search

Plastic-Scanner / PSplot

A lightweight tool for obtaining and visualising the discrete near-infrared (NIR) data using the Plastic Scanner
GNU General Public License v3.0
7 stars 5 forks source link

Packaging, cross-platform, easier installation #13

Open jurc192 opened 2 years ago

jurc192 commented 2 years ago

It is tedious for non-tech users to install python, pip, pipenv and all dependencies to make spectraplot run. How can we make it simpler? Even if it's in early development and probably does not deserve a "stable release", how can we distribute it to people? One option I see is packaging it (wheel?) for installation with pip. Another is making a single-click-executables.

If anyone has experience with python packaging, I'll be glad to have a chat :)

JoostScheffer commented 1 year ago

Maybe this can be used: https://build-system.fman.io/

Here are 2 tutorials:

  1. https://github.com/mherrmann/fbs-tutorial
  2. https://pythonbasics.org/compile-pyqt-to-exe/
JoostScheffer commented 1 year ago

There is also some information on python.org: https://packaging.python.org/en/latest/overview/#bringing-your-own-python-executable

JoostScheffer commented 1 year ago

pyqtdeploy could be used for packing But I don't now how well it will handle the inclusion of PyQtDataVisualisation. Then you can also deploy to iphone or android.

kegsay commented 7 months ago

I think the easiest option would be to package up this project as a Docker container and then run a web UI which you can connect to over localhost.

End users can then just run the docker container and use a normal web browser to view the plots.

This would of course be quite a large change as it's changing languages and would make it harder to collect live data and plot it immediately like you can do now. However, it would mean that you could even host the plot UI on GitHub Pages and just have an "open file" dialog to load the data, making it absolutely trivial for end users to use (just visit X website and load your file).