Morsecode GPT to Allows the Audience to Interact With an AI Program Using Morse Code to Communicate.

I explain "Morsecode GPT" work which allows the audience to interact with an AI program using Morse code to communicate for exhibiting in Orlando Maker Faire 2024.

I liked the idea of ​​wireless communication like Morse code in the movie. I originally wanted to exhibit this idea at there. Since LLM (Large Language Model) is widely open source, I thought I could use it to create a production that looks like people are communicating. I'm sharing the development code here.

I made two program modules using Arduino and Python. I explain about them like below (I will take the time to update the content further).

Before following this, you should install ollama, Llama2, Arduino IDE, Python.

1. Download Ollama on Windows

1. Initialization (setup phase)

• The Arduino initializes the LCD screen and sets up pins for the Morse code button, new text button, and the buzzer.

A welcome message, "Hi. Morse world!" is displayed on the LCD for 3 seconds.

2. Reading Button States (loop phase)

• The main loop() function continuously reads the state of the Morse code button (button_morse) and the new text button (button_new_text).

• If button_morse is pressed, the buzzer emits a tone, simulating the pressing of the Morse key.

• The duration the button is pressed (time_pressed_button) is measured to differentiate between dots (short press) and dashes (long press).

3. Morse Code Conversion

• Once the button is released, the recorded time_pressed_button value determines if it represents a dot (short press) or a dash (long press).

• These Morse symbols are appended to morse_code.

• If there is a pause between button presses (detected via time_not_push_button), the current Morse code is converted to a text character using complate_morse_code() and added to morse_text.

4. Displaying and Resetting

• If the button_new_text is pressed, the current morse_text is sent to the serial monitor for further processing and the code resets (morse_code and morse_text are cleared).

• The LCD screen displays the current Morse code sequence and the resulting text.

1. Initialize and Setup:

• Initialize pygame for creating a graphical display window.

• Set up the display window size, title, and text fonts.

• Define colors for text rendering.

2. Serial Port Configuration:

• Find available COM ports and open one for serial communication at 9600 baud rate.

• Raise an exception if no COM port is found.

3. Language Model Integration:

• Set up a language model chain using langchain, Ollama, and a joke prompt template.

4. Main Program Loop:

• Enter the main loop, which runs until the user quits.

• Check for user events like window closure.

• Read Morse code data from the serial port.

• If data is received, store it in morse_string.

5. Processing and Display:

• Pass morse_string to the language model to get a response (a joke).

• Clear the display window and render the Morse message.

• Render the generated joke, line by line, on the screen.

• Refresh the display with pygame.display.flip().

6. Loop Control and Exit:

• Add a short delay (time.sleep(0.1)) to control loop speed.

• Repeat the loop to check for new serial data and user events.

• Exit the loop when the user closes the window.

7. Cleanup:

• Close the serial port connection.

• Clean up and close the pygame window with pygame.quit().

P.S. Actually, my grandfather suddenly passed away, so I had to inform this situation to 2024 Orlando Maker faire's staff and go back home. My grandfather loved me, and I had fond memories of going out with him a lot when I was young. For this reason, I shared my work here with regret.