Tutorial on how to use Streamlit, Langchain,and SerpAPI to search the web with AI

langchain_serpapi_web_search_tutorial.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "b3f7dbe3",
   "metadata": {},
   "source": [
    "\n",
    "# LangChain Search with Streamlit\n",
    "\n",
    "In this tutorial, we will guide you through the process of setting up and running a Streamlit app for LangChain Search. The app allows users to input their OpenAI and SERP API keys, enter a search query, and retrieve results using LangChain's capabilities.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "821d332d",
   "metadata": {},
   "source": [
    "\n",
    "## Installation of Required Packages\n",
    "\n",
    "Before we begin, let's ensure we have all the necessary packages installed.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14c0a9dc",
   "metadata": {},
   "source": [
    "\n",
    "## Creating the Streamlit App\n",
    "\n",
    "We'll define our Streamlit app and save the code to a local file named `app.py`.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c04340d9",
   "metadata": {},
   "source": [
    "\n",
    "## Setting Up LocalTunnel\n",
    "\n",
    "To make our Streamlit app accessible over the internet, we'll set up LocalTunnel.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c2827d1e",
   "metadata": {},
   "source": [
    "\n",
    "## Running the Streamlit App\n",
    "\n",
    "Let's run our Streamlit app in the background. It will start serving on the default port, 8501.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2c2d8496",
   "metadata": {},
   "source": [
    "\n",
    "## Exposing the App\n",
    "\n",
    "Finally, we'll expose our Streamlit app to the internet using LocalTunnel on port 8501.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6121b9f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "!pip install streamlit\n",
    "!pip install langchain\n",
    "!pip install openai\n",
    "!pip install google-search-results\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ce583d11",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "%%writefile app.py\n",
    "import streamlit as st\n",
    "from langchain.llms.openai import OpenAI\n",
    "from langchain.agents import load_tools, initialize_agent\n",
    "\n",
    "# Streamlit app\n",
    "st.title('LangChain Search')\n",
    "\n",
    "# Get OpenAI API key, SERP API key and search query\n",
    "openai_api_key = st.text_input(\"OpenAI API Key\", type=\"password\")\n",
    "serpapi_api_key = st.text_input(\"SERP API Key\", type=\"password\")\n",
    "search_query = st.text_input(\"Search Query\")\n",
    "\n",
    "# Check if the 'Search' button is clicked\n",
    "if st.button(\"Search\"):\n",
    "    # Validate inputs\n",
    "    if not openai_api_key.strip() or not serpapi_api_key.strip() or not search_query.strip():\n",
    "        st.write(f\"Please provide the missing fields.\")\n",
    "    else:\n",
    "        try:\n",
    "            # Initialize the OpenAI module, load the SerpApi tool, and run the search query using an agent\n",
    "            llm=OpenAI(temperature=0, openai_api_key=openai_api_key, verbose=True)\n",
    "            tools = load_tools([\"serpapi\"], llm, serpapi_api_key=serpapi_api_key)\n",
    "            agent = initialize_agent(tools, llm, agent=\"zero-shot-react-description\", verbose=True)\n",
    "\n",
    "            result = agent.run(search_query)\n",
    "            \n",
    "            st.write(result)\n",
    "        except Exception as e:\n",
    "            st.write(f\"An error occurred: {e}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7c9da7f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "!npm install localtunnel\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "39f00d04",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "!streamlit run app.py &>/content/logs.txt &\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f89ec071",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "!npx localtunnel --port 8501\n"
   ]
  }
 ],
 "metadata": {},
 "nbformat": 4,
 "nbformat_minor": 5
}