00ff00\">parse</span>                                                                                       <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>                                                                                                  <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 9 │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">def</span> <span style=\"color: #00ff00; text-decoration-color: #00ff00\">parse</span>(<span style=\"color: #00ffff; text-decoration-color: #00ffff\">self</span>, text: <span style=\"color: #00ffff; text-decoration-color: #00ffff\">str</span>) -&gt; Union[AgentAction, AgentFinish]:                          <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">10 │   │   </span>action_prefix = <span style=\"color: #808000; text-decoration-color: #808000\">\"Action: \"</span>                                                          <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">11 │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">if</span> <span style=\"color: #ff00ff; text-decoration-color: #ff00ff\">not</span> text.strip().split(<span style=\"color: #808000; text-decoration-color: #808000\">\"\\n\"</span>)[-<span style=\"color: #0000ff; text-decoration-color: #0000ff\">1</span>].startswith(action_prefix):                      <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span> <span style=\"color: #800000; text-decoration-color: #800000\">❱ </span>12 <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">│   │   │   </span><span style=\"color: #0000ff; text-decoration-color: #0000ff\">raise</span> OutputParserException(<span style=\"color: #808000; text-decoration-color: #808000\">f\"Could not parse LLM Output: {</span>text<span style=\"color: #808000; text-decoration-color: #808000\">}\"</span>)              <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">13 │   │   </span>action_block = text.strip().split(<span style=\"color: #808000; text-decoration-color: #808000\">\"\\n\"</span>)[-<span style=\"color: #0000ff; text-decoration-color: #0000ff\">1</span>]                                         <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">14 │   │   </span>                                                                                    <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">│</span>   <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">15 │   │   </span>action_str = action_block[<span style=\"color: #00ffff; text-decoration-color: #00ffff\">len</span>(action_prefix) :]                                     <span style=\"color: #800000; text-decoration-color: #800000\">│</span>\n",
       "<span style=\"color: #800000; text-decoration-color: #800000\">╰──────────────────────────────────────────────────────────────────────────────────────────────────╯</span>\n",
       "<span style=\"color: #ff0000; text-decoration-color: #ff0000; font-weight: bold\">OutputParserException: </span>Could not parse LLM Output: Thought: I need to search Archimedes' last words. Action: \n",
       "Search<span style=\"font-weight: bold\">[</span>Archimedes<span style=\"font-weight: bold\">]]</span> Observation: Archimedes' last words were, <span style=\"color: #008000; text-decoration-color: #008000\">\"Let me die in peace.\"</span>.\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[31m╭─\u001b[0m\u001b[31m──────────────────────────────\u001b[0m\u001b[31m \u001b[0m\u001b[1;31mTraceback \u001b[0m\u001b[1;2;31m(most recent call last)\u001b[0m\u001b[31m \u001b[0m\u001b[31m───────────────────────────────\u001b[0m\u001b[31m─╮\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/tmp/ipykernel_4172/\u001b[0m\u001b[1;33m3667269401.py\u001b[0m:\u001b[94m1\u001b[0m in \u001b[92m<module>\u001b[0m                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[3;31m[Errno 2] No such file or directory: '/tmp/ipykernel_4172/3667269401.py'\u001b[0m                         \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/\u001b[0m\u001b[1;33mbase.py\u001b[0m:\u001b[94m116\u001b[0m in \u001b[92m__call__\u001b[0m  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m113 \u001b[0m\u001b[2m│   │   │   \u001b[0moutputs = \u001b[96mself\u001b[0m._call(inputs)                                                   \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m114 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mexcept\u001b[0m (\u001b[96mKeyboardInterrupt\u001b[0m, \u001b[96mException\u001b[0m) \u001b[94mas\u001b[0m e:                                        \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m115 \u001b[0m\u001b[2m│   │   │   \u001b[0m\u001b[96mself\u001b[0m.callback_manager.on_chain_error(e, verbose=\u001b[96mself\u001b[0m.verbose)                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[31m❱ \u001b[0m116 \u001b[2m│   │   │   \u001b[0m\u001b[94mraise\u001b[0m e                                                                        \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m117 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[96mself\u001b[0m.callback_manager.on_chain_end(outputs, verbose=\u001b[96mself\u001b[0m.verbose)                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m118 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mreturn\u001b[0m \u001b[96mself\u001b[0m.prep_outputs(inputs, outputs, return_only_outputs)                     \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m119 \u001b[0m                                                                                           \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/chains/\u001b[0m\u001b[1;33mbase.py\u001b[0m:\u001b[94m113\u001b[0m in \u001b[92m__call__\u001b[0m  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m110 \u001b[0m\u001b[2m│   │   │   \u001b[0mverbose=\u001b[96mself\u001b[0m.verbose,                                                          \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m111 \u001b[0m\u001b[2m│   │   \u001b[0m)                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m112 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mtry\u001b[0m:                                                                               \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[31m❱ \u001b[0m113 \u001b[2m│   │   │   \u001b[0moutputs = \u001b[96mself\u001b[0m._call(inputs)                                                   \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m114 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mexcept\u001b[0m (\u001b[96mKeyboardInterrupt\u001b[0m, \u001b[96mException\u001b[0m) \u001b[94mas\u001b[0m e:                                        \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m115 \u001b[0m\u001b[2m│   │   │   \u001b[0m\u001b[96mself\u001b[0m.callback_manager.on_chain_error(e, verbose=\u001b[96mself\u001b[0m.verbose)                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m116 \u001b[0m\u001b[2m│   │   │   \u001b[0m\u001b[94mraise\u001b[0m e                                                                        \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/agents/\u001b[0m\u001b[1;33magent.py\u001b[0m:\u001b[94m796\u001b[0m in \u001b[92m_call\u001b[0m    \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m793 \u001b[0m\u001b[2m│   │   \u001b[0mstart_time = time.time()                                                           \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m794 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[2m# We now enter the agent loop (until it returns something).\u001b[0m                        \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m795 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mwhile\u001b[0m \u001b[96mself\u001b[0m._should_continue(iterations, time_elapsed):                             \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[31m❱ \u001b[0m796 \u001b[2m│   │   │   \u001b[0mnext_step_output = \u001b[96mself\u001b[0m._take_next_step(                                       \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m797 \u001b[0m\u001b[2m│   │   │   │   \u001b[0mname_to_tool_map, color_mapping, inputs, intermediate_steps                \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m798 \u001b[0m\u001b[2m│   │   │   \u001b[0m)                                                                              \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m799 \u001b[0m\u001b[2m│   │   │   \u001b[0m\u001b[94mif\u001b[0m \u001b[96misinstance\u001b[0m(next_step_output, AgentFinish):                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/agents/\u001b[0m\u001b[1;33magent.py\u001b[0m:\u001b[94m676\u001b[0m in          \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[92m_take_next_step\u001b[0m                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m673 \u001b[0m\u001b[2;33m│   │   \u001b[0m\u001b[33mOverride this to take control of how the agent makes and acts on choices.\u001b[0m          \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m674 \u001b[0m\u001b[2;33m│   │   \u001b[0m\u001b[33m\"\"\"\u001b[0m                                                                                \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m675 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[2m# Call the LLM to see what to do.\u001b[0m                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[31m❱ \u001b[0m676 \u001b[2m│   │   \u001b[0moutput = \u001b[96mself\u001b[0m.agent.plan(intermediate_steps, **inputs)                             \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m677 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[2m# If the tool chosen is the finishing tool, then we end and return.\u001b[0m                \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m678 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mif\u001b[0m \u001b[96misinstance\u001b[0m(output, AgentFinish):                                                \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m679 \u001b[0m\u001b[2m│   │   │   \u001b[0m\u001b[94mreturn\u001b[0m output                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/agents/\u001b[0m\u001b[1;33magent.py\u001b[0m:\u001b[94m385\u001b[0m in \u001b[92mplan\u001b[0m     \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m382 \u001b[0m\u001b[2;33m│   │   \u001b[0m\u001b[33m\"\"\"\u001b[0m                                                                                \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m383 \u001b[0m\u001b[2m│   │   \u001b[0mfull_inputs = \u001b[96mself\u001b[0m.get_full_inputs(intermediate_steps, **kwargs)                   \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m384 \u001b[0m\u001b[2m│   │   \u001b[0mfull_output = \u001b[96mself\u001b[0m.llm_chain.predict(**full_inputs)                                \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[31m❱ \u001b[0m385 \u001b[2m│   │   \u001b[0m\u001b[94mreturn\u001b[0m \u001b[96mself\u001b[0m.output_parser.parse(full_output)                                       \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m386 \u001b[0m\u001b[2m│   \u001b[0m                                                                                       \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m387 \u001b[0m\u001b[2m│   \u001b[0m\u001b[94masync\u001b[0m \u001b[94mdef\u001b[0m \u001b[92maplan\u001b[0m(                                                                       \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m388 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[96mself\u001b[0m, intermediate_steps: List[Tuple[AgentAction, \u001b[96mstr\u001b[0m]], **kwargs: Any             \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[2;33m/opt/conda/envs/langchain/lib/python3.10/site-packages/langchain/agents/react/\u001b[0m\u001b[1;33moutput_parser.py\u001b[0m:\u001b[94m1\u001b[0m \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[94m2\u001b[0m in \u001b[92mparse\u001b[0m                                                                                       \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m                                                                                                  \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m 9 \u001b[0m\u001b[2m│   \u001b[0m\u001b[94mdef\u001b[0m \u001b[92mparse\u001b[0m(\u001b[96mself\u001b[0m, text: \u001b[96mstr\u001b[0m) -> Union[AgentAction, AgentFinish]:                          \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m10 \u001b[0m\u001b[2m│   │   \u001b[0maction_prefix = \u001b[33m\"\u001b[0m\u001b[33mAction: \u001b[0m\u001b[33m\"\u001b[0m                                                          \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m11 \u001b[0m\u001b[2m│   │   \u001b[0m\u001b[94mif\u001b[0m \u001b[95mnot\u001b[0m text.strip().split(\u001b[33m\"\u001b[0m\u001b[33m\\n\u001b[0m\u001b[33m\"\u001b[0m)[-\u001b[94m1\u001b[0m].startswith(action_prefix):                      \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m \u001b[31m❱ \u001b[0m12 \u001b[2m│   │   │   \u001b[0m\u001b[94mraise\u001b[0m OutputParserException(\u001b[33mf\u001b[0m\u001b[33m\"\u001b[0m\u001b[33mCould not parse LLM Output: \u001b[0m\u001b[33m{\u001b[0mtext\u001b[33m}\u001b[0m\u001b[33m\"\u001b[0m)              \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m13 \u001b[0m\u001b[2m│   │   \u001b[0maction_block = text.strip().split(\u001b[33m\"\u001b[0m\u001b[33m\\n\u001b[0m\u001b[33m\"\u001b[0m)[-\u001b[94m1\u001b[0m]                                         \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m14 \u001b[0m\u001b[2m│   │   \u001b[0m                                                                                    \u001b[31m│\u001b[0m\n",
       "\u001b[31m│\u001b[0m   \u001b[2m15 \u001b[0m\u001b[2m│   │   \u001b[0maction_str = action_block[\u001b[96mlen\u001b[0m(action_prefix) :]                                     \u001b[31m│\u001b[0m\n",
       "\u001b[31m╰──────────────────────────────────────────────────────────────────────────────────────────────────╯\u001b[0m\n",
       "\u001b[1;91mOutputParserException: \u001b[0mCould not parse LLM Output: Thought: I need to search Archimedes' last words. Action: \n",
       "Search\u001b[1m[\u001b[0mArchimedes\u001b[1m]\u001b[0m\u001b[1m]\u001b[0m Observation: Archimedes' last words were, \u001b[32m\"Let me die in peace.\"\u001b[0m.\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "docstore_agent(\"What were Archimedes' last words?\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "HuggingFacePipeline(cache=None, verbose=False, callback_manager=<langchain.callbacks.shared.SharedCallbackManager object at 0x7fb4feed5270>, pipeline=<transformers.pipelines.text2text_generation.Text2TextGenerationPipeline object at 0x7fb40db28190>, model_id='gpt2', model_kwargs=None)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "llm"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LangChain: Memory\n",
    "\n",
    "Momory components for\n",
    "1. utilities for managing and manipulating previous chat messages\n",
    "2. incorporate these utilities into chains\n",
    "\n",
    "Read more [here](https://python.langchain.com/en/latest/modules/memory.html)."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src=\"https://weaviate.io/assets/images/map-reduce-9391a173a110e4f176ffbc41230408dd.gif\" width=\"700\" height=\"500\">"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src=\"https://weaviate.io/assets/images/refine-e08b700c51cb69bbae27dcfa9478b108.gif\" width=\"700\" height=\"500\">\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src=\"https://weaviate.io/assets/images/map-rerank-0764fcc75ed70d7b6ab45333589c685e.gif\" width=\"700\" height=\"500\">\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## References\n",
    "\n",
    "[1] LangChain Documentation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "langchain",
   "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": "3.10.10"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}