pgljupyter

PlantGL & L-Py jupyter widgets

Quick Examples

PlantGL

• simple PlantGL shapes - spheres @nbviewer

L-Py

• a tree model - leuwenberg @nbviewer

• plot magic - leuwenberg @nbviewer

• champignon @nbviewer

• magics with scalar parameters - broccoli @nbviewer

• magics with curve parameters - sweep surface @nbviewer

Usage

from pgljupyter import SceneWidget, LsystemWidget

SceneWidget

Renderer for PlantGL Scenes, Shapes, Geometries

Arguments:

• arg0 list | plantgl.Shape | plantgl.Scene: a list of shapes or scenes or a single object
• position tuple (float, float, float): x, y, z position of arg0 (default (0, 0, 0))
• scale float: scale factor for arg0 (default 1)
• size_display tuple (int, int): width and height of the canvas (minimum 400)
• size_world float: extend on the 3D scene in all directions

Example:

from openalea.plantgl.all import Scene, Sphere
from pgljupyter import SceneWidget

s = Scene([Sphere()])
sw = SceneWidget(s)
sw

Tutorial:

• Display of spheres @nbviewer

LsystemWidget

Renderer for lpy.Lstrings derived from lpy.Lsystem code

Arguments:

• arg0 string: file name / path of lpy code file
• unit string enum: the unit used in the Lsystem model ('m', 'dm', 'cm', 'mm', default 'm')
• scale float: scale factor for arg0 (default 1)
• animate bool: if True runs animation automatically
• size_display tuple (int, int): width and height of the canvas (minimum 400)
• size_world float: extend on the 3D scene in all directions

Important property/methods:

• editor : display an editor of the graphical parameters
• get_lstring(self) : return the lstring corresponding to the current step of the simulation display in the widget.
• get_namespace(self) : return the namespace of variables of the simulation

Example:

from pgljupyter import LsystemWidget

lw = LsystemWidget('myfile.lpy')
lw

lw.editor

Tutorial:

• Display of the simulation of a lpy file : @nbviewer

%%lpy

A cell magic to inline L-Py code in a notebook

Arguments:

• --size, -s int,int: same as size_display
• --world, -w float: same as size_world
• --unit, -u enum: same as unit
• --params, -p LsystemParameters: name of LsystemParameters instance
• --animate, -a True: runs animation automatically

Example:

# activated by importing pgljupyter
import pgljupyter

%%lpy -u cm
from openalea.plantgl.all import *
Axiom: ;(1)+(10)_(2)F(10);(0)@g(Paraboloid(10,10,2,False))
derivation length: 100
production:
F(x) --> F(x+0.2)
@g(p) --> @g(Paraboloid(p.radius + 0.15, p.height + 0.2, 2, False))
_(x) --> _(x+0.02)

Example with graphical parameters:

# activated by importing pgljupyter
import pgljupyter
from openalea.lpy.lsysparameters import LsystemParameters

lp = LsystemParameters()
lp.add_scalar(name='a', value=1, minvalue=0, maxvalue=100)
lp.add_function(name='b') # default graphical function created

%%lpy -p lp

Axiom: SetGuide(b,a)F(a)

Tutorials:

• simple L-systems @nbviewer
• graphical scalar parameters - broccoli @nbviewer
• graphical curve parameters - sweep surface @nbviewer

%lpy_plot

A line magic to plot all derivation steps of an Lsystem on a square plane

Arguments:

• arg0, string: L-Py file
• --size, -s int,int: width and hight of the canvas
• --cell, -c float: size of cell for a single derivation step

Example:

# activated by importing pgljupyter
import pgljupyter

%lpy_plot myfile.lpy

Tutorial:

• Display of the simulation from a lpy file :@nbviewer

Installation

Install with pip - inside conda env

Building JupyterLab extensions requires nodejs. In case it is not available on your system add nodejs to the conda create command. Some examples also require matplotlib, xarray-simlab and rpy2.

conda create -y -n pgl -c fredboudon -c conda-forge \
    openalea.lpy jupyterlab ipywidgets ipython=7 pip

conda activate pgl
jupyter labextension install @jupyter-widgets/jupyterlab-manager
pip install pgljupyter

Build, install and run from source

• install lpy, plantgl, jupyterlab, widgets and widgetsextension

conda create -y -n pgl -c fredboudon -c conda-forge openalea.lpy jupyterlab ipywidgets
conda activate pgl
jupyter labextension install --no-build @jupyter-widgets/jupyterlab-manager

• install emsdk: https://emscripten.org/docs/getting_started/downloads.html

git clone https://github.com/emscripten-core/emsdk.git
cd emsdk
./emsdk install 2.0.20
./emsdk activate 2.0.20
source ./emsdk_env.sh
cd ..

• get pgljupyter source

git clone https://github.com/jvail/plantgl-jupyter.git
cd plantgl-jupyter

• fetch plantgl and install pgljs deps

git submodule update --init --recursive
cd src/pgljs
npm install
cd ../..

• install pgljupyter deps and build (requires activation of emsdk i.e. source ./emsdk_env.sh)

npm install
npm run build:all

• install python modules and jupyter extensions

pip install -e .
jupyter nbextension install --sys-prefix --overwrite --py pgljupyter
jupyter nbextension enable --sys-prefix --py pgljupyter
jupyter labextension install .
jupyter lab clean

• run the lab

jupyter lab --notebook-dir=./examples

Docker

Run jupyter as docker container locally. Replace /examples with the path to your notebooks. Tag latest might not always be up-to-date since docker is primarily used for binder

docker pull jvail/plantgl-jupyter:1.2.0
docker run --rm \
    -p 8888:8888 \
    -v $PWD/examples:/home/jovyan/work jvail/plantgl-jupyter:1.2.0 \
    jupyter lab