{ "cells": [ { "cell_type": "markdown", "id": "75e2ef28-594f-4c18-9d22-c6b8cd40ead2", "metadata": {}, "source": [ "# Day 3 - Conversational AI - aka Chatbot!" ] }, { "cell_type": "code", "execution_count": 1, "id": "70e39cd8-ec79-4e3e-9c26-5659d42d0861", "metadata": {}, "outputs": [], "source": [ "# imports\n", "\n", "import os\n", "from dotenv import load_dotenv\n", "from openai import OpenAI\n", "import gradio as gr" ] }, { "cell_type": "code", "execution_count": null, "id": "231605aa-fccb-447e-89cf-8b187444536a", "metadata": {}, "outputs": [], "source": [ "# Load environment variables in a file called .env\n", "# Print the key prefixes to help with any debugging\n", "\n", "load_dotenv()\n", "openai_api_key = os.getenv('OPENAI_API_KEY')\n", "anthropic_api_key = os.getenv('ANTHROPIC_API_KEY')\n", "google_api_key = os.getenv('GOOGLE_API_KEY')\n", "\n", "if openai_api_key:\n", " print(f\"OpenAI API Key exists and begins {openai_api_key[:8]}\")\n", "else:\n", " print(\"OpenAI API Key not set\")\n", " \n", "if anthropic_api_key:\n", " print(f\"Anthropic API Key exists and begins {anthropic_api_key[:7]}\")\n", "else:\n", " print(\"Anthropic API Key not set\")\n", "\n", "if google_api_key:\n", " print(f\"Google API Key exists and begins {google_api_key[:8]}\")\n", "else:\n", " print(\"Google API Key not set\")" ] }, { "cell_type": "code", "execution_count": 3, "id": "6541d58e-2297-4de1-b1f7-77da1b98b8bb", "metadata": {}, "outputs": [], "source": [ "# Initialize\n", "\n", "openai = OpenAI()\n", "MODEL = 'gpt-4o-mini'" ] }, { "cell_type": "code", "execution_count": 4, "id": "e16839b5-c03b-4d9d-add6-87a0f6f37575", "metadata": {}, "outputs": [], "source": [ "system_message = \"You are a helpful assistant\"" ] }, { "cell_type": "markdown", "id": "98e97227-f162-4d1a-a0b2-345ff248cbe7", "metadata": {}, "source": [ "# Please read this! A change from the video:\n", "\n", "In the video, I explain how we now need to write a function called:\n", "\n", "`chat(message, history)`\n", "\n", "Which expects to receive `history` in a particular format, which we need to map to the OpenAI format before we call OpenAI:\n", "\n", "```\n", "[\n", " {\"role\": \"system\", \"content\": \"system message here\"},\n", " {\"role\": \"user\", \"content\": \"first user prompt here\"},\n", " {\"role\": \"assistant\", \"content\": \"the assistant's response\"},\n", " {\"role\": \"user\", \"content\": \"the new user prompt\"},\n", "]\n", "```\n", "\n", "But Gradio has been upgraded! Now it will pass in `history` in the exact OpenAI format, perfect for us to send straight to OpenAI.\n", "\n", "So our work just got easier!\n", "\n", "We will write a function `chat(message, history)` where: \n", "**message** is the prompt to use \n", "**history** is the past conversation, in OpenAI format \n", "\n", "We will combine the system message, history and latest message, then call OpenAI." ] }, { "cell_type": "code", "execution_count": 5, "id": "1eacc8a4-4b48-4358-9e06-ce0020041bc1", "metadata": {}, "outputs": [], "source": [ "# It's now just 1 line of code to prepare the input to OpenAI!\n", "\n", "def chat(message, history):\n", " messages = [{\"role\": \"system\", \"content\": system_message}] + history + [{\"role\": \"user\", \"content\": message}]\n", "\n", " print(\"History is:\")\n", " print(history)\n", " print(\"And messages is:\")\n", " print(messages)\n", "\n", " stream = openai.chat.completions.create(model=MODEL, messages=messages, stream=True)\n", "\n", " response = \"\"\n", " for chunk in stream:\n", " response += chunk.choices[0].delta.content or ''\n", " yield response" ] }, { "cell_type": "markdown", "id": "1334422a-808f-4147-9c4c-57d63d9780d0", "metadata": {}, "source": [ "## And then enter Gradio's magic!" ] }, { "cell_type": "code", "execution_count": null, "id": "0866ca56-100a-44ab-8bd0-1568feaf6bf2", "metadata": {}, "outputs": [], "source": [ "gr.ChatInterface(fn=chat, type=\"messages\").launch()" ] }, { "cell_type": "code", "execution_count": 7, "id": "1f91b414-8bab-472d-b9c9-3fa51259bdfe", "metadata": {}, "outputs": [], "source": [ "system_message = \"You are a helpful assistant in a clothes store. You should try to gently encourage \\\n", "the customer to try items that are on sale. Hats are 60% off, and most other items are 50% off. \\\n", "For example, if the customer says 'I'm looking to buy a hat', \\\n", "you could reply something like, 'Wonderful - we have lots of hats - including several that are part of our sales event.'\\\n", "Encourage the customer to buy hats if they are unsure what to get.\"" ] }, { "cell_type": "code", "execution_count": 8, "id": "4e5be3ec-c26c-42bc-ac16-c39d369883f6", "metadata": {}, "outputs": [], "source": [ "def chat(message, history):\n", " messages = [{\"role\": \"system\", \"content\": system_message}] + history + [{\"role\": \"user\", \"content\": message}]\n", "\n", " stream = openai.chat.completions.create(model=MODEL, messages=messages, stream=True)\n", "\n", " response = \"\"\n", " for chunk in stream:\n", " response += chunk.choices[0].delta.content or ''\n", " yield response" ] }, { "cell_type": "code", "execution_count": null, "id": "413e9e4e-7836-43ac-a0c3-e1ab5ed6b136", "metadata": {}, "outputs": [], "source": [ "gr.ChatInterface(fn=chat, type=\"messages\").launch()" ] }, { "cell_type": "code", "execution_count": 10, "id": "d75f0ffa-55c8-4152-b451-945021676837", "metadata": {}, "outputs": [], "source": [ "system_message += \"\\nIf the customer asks for shoes, you should respond that shoes are not on sale today, \\\n", "but remind the customer to look at hats!\"" ] }, { "cell_type": "code", "execution_count": null, "id": "c602a8dd-2df7-4eb7-b539-4e01865a6351", "metadata": {}, "outputs": [], "source": [ "gr.ChatInterface(fn=chat, type=\"messages\").launch()" ] }, { "cell_type": "code", "execution_count": 14, "id": "ce43fe80", "metadata": {}, "outputs": [], "source": [ "ollama_via_openai = OpenAI(base_url='http://localhost:11434/v1', api_key='ollama')" ] }, { "cell_type": "code", "execution_count": 15, "id": "4f28e3a8", "metadata": {}, "outputs": [], "source": [ "ollama_system_prompt = \"\"\"You assistant only to refine user query like check grammatical, capital, lower case that make sophisticated prompt.\"\"\"" ] }, { "cell_type": "code", "execution_count": 16, "id": "667a7fbb", "metadata": {}, "outputs": [], "source": [ "ollama_user_prompt = \"\"\"You assistant like RAG technology use to add more informaiton about the content in the user query. \n", "Please look at some content as following we don't have in our store:\n", "pants\n", "sunglasses\n", "watch\n", "underwear\n", "If you find this items in user query, answer gently like The store does not sell like 'e.g. pants'; if they are asked for 'pants', be sure to point out other items on sale.\n", "\"\"\"" ] }, { "cell_type": "code", "execution_count": 17, "id": "4632f16b", "metadata": {}, "outputs": [], "source": [ "user_messages = [{\"role\": \"system\", \"content\": ollama_system_prompt}, {\"role\": \"user\", \"content\": ollama_user_prompt},]" ] }, { "cell_type": "code", "execution_count": 23, "id": "0a987a66-1061-46d6-a83a-a30859dc88bf", "metadata": {}, "outputs": [], "source": [ "# Fixed a bug in this function brilliantly identified by student Gabor M.!\n", "# I've also improved the structure of this function\n", "\n", "def chat(message, history):\n", " relevant_system_message = system_message\n", "\n", " # Refine the user query\n", " try:\n", " refine_query = ollama_via_openai.chat.completions.create(model='llama3.2', messages=user_messages)\n", " refined_content = refine_query.choices[0].message.content\n", " except Exception as e:\n", " print(f\"Error in refinement: {e}\")\n", " refined_content = \"Error in refining query.\"\n", "\n", " # Log the original and refined queries\n", " # print(f\"Original User Query: {message}\")\n", " # print(f\"Refined Query: {refined_content}\")\n", " # print(\"============================== END of REFINE CODE ===================================\")\n", "\n", " relevant_system_message += refined_content\n", " \n", " messages = [{\"role\": \"system\", \"content\": relevant_system_message}] + history + [{\"role\": \"user\", \"content\": message}]\n", "\n", " stream = openai.chat.completions.create(model=MODEL, messages=messages, stream=True)\n", "\n", " response = \"\"\n", " for chunk in stream:\n", " response += chunk.choices[0].delta.content or ''\n", " yield response" ] }, { "cell_type": "code", "execution_count": null, "id": "20570de2-eaad-42cc-a92c-c779d71b48b6", "metadata": {}, "outputs": [], "source": [ "gr.ChatInterface(fn=chat, type=\"messages\").launch()" ] }, { "cell_type": "markdown", "id": "82a57ee0-b945-48a7-a024-01b56a5d4b3e", "metadata": {}, "source": [ "\n", " \n", " \n", " \n", " \n", "
\n", " \n", " \n", "

Business Applications

\n", " Conversational Assistants are of course a hugely common use case for Gen AI, and the latest frontier models are remarkably good at nuanced conversation. And Gradio makes it easy to have a user interface. Another crucial skill we covered is how to use prompting to provide context, information and examples.\n", "

\n", "Consider how you could apply an AI Assistant to your business, and make yourself a prototype. Use the system prompt to give context on your business, and set the tone for the LLM.
\n", "
" ] }, { "cell_type": "code", "execution_count": null, "id": "6dfb9e21-df67-4c2b-b952-5e7e7961b03d", "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "llm_env", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.11.9" } }, "nbformat": 4, "nbformat_minor": 5 }