Adding a third level of compounds distinction in chemnum?

Question

Essentially I would like to add a "sub-sub-label" to my compounds' numbering using chemnnum.

Why and how exactly I want this

I'm describing a series of chiral compounds, numbering them 1a, 1b, and so on. Sometimes I have also an another diastereomer of a compound: lets say 1b has epimerized in 2nd position - I would like to name it 2-epi-1b (supervisor's preference).

How I'm doing it now

Currently I'm just reusing existing references to 1b with bold prefix, like this: \textbf{2-epi-}\refcmpd{one.b}, but I find this to be not convinient and error-prone.

How I would like to do it

I was thinking about decreasing the appropriate sub-counter by 1 directly after first \cmpd{one.b} and immediately declaring an epimerized compound like this: \labelcmpd[pre-label-code=\textbf{2-epi-}]{one.b:epi}. Unfortunately, I've learned that there is no sub-counter, as chemnum stores current ID number of sub-label as an expl3 ints and I have no Idea how to interact with them.

The question(s)

Is it possible to manually decrease internal sub-label counter by 1 in chemnum? If yes, how to do it? Or maybe there is a better/alternative way to introduce a "sub-sub-label" to the numbering of my compounds?

Answer 1

After reading the source code of chemnum package and learning some expl3 programming, I've came up with a hacky way to do this. I've created a new command that defines a label (but does not print it, like \lebelcmpd), mimmicking a given label and sub label. The code:

\ExplSyntaxOn
\int_new:N \l__chemnum_tmpnum_int

% #1: options, #2: main ID, #3: sub ID, #4: new main ID
\NewDocumentCommand {\sublabelcmpd} {O{ }mmm} {
    % stash main counter value
    \int_set:Nn \l__chemnum_tmpnum_int {\value{cmpdmain}}
    % set main counter, that it will produce #2 label again
    \setcounter {cmpdmain} {\int_eval:n {\cmpdproperty{#2}{number}-1}}
    % define new compound disguised as #2 with dummy sub compound
    \chemnum_cmpd:nnnn {\c_true_bool} {\c_false_bool} {#1} {#4.subundefined}
    % set sub counter to produce desired sub label #3
    \int_set:cn {g__chemnum_compound_#4_subcompound_int} {\subcmpdproperty{#2}{#3}{number}-1}
    % define new sub compound disguised as #2.#3
    \chemnum_cmpd:nnnn {\c_true_bool} {\c_false_bool} {#1} {#4.#3}
    % revert previous main counter state
    \setcounter {cmpdmain} {\l__chemnum_tmpnum_int}
}
\ExplSyntaxOff

This way I can provide different/additional options for what seems to be a previously used label. This method requires a new main label, because options are associated with main label and cannot be changed. The syntax is \sublabelcmpd[<options>]{<main ID>}{<sub ID>}{<new main ID>}, new compound can be then simply referenced with \cmpd{<new main ID>}{<sub ID>}. To give an example:

Compounds \cmpd{one.a} and \cmpd{one.b} are defined as usual.
Then a new compound is defined using
\verb#\sublabelcmpd[pre-label-code=\textbf{2-epi-}]{one}{b}{epi:one}#.
\sublabelcmpd[pre-label-code=\textbf{2-epi-}]{one}{b}{epi:one}
This newly defined compound will have the same label as \cmpd{one.b},
but with additional options, and can be referenced normally: \cmpd{epi:one.b}.

will produce: