rcwhytock/wildlife-explorer

Wildlife Explorer

A desktop application that allows bioconservation researchers to classify camera trap animal images and analyze the results.

The application is multi-platform. We support MacOS, Linux and Windows 10. Primary focus is on Windows 10.

Maintainers

• Marek Rogala
• Kamil Żyła
• Tomasz Dziopa
• Wojciech Dziwulski

Development

Basic information

• We're using Electron. The app is based on https://github.com/electron-react-boilerplate/electron-react-boilerplate], which internally uses https://github.com/electron-userland/electron-builder.
• Useful files (sample data, biomonitoring stations, icons) can be found on our Drive (Appsilon internal).

Dev environment setup

• Install Yarn (on Ubuntu: https://linuxize.com/post/how-to-install-yarn-on-ubuntu-18-04/) -- for node packages management, we will use https://yarnpkg.com/. This is recommended by Electron Builder (npm can lead to errors).
• Editor configuration: https://electron-react-boilerplate.js.org/docs/editor-configuration
• To test on Windows use virtual machine image from https://developer.microsoft.com/en-us/windows/downloads/virtual-machines/ (install VirtualBox 6 on Ubuntu following this guide: https://tecadmin.net/install-virtualbox-on-ubuntu-18-04/)

Linting

We check for appropriate linting before each commit. You will be able to fix some lint errors automatically using yarn lint-fix.

Running the app

• yarn installs dependencies.
• yarn dev runs the app in development mode (with developer tools, hot swapping, ...).
• yarn start runs the app in production mode (bundled and optimized).

Prepend yarn cross-env DEBUG_PROD=true to a command to enable debugging in production. For example: yarn cross-env DEBUG_PROD=true yarn package.

Packaging for production

yarn package packages the app for install. We run it with GitHub actions and store artifacts, so you can simply download an artifact from there. Model packaging: go to models_runner and follow the README there (both for Windows and Linux).

License

AGPLv3

Definition of Done:

• Change has been tested (manually or with automated tests). No existing functionality is broken.
• Continuous integration checks (linter, project build) pass.
• README is updated with all information related to the change.
• All code has been peer-reviewed before merging into any main branch.
• Change is tested in installed version on Windows and Linux.