langchain/docs/extras/integrations/toolkits/csv.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "7094e328",
   "metadata": {},
   "source": [
    "# CSV Agent\n",
    "\n",
    "This notebook shows how to use agents to interact with a csv. It is mostly optimized for question answering.\n",
    "\n",
    "**NOTE: this agent calls the Pandas DataFrame agent under the hood, which in turn calls the Python agent, which executes LLM generated Python code - this can be bad if the LLM generated Python code is harmful. Use cautiously.**\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "827982c7",
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain.agents import create_csv_agent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "caae0bec",
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain.llms import OpenAI\n",
    "from langchain.chat_models import ChatOpenAI\n",
    "from langchain.agents.agent_types import AgentType"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bd806175",
   "metadata": {},
   "source": [
    "## Using ZERO_SHOT_REACT_DESCRIPTION\n",
    "\n",
    "This shows how to initialize the agent using the ZERO_SHOT_REACT_DESCRIPTION agent type. Note that this is an alternative to the above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a1717204",
   "metadata": {},
   "outputs": [],
   "source": [
    "agent = create_csv_agent(\n",
    "    OpenAI(temperature=0),\n",
    "    \"titanic.csv\",\n",
    "    verbose=True,\n",
    "    agent_type=AgentType.ZERO_SHOT_REACT_DESCRIPTION,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c31bb8a6",
   "metadata": {},
   "source": [
    "## Using OpenAI Functions\n",
    "\n",
    "This shows how to initialize the agent using the OPENAI_FUNCTIONS agent type. Note that this is an alternative to the above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "16c4dc59",
   "metadata": {},
   "outputs": [],
   "source": [
    "agent = create_csv_agent(\n",
    "    ChatOpenAI(temperature=0, model=\"gpt-3.5-turbo-0613\"),\n",
    "    \"titanic.csv\",\n",
    "    verbose=True,\n",
    "    agent_type=AgentType.OPENAI_FUNCTIONS,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "46b9489d",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Error in on_chain_start callback: 'name'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[32;1m\u001b[1;3m\n",
      "Invoking: `python_repl_ast` with `df.shape[0]`\n",
      "\n",
      "\n",
      "\u001b[0m\u001b[36;1m\u001b[1;3m891\u001b[0m\u001b[32;1m\u001b[1;3mThere are 891 rows in the dataframe.\u001b[0m\n",
      "\n",
      "\u001b[1m> Finished chain.\u001b[0m\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'There are 891 rows in the dataframe.'"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agent.run(\"how many rows are there?\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a96309be",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Error in on_chain_start callback: 'name'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[32;1m\u001b[1;3m\n",
      "Invoking: `python_repl_ast` with `df[df['SibSp'] > 3]['PassengerId'].count()`\n",
      "\n",
      "\n",
      "\u001b[0m\u001b[36;1m\u001b[1;3m30\u001b[0m\u001b[32;1m\u001b[1;3mThere are 30 people in the dataframe who have more than 3 siblings.\u001b[0m\n",
      "\n",
      "\u001b[1m> Finished chain.\u001b[0m\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'There are 30 people in the dataframe who have more than 3 siblings.'"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agent.run(\"how many people have more than 3 siblings\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "964a09f7",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Error in on_chain_start callback: 'name'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[32;1m\u001b[1;3m\n",
      "Invoking: `python_repl_ast` with `import pandas as pd\n",
      "import math\n",
      "\n",
      "# Create a dataframe\n",
      "data = {'Age': [22, 38, 26, 35, 35]}\n",
      "df = pd.DataFrame(data)\n",
      "\n",
      "# Calculate the average age\n",
      "average_age = df['Age'].mean()\n",
      "\n",
      "# Calculate the square root of the average age\n",
      "square_root = math.sqrt(average_age)\n",
      "\n",
      "square_root`\n",
      "\n",
      "\n",
      "\u001b[0m\u001b[36;1m\u001b[1;3m5.585696017507576\u001b[0m\u001b[32;1m\u001b[1;3mThe square root of the average age is approximately 5.59.\u001b[0m\n",
      "\n",
      "\u001b[1m> Finished chain.\u001b[0m\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'The square root of the average age is approximately 5.59.'"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agent.run(\"whats the square root of the average age?\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "09539c18",
   "metadata": {},
   "source": [
    "### Multi CSV Example\n",
    "\n",
    "This next part shows how the agent can interact with multiple csv files passed in as a list."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "15f11fbd",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Error in on_chain_start callback: 'name'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[32;1m\u001b[1;3m\n",
      "Invoking: `python_repl_ast` with `df1['Age'].nunique() - df2['Age'].nunique()`\n",
      "\n",
      "\n",
      "\u001b[0m\u001b[36;1m\u001b[1;3m-1\u001b[0m\u001b[32;1m\u001b[1;3mThere is 1 row in the age column that is different between the two dataframes.\u001b[0m\n",
      "\n",
      "\u001b[1m> Finished chain.\u001b[0m\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'There is 1 row in the age column that is different between the two dataframes.'"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agent = create_csv_agent(\n",
    "    ChatOpenAI(temperature=0, model=\"gpt-3.5-turbo-0613\"),\n",
    "    [\"titanic.csv\", \"titanic_age_fillna.csv\"],\n",
    "    verbose=True,\n",
    "    agent_type=AgentType.OPENAI_FUNCTIONS,\n",
    ")\n",
    "agent.run(\"how many rows in the age column are different between the two dfs?\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f2909808",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}