[pgf-tikz] Need help to understand the parser (and to draw parse trees)

[Eduardo Nahum Ochs]

Hi list,

I'm looking for some help to create parse trees for TikZ - not parse
trees _in_ TikZ, parse trees _for_ Tikz, that would show how certain
TikZ commands are parsed...

Let me try to explain. I've been working on a certain set of tools for
Lpeg that can display the parse tree of an expression like this one

  if (x>9) { x=0; y=y+1; }

both as a tree in ascii art, like this one,

  Stmt__.________________.
  |     |                |
  if    Expr__._____.    Stmt
        |     |     |    |
        Id    Optr  Num  StmtList_______.
        |     |     |    |              |
        x     >     9    Stmt__.__.     Stmt__.__.
                         |     |  |     |     |  |
                         Id    =  Expr  Id    =  Expr__._____.
                         |        |     |        |     |     |
                         x        Num   y        Id    Optr  Num
                                  |              |     |     |
                                  0              y     +     1

and as an underbrace diagram, like in this screenshot:

  http://anggtwu.net/IMAGES/emacsconf2023-repls-b.jpg
  http://anggtwu.net/emacsconf2023.html

There are tons of examples in the TikZ manual that I don't know how to
parse. Here's an example:

  \documentclass[]{book}
  \usepackage{tikz}
  \usetikzlibrary{arrows.meta, positioning}
  %
  \begin{document}
  \begin{tikzpicture}[ auto,bend right ]
    \node (a) at (0:1)   {A};
    \node (b) at (120:1) {B};
    \node (c) at (240:1) {C};
    \draw [->] (a) to                  (b);
    \draw [->] (b) to node        {BC} (c);
    \draw [->] (c) to node [swap] {CA} (a);
  \end{tikzpicture}
  \end{document}

I took it from the section "\subsection{Adding Labels Next to Lines}"
in pgfmanual-en-tutorial-nodes.tex and modified it a bit. I _guess_
that its third \node its parsed in this way:

  Node__.__________________________________.
  |     |                                  |
  node  NameAtCoordinate__.___.            Contents
        |                 |   |            |
        Name              at  Coordinate   {C}
        |                     |
        (c)                   CanvasPolar
                              |
                              (240:1)

  http://anggtwu.net/IMAGES/2024-ParseTikz1-1.png

So: I know several places in the TikZ manual in which it explains
\node, \draw, and \path in different levels of detail, and some of
these places, like the \section{Syntax for Path Specifications} in
pgfmanual-en-tikz-paths.tex, even try to explain what is the exact
language that TikZ is able to parse... but the TikZ parser is still a
mystery to me, and in constructs like these ones...

  \subsection{The Move-To Operation}
  \subsubsection{Horizontal and Vertical Lines}

...I don't have any good guesses on how to name certain subtrees of my
parse trees.

Questions: what are the internal names of move-to, -|, and |-? Where
does the parser define support for them? Also, suppose that my
approach works, and that I manage to produce, say, 20 nice underbrace
diagrams showing how to parse TikZ commands that I found hard to
understand... then I will need names for many non-terminals in the
TikZ language, and I will need a good way to choose those names. Do
you have any other suggestions on how to choose those names besides -
this is just a wild guess - looking at the tex macros with names of
the form such and such in tikz.code.tex? Also, btw, is the parser
written by hand?

  Thanks in advance!
    Eduardo Ochs
    eduardoochs at gmail.com
    http://anggtwu.net/#eev
    http://anggtwu.net/eev-tikz.html


P.S.: I'm sending this from a secondary e-mail because mailmain says
that in my gmail address "You already have a subscription pending
confirmation"... anyone knows how to solve that?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://tug.org/pipermail/pgf-tikz/attachments/20240206/73f50ce5/attachment.htm>