Please consult the documentation for general usage.
We use uv to manage virtual environments. First, install uv, then activate the virtual environment:
uv sync --all-extras
. .venv/bin/activateThe package will install a script allowing the production server to be launched with InsightBoard. To launch the dashboard manually run python -m InsightBoard. To launch the dashboard in debug-mode (which uses Dash's bundled Flash server) run python -m InsightBoard --debug. If you wish to launch InsightBoard from within python, import InsightBoard then call InsightBoard.main().
Use Semantic Versioning. Version management is handled by setuptools-scm, which extracts the current version number from the git tags.
Note that setuptools-scm writes to a src/InsightBoard/version.py file which should not be checked in to git (the file is listed in .gitignore), but will allow packaged releases to query InsightBoard.__version__ to get the current version number. During development this will look peculiar (e.g. 0.1.0.dev...).
To create a new release, navigate to github - Releases, then select Draft a new release. Create a new tag in the format vX.Y.Z (e.g. v0.1.0), click Generate release notes, then Publish release. This will trigger a github action that will build the Python wheels and upload the package to PyPI. It is worth keeping an eye on this action to ensure it completes successfully, and making any necessary changes if it does not.
Experimental
To build the application as a standalone executable, we use PyInstaller. This will create a dist/InsightBoard folder containing the executable and all necessary dependencies. To build the application, run ./dev/build_app.sh.