Divoom PC Monitor Python - Implementation for Linux Systems

This repository contains the Python code that interfaces with the Divoom PC Monitor application on the Divoom Times Gate or Divoom Pixoo64 to monitor system information for Linux systems.

Overview

This project is a Python implementation of the Divoom PC Monitor originally written in C# by DivoomDevelop. It allows users to monitor various system metrics and display them on their Divoom Times Gate or Divoom Pixoo64.

Features

• System Metrics Displayed:
  • CPU usage percentage
  • CPU temperature
  • RAM usage
  • GPU usage (NVIDIA, AMD, and Intel)
  • GPU temperature (NVIDIA, AMD, and Intel)
  • HDD usage (dynamic drive selection based on highest usage)
• Easy Setup: Configuration script simplifies the setup process.
• Scheduled Task Support: Users can copy the SystemD config files for easy systemctl support.

Requirements

The project needs both Python packages and a few Linux command-line tools that are called directly by the scripts.

Hardware and network

• A Divoom Times Gate or Pixoo64 on the same LAN as the Linux host
• Internet access during divoom_setup.py so the script can query Divoom cloud endpoints
• A Linux system with Python 3

Python packages

The monitor needs psutil and requests.

On Fedora, install them with the system pip. When the system site-packages directory is not writable, pip installs them for your user:

pip install -r requirements.txt

On Debian and Ubuntu, install them into a virtualenv:

python3 -m venv .venv
.venv/bin/pip install -r requirements.txt

Required Linux packages / binaries

• lspci from pciutils Setup uses it to list GPUs. The monitor uses it only when config.env does not already name a GPU.

# Fedora
sudo dnf install pciutils

# Debian / Ubuntu
sudo apt install pciutils

# Arch Linux
sudo pacman -S pciutils

CPU percentage, memory, disk busy time, and most temperatures come from psutil. Disk names come from /sys/block. iostat is not used.

GPU vendor-specific requirements

• NVIDIA: nvidia-smi, usually shipped with the proprietary driver.
• AMD: the kernel gpu_busy_percent and hwmon files. rocm-smi is used only when those files are missing.
• Intel: the same kernel files. nvtop -s is the fallback, and a JSON null from nvtop is left blank instead of being shown as 0%.

lm-sensors (sensors) is optional. It is consulted only when psutil cannot read a CPU temperature.

Optional system software

• systemd Only required if you want to run the monitor as a service with the provided unit files.

Usage

1. Configuration:

   • Run divoom_setup.py to create the necessary configuration file and established the Divoom PC Monitor app on a LCD face.

2. Running the Script:

   • Execute divoom_pc_monitor.py
   • The interval is the time from the start of one update to the start of the next. Each update spends about one second measuring CPU and disk.
   • --testing collects the same readings and does not send them to the device.
   • Run python divoom_pc_monitor.py --help for the full argument list.

Installation

1. Clone the repository:

   git clone https://github.com/KallanX/Divoom-PC-Monitor-Python.git
   cd Divoom-PC-Monitor-Python
   

2. Install the Python dependencies.

   Fedora uses the system interpreter:

   pip install -r requirements.txt
   

   Debian and Ubuntu use a virtualenv:

   python3 -m venv .venv
   .venv/bin/pip install -r requirements.txt
   

3. Run the setup. It writes config.env next to the scripts, including the CPU vendor and the GPU you select. Re-run it if you add a drive or change graphics cards.

   # Fedora
   python divoom_setup.py
   
   # Debian / Ubuntu
   .venv/bin/python divoom_setup.py
   

   ⚠️ Attention: If the PC Monitor app (625) is currently displayed on a 'LCD face', the setup output will include 'Divoom PC Monitor'. This message doesn't restrict which 'LCD face' you can ultimately use; the app can be assigned to any face regardless of its initial state.

4. Execute main monitor:

   # Fedora
   python divoom_pc_monitor.py -i 10 -v
   
   # Debian / Ubuntu
   .venv/bin/python divoom_pc_monitor.py -i 10 -v
   

5. (Optional) Install the systemd service. Edit User, ExecStart, and WorkingDirectory in the distro unit, then copy it. On Fedora, User must be the account that ran pip install, because those packages are installed for that user. On Debian and Ubuntu, ExecStart must use the virtualenv's Python.

   cp systemd-configs/<DISTRO>/divoom_monitor.service /etc/systemd/system/
   sudo systemctl daemon-reload
   sudo systemctl enable divoom_monitor.service
   sudo systemctl start divoom_monitor.service
   

Work In Progress

1. Intel Arc still depends on the kernel exposing gpu_busy_percent and a hwmon temperature. Where those files are missing, readings depend on nvtop and may stay blank.
2. An on-going bug hunt and optimization. User contributions/recommendations are welcome.

Author

This project was created by Keith Carichner Jr. due to the lack of a competent implementation that actually functioned and was well-written/documented.

License

This project is licensed under the MIT License - see the LICENSE file for details.