Instead of using SVG RAW code (which could be tricky when one needed to modify some elements), a new approach only using the {ggplot2} package (along with the {ggtext} extension) was developped.
This approach is based on simple and reproducible ggplot syntax and based on a few reference dataframes containing coordinates for lines and boxes (nodes) which are stored in the package as internal data that can be exported on your computer by using :
The node labels used in the source code which produces the colored
trees are found in another internal object, reference_list
,
which is a list gathering several tibbles with reference info. By
modifying the name
attribute in $indic_dim
,
$indic_prop
and $properties_nodes
, you can
translate the colored trees which will be produced.
The algorithm for producing colored trees was the result of a long work involving the testing of many methods and R packages. The first solution chosen was to draw the models in “white” on a vectorial drawing software (Inkscape) in svg format. SVG is a markup language (similar to XML or HTML) with which it is fairly easy to apply search and replace algorithms. A rectangle tag (corresponding to a node of the illuminated tree) can be written :
<rect
style="fill:#ffffff;fill-opacity:1;stroke:#001800;stroke-width:0.1"
id="rect2-16"
width="17.833706"
height="4.9643545"
x="370.30774"
y="143.9761"
inkscape:label="#rect2-16" />
So all we had to do was find the tag for each rectangle, assign each rectangle identifier to its corresponding indicator, then we can search-replace the argument “fill:#ffffff” by the appropriate color corresponding to the evaluation :
IDEATools then converted this modified SVG source code to a proper
PNG/PDF image using the {rsvg}
library.
The templates were stored as internal object R named “canvas”. It was a list containing in each element the complete SVG source code of each template.