This is a Python-Qt based application that uses the GPT-xxx model to create a helpful assistant based on opeAI GPTs that can answer any question you ask.
This is a simple demo of using openAI API to simulate chatGPT.
All like screenshots, here one:
Everything
Install the required packages via pip:
pip install PyQt5 openai
The user interface is built using Qt Designer which creates an *.ui files (e.g.assistent.ui). This file needs to be translated into Python code using the following command:
pyuic5 assistent.ui -o assistent.py
-
Set your OpenAI API Key as an environment variable:
export OPENAI_API_KEY=your_api_key_here -
Run the Python script:
python main.pyOptionally, you can enable debug or info logs by passing "debug" or "info" as an argument:
python main.py debugpython main.py info
The code consists of several classes:
openAIBackgroundTask: A QThread subclass that handles incoming stream from GPT-4 model in the background.MainWindow: The main window class that handles the UI and user interactions.
The MainWindow class setups up the UI, connects buttons to their corresponding methods, and allows for window resizing and dragging. It also includes methods for handling text input and output and submitting user prompts to the GPT-4 model.
When the user submits a prompt, the submit_action method is called, which creates an openai_arg dictionary with the required parameters for the GPT-xxx model. The openAIBackgroundTask is then started in a separate thread, and its progress signal is connected to the update_text_message method, which updates the UI with the assistant's response.
The main part of the code sets up logging, retrieves the OpenAI API key from the environment variable, and initializes the QApplication and MainWindow.
GPT model can be changed in the code
MODEL = "gpt-3.5-turbo"
#MODEL = "gpt-4"
No license, use it as you like.