• python >= 3.5 (with dev headers)
• setuptools (pip install setuptools)

Edit local.properties in the project root:

• architecture - set to arm64 or x86_64 depending on your Python architecture.
• enable_python_package - set to true for working with Python package.
• python.include_path - path to Python include directory where Python.h located. To get this path you can run python -c "from sysconfig import get_paths as gp; print(gp()['include'])".
• python.bin_path - path to your Python bin directory. To get this path you can run python -c "from sysconfig import get_paths as gp; print(gp()['scripts'])".

Gradle build in the sibling project python-package-build

or just run shell command (if the changes are only in the Python code):

python setup.py update_js bdist_wheel

• change dir to lets-plot/python-package
• create / activate Python environment (if needed)
• run shell command:

pip install --no-index --find-links=dist/ lets-plot --no-deps --force-reinstall

When the "dev" version is built, lets-plot embeds current "dev" js into Jupyter notebook.

If necessary, this default can be temporarily overwritten by editing dev_xxx settings in _global_settings.py

The "dev" version of js library can be served from dist folder of js-package project like:

# Start local web-server to serve dev js script:
$ cd lets-plot
$ python -m http.server 8080

from lets_plot import *
data = dict(time=['Lunch', 'Lunch', 'Dinner', 'Dinner', 'Dinner'])
p = ggplot(data) + geom_bar(aes(x='time', fill='..count..'))
p += scale_fill_hue()
p