Pink Programming Chatbot (Using Flask + Neo4j)

Welcome to the Pink Programming Chatbot workshop! Today we will build a chatbot that stores users, chats, and messages in Neo4j. Your goal is to complete the Cypher TODOs in db.py to get a working chatbot. Good Luck!

Overview

• This code base is a Flask web app scaffold for a chatbot that stores users, chats, and messages in Neo4j.
• Your goal is to complete the Cypher TODOs in db.py to enable user auth, chat creation, and viewing history.
• Use OpenAI through the Cypher GenAI Plugin integration.

Quick Start with Codespaces

• Create an account on https://console.neo4j.io/ and create a free Aura instance (make sure to save the password).
• Important: Connect to the database (click query) and run the command: ALTER DATABASE <db_name> SET DEFAULT LANGUAGE CYPHER 25 in order to get the correct Cypher version.

• In the GitHub repository click the green "Use this template" button and open the project in a codespace.

• Copy .env.example to .env and set values.
  • NEO4J_USERNAME: This should be the same name as your Aura instance.
  • NEO4J_CONNECTION_URI: Inspect your Aura instance and copy the connection URI.
  • NEO4J_PASSWORD: paste password that you got during the instance creation.

• Open the Run and Debug menu and start the "Flask: run app.py" configuration.

• It opens a terminal and starts the Flask development server, click on the link to open the app in a new browser tab.

• Try registering a new user, you should see Cypher Error page which means you're ready to start the exercises!

Workshop Tasks (Cypher TODOs in db.py)

• Create/Login User
  • create_user(username, password_hash): create a :User with properties and return it.
  • fetch_user_by_username(username): find a :User by username and return it.
• Chat With Chatbot
  • create_chat_and_first_message(username, message): create a :Chat and the first user :Message.
  • create_message(chat_id, role, content, previous_ai_chat_id): append messages (both user and ai) to a chat.
• See Previous Chats
  • list_user_chats(username): get all chats for the user, return a summary in a list.
  • fetch_chat(chat_id): return full chat with messages ordered by time.
  • get_ai_response(chat_id, previous_chat_id, message): get the response from the OpenAI API.
  • get_previous_ai_chat_id(chat_id): find the previous AI chat id for a given chat id to send to OpenAI.

Pages

• /register and /login: basic authentication using Flask sessions.
• /chat/new: start a new chat by sending the first message.
• /chats: list previous chats.
• /chats/<chat_id>: view and continue a chat.

Clean Database

You may want to reset your database to start fresh, do this in the Aura console and run "MATCH (n) DETACH DELETE n"

Database Model

Possible Extensions

• Add a delete chat option
• Add a delete User option