using regex to extract two values (send to macro variables) of unique multiple occurences

Question

I read several key values from a document with bible passages written in spanish like this

1 Corintios 2:4-5; Mateo 15:1-10;Salmos 118:16

I want to extract the citations in each reference (could be one or more than one) and for each citation inside the reference I need to split it to get the "book name" and "chapter" in order to send both to special index. Book names may be preceeded by numbers. For each occurrence I want to "store" the values into l_hpbook and l_hpchapter variables. I am able to use the right regex expression, but not able to store the values in the respective macro variable lp_hpbook and l_hpchapter. Any hints? (functionality not available in bibleref, there is not even an spanish package)

\documentclass{article}
\usepackage{expl3}
\usepackage{l3regex}

\ExplSyntaxOn
\tl_new:N \l_hpsplit_tl

\cs_new:Npn \hpsplit #1 {
% Purpose: Extract all words that are followed by a space + number
%          optionally preceded by a number and space
% Return: All matches are stored in a global token \hpmatches_tl
\str_new:N \l_hpbook
\int_new:N \l_hpchapter 

\tl_new:N \l_holdmatches

\int_new:N \l_hpmatch_int
\tl_set:Nn \l_hpsplit_tl {#1}

% This statement DOES NOT work .  Count of matches not stored in variable
% Count \l_hpmatch_int % Output generates error: Missing number
\regex_count:nnN {\d+\s\w+\s\d+|\w+\s\d+ } { #1 } \l_hpmatch_int
%    Count \l_hpmatch_int % Output generates error: Missing number

% Splits multiple citations: book name and chapter number retrieved
% This statement WORKS!
\regex_extract_all:nnN{ \d+\s\w+\s\d+|\w+\s\d+} { #1 } \l_holdmatches%
\seq_map_inline:Nn \l_holdmatches { 
    item:~##1\par% 

    % Extract the book name of a single citation
    \regex_extract_once:nnN{ \d+\s\w+|\w+} { ~##1 } \l_hpbook%
%       \l_hpbook % this one generates error

    % Extract the book's chapter from a single citation
    % by finding the last occurrence of space and digits
    \regex_extract_once:nnN{(?!.*\s\d)\d+} { ~##1 } \l_hpchapter%
%       \l_hpchapter   % this one generates error if uncommented. 

% I want to use these two variables as follows:
% \index{\lp_hpbook!lp_hpchapter!\articlename}
}%

\tl_use:N \l_hpsplit_tl

%    \hpvar{\seq_map_inline:Nn \l_holdmatch {~##1\par }}%
}
\ExplSyntaxOff

\begin{document}
    % Will read several references from book like this
    % Need to obtain 'book name' separate from chapter
    \hpsplit{1 Corintios 2:4-5; Mateo 15:1-10;Salmos 118:16}
 \end{document}

Answer 1

In my opinion it is best to first split the argument at semicolons and then extract the data from each item.

\documentclass{article}
\usepackage{xparse,l3regex}

\ExplSyntaxOn
\NewDocumentCommand{\hpsplit}{sm}
 {
  \IfBooleanTF{#1}
   { \hpcolos_hpsplit:V #2 }
   { \hpcolos_hpsplit:n { #2 } }
 }

\seq_new:N \l__hpcolos_hpsplit_seq
\seq_new:N \l__hpcolos_entry_seq

\cs_new_protected:Nn \hpcolos_hpsplit:n
 {
  \seq_set_split:Nnn \l__hpcolos_hpsplit_seq { ; } { #1 }
  \seq_map_function:NN \l__hpcolos_hpsplit_seq \hpcolos_hpentry:n
 }
\cs_generate_variant:Nn \hpcolos_hpsplit:n { V }

\cs_new_protected:Nn \hpcolos_hpentry:n
 {
  \regex_split:nnN { \s* ([0-9]+)[\:0-9-]* \Z } { #1 } \l__hpcolos_entry_seq
  \hpcolos_index_entry:
 }

\cs_new_protected:Nn \hpcolos_index_entry:
 {
  Book~is~``\seq_item:Nn \l__hpcolos_entry_seq { 1 }''
  ---
  Chapter~is~``\seq_item:Nn \l__hpcolos_entry_seq { 2 }''
  \par
 }
\cs_generate_variant:Nn \tl_trim_spaces:n { f }
\ExplSyntaxOff

\begin{document}

\hpsplit{1 Corintios 2:4-5; Mateo 15:1-10;Salmos 118:16}

\end{document}

I defined the final function \hpcolos_index_entry: to print the data, define it as you prefer.

The trick is using \regex_split:nnN that splits the given token list according to the regex: the matching part in the capturing group will be remembered, the rest discarded. So the first case will set the sequence to contain 1 Corintios and 2. The space before 2 and the verse numbers will be removed because they're not in the capturing group.

If you want to pass a macro to the command, use \hpsplit*{\foo}, where \foo expands to a list of Bible references as above.