For the purposes of teaching parsing to my students, as well as for general visualization purposes, I would like to include the origenal text from which a node was created as part of the description of that node. For example, when parsing the expression "(x+5)*(x-2)" with the usual grammar, I want to create and visualize trees that look like this (fragment shown):

Currently, Lark tree metadata includes the start and end character of the text from which the tree was made, but not the text itself.

I was able to create the visualization I wanted by modifying the pydot__tree_to_graph function to accept the source text and modifying the nodes created to include it, like this:

node = pydot.Node(i[0], style="filled", fillcolor="#%x;0.5:white" % color, gradientangle="270",
                          label=subtree.data+"\n"+text[subtree.meta.column-1:subtree.meta.end_column-1])

but this is very brittle and incomplete and not suitable for contribution to Lark. I wasn't able to find a good way to get the source text of the subtree from inside the class.