CLI Chat with Python

Posted on 28 July 2023.

learn practice programming ai cli

About

In this project we will create a very simple CLI application that we can use to have a conversation with ChatGPT. We will be using Python as the programming language. We will use the openai library.

NOTE: This project will require you to have a paid openai account. This means that you will an API KEY that you can find on your openai account. Then you will have to run the command

export OPENAI_API_KEY=sk-...

Walkthrough

Setup

I usually like to setup my Python projects by using venv, a tool that allows you to create lightweight virtual environments.

First, create a folder for the project

mkdir chatgpt-cli

Then enter the directory and create a virtual environment

cd chatgpt-cli
python -m venv .venv

Next you will have to activate the environment, depending on your shell you will have to use one of the following commands:

1. bash -> source .venv/bin/activate
2. zsh -> source .venv/bin/activate
3. fish -> source .venv/bin/activate.fish

Now that we have a virtual environment we can install the openai package by using

pip install openai

First CLI

With all the dependencies installed we can start working on our CLI application. First create a main.py file and then open it in your code editor.

touch main.py

Next step is to import the dependencies using

import openai

Next let’s setup a system prompt that we can use with ChatGPT.

SYSTEM = "You are a helpful AI assistant that can answer questions provided by the user"

The next step is to take some input from the user with input

prompt = input()

Finally we can create the completion object and get the result back

chat_completion = openai.ChatCompletion.create(
    model="gpt-3.5-turbo",
    messages=[
        {"role": "system", "content": SYSTEM},
        {"role": "user", "content": prompt}
    ]
)

this will create a completion that will store as history the system prompt that we have created and the message from the user. And then it will send it to the API. We get back a JSON response in the form of a dictionary that we can display

print(chat_completion.choices[0].message.content)

And next we can use the script from terminal with python main.py to see the results. You have to enter a prompt and then press enter, after a bit of time you will see the response.

Improve CLI

We can make the initial prompt fancier by using something like

input("User: ")

and for the print statement we can use

print(f"GPT: {chat_completion.choices[0].message.content)}")

However we can only pass a single message into the application. Well, we can use a list to keep the history of the conversation between us and the chatbot.

First we can add a list with the initial prompt after we create the SYSTEM variable

HISTORY = [{"role": "system", "content": SYSTEM}]

The next step requires to modify the script such that we have an infinite loop waiting for user input, adds the user input in the history and then adds the ChatGPT response in the history (an important step that sometimes you can forget about).

while True:
    prompt = input("User: ")

    HISTORY.append({"role": "user", "content": prompt})

    chat_completion = openai.ChatCompletion.create(
        model="gpt-3.5-turbo",
        messages=HISTORY,
    )

    response = chat_completion.choices[0].message.content

    HISTORY.append({"role": "assistant", "content": response})

    print(f"GPT: {response}")

The other change than before is that instead of the hardcoded list we also pass as messages the HISTORY list that we created. This way we can keep track of all the previous messages.

we can also add a keyword that stops the application, for example exit. Finally we should have the following script, which does the conversation and quits the application when we type exit.

import openai

SYSTEM = "You are a helpful AI assistant that can answer questions provided by the user"
HISTORY = [{"role": "system", "content": SYSTEM}]

while True:
    prompt = input("User: ")

    if prompt == "exit":
        break

    HISTORY.append({"role": "user", "content": prompt})

    chat_completion = openai.ChatCompletion.create(
        model="gpt-3.5-turbo",
        messages=HISTORY,
    )

    response = chat_completion.choices[0].message.content

    HISTORY.append({"role": "assistant", "content": response})

    print(f"GPT: {response}")

Conclusion

This project teaches how to use the openai library in it’s basic form, and how to think of a chatbot API endpoint. By saving the history of the messages and keeping track of the responses.

• GitHub Repo