Command to uppercase the first letter of each word in a sentence

Question

I'm writing a textboox in latex, and I have things like

\textit{some definition}\index{Some Definition}

all over the place. I'd like to combine them into something like

\define{some definition}

and have it expand automagically. So I can have

\newcommand{\define}[2]{\textit{#1}\index{#2}}

which works, but forces me to include both arguments. The second argument is almost always the first argument with the first letter of each word uppercased. How can I write a command which uppercases the first character of each word?

Suggestions of better indexing best practices would also be appreciated.

Answer 1

\documentclass{article}
\usepackage{makeidx}\makeindex
\newcommand*{\formatfirst}[1]{\MakeUppercase{#1}}
\makeatletter
\newcommand*{\mymacro}[1]{%
  \expandafter\formatfirst\expandafter{\@car #1\@empty\@nil}%
  \@cdr #1\@empty\@nil}
\newcommand*\myMakeUpperCase[1]{%
  \def\@myuppercasewords{\myuppercase@i#1 \@nil}%
    {\itshape\@myuppercasewords}\index{#1@\@myuppercasewords}}
\def\myuppercase@i#1 #2\@nil{%
  \mymacro{#1}%
  \ifx\\#2\\%
  \else
    \@ReturnAfterFi{%
      \space
      \myuppercase@i#2\@nil
    }%
  \fi} 
\long\def\@ReturnAfterFi#1\fi{\fi#1} 
\makeatother

\begin{document}
foo
\myMakeUpperCase{capital letter} bar
\myMakeUpperCase{Next one} baz
\myMakeUpperCase{two words}

\printindex

\end{document} 

Answer 2

The mfirstuc package from the glossaries bundle provides the handy \capitalisewords command which could be used here. I adapted Herbert's example and the output is the same:

\documentclass{article}

\usepackage{mfirstuc}
\usepackage{makeidx}\makeindex

\makeatletter
\newcommand*\define[1]{%
  \textit{#1}%
  \index{#1@\protect\capitalisewords{#1}}%
   }
\makeatother

\begin{document}
foo
\define{capital letter} bar
\define{Next one} baz
\define{two words}

\printindex

\end{document}

Answer 3

I am not sure if I understand your question correctly, but a short macro as shown below, will capitalize the first word you type and place it in the index.

\documentclass[11pt]{article} 
\usepackage{index}
\makeindex
\begin{document}
\def\Index#1{\def\x##1##2{\MakeUppercase{##1}{##2}}\textit{\x#1} \index{\x#1}} 
\Index{alpha}

\def\indeX#1#2{\def\x##1##2{\MakeUppercase{##1}{##2}}\x#1 \index{\x#2}} 
\Index{alpha}

\indeX{test}{this}

\printindex
\end{document}

Answer 4

My pure LaTex3 solution:

\documentclass{minimal}
\usepackage{xparse}
\usepackage[T1]{fontenc}

\ExplSyntaxOn

\NewDocumentCommand{\ucFirst}{m} {
  \sphakka_ucfirst:f {#1}
}
\cs_new_protected:Npn \sphakka_ucfirst:f #1 {
  \exp_last_unbraced:Nx \tl_to_uppercase:n {\tl_head:f {#1}}
  % notice the ~ to put back blank padding
  \tl_tail:f {#1}~
}

\NewDocumentCommand{\ucFirstMore}{m} {
  \sphakka_ucfirstmore:f {#1}
}
\cs_new_protected:Npn \sphakka_ucfirstmore:n #1 {
  % must convert the TL into a SEQ
  \seq_set_split:Nnn \l_sphakka_seq {~} {#1}
  \seq_map_function:NN \l_sphakka_seq \sphakka_ucfirst:f
}
\cs_generate_variant:Nn \sphakka_ucfirstmore:n {f}

\ExplSyntaxOff

\begin{document}

\def\word{bar}
\def\sentence{foo bar baz}

[Plain word] quux => \ucFirst{quux}\par
[Word via macro]~\word~=> \ucFirst{\word}\par
[Plain sentence] foo bar baz => \ucFirstMore{foo bar baz}\par
[Sentence via macro]~\sentence~=> \ucFirstMore \sentence \par

\end{document}

Though simple, the above solution leaves a trailing non-breakable space which is rather tricky to get rid of. Thus, I wrote a variant that stores capitalized tokens into another seq, then glues them by spaces:

\documentclass{minimal}
\usepackage{xparse}
\usepackage[T1]{fontenc}

\ExplSyntaxOn

\NewDocumentCommand \ucFirst {>{\TrimSpaces}m} {
  % Won't trim explicit trailing `\space's... why?
  \sphakka_ucfirst:f {#1}
}
\cs_new_protected:Npn \sphakka_ucfirst:f #1 {
  \exp_last_unbraced:Nf \tl_to_uppercase:n {\tl_head:f {#1}}
  \tl_tail:f {#1}
}

\NewDocumentCommand \ucFirstMore {>{\TrimSpaces}m} {
  \sphakka_ucfirstmore:f {#1}
}
\seq_new:N \l_out_seq
\cs_new_protected:Npn \sphakka_ucfirstmore:n #1 {
  % split by blanks. Can anybody tell me why `\space` doesn't work?
  \seq_set_split:Nnn \l_in_seq {~} {#1}
  % remove empty/blank items
  \seq_remove_all:Nn \l_in_seq {}
  \seq_remove_all:Nn \l_in_seq {\space}
  % capitalize each token and store it into another seq
  \seq_map_inline:Nn \l_in_seq {
    \seq_put_right:Nn \l_out_seq {\sphakka_ucfirst:f {##1}}
  }
  \seq_use:Nnnn \l_out_seq {\space}{\space}{\space}
  \seq_clear:N \l_out_seq
}
\cs_generate_variant:Nn \sphakka_ucfirstmore:n {f}


\ExplSyntaxOff

\begin{document}

\edef\word{\space  quux\space }
\edef\sentence{  \space  foo  bar baz \space }

\begin{tabular}[c]{p{.6\linewidth}l}
  \verb|\word|                                          & \verb|`\space  quux\space '|\\
  \verb|\sentence|                                      & \verb|`\space  foo  bar baz \space '|\\
  \\
  \verb|\ucFirst{\space  quux\space }|                  & `\ucFirst{\space  quux\space }'\\
  \verb|\ucFirst{\word}|                                & `\ucFirst{\word}' \\
  \verb|\ucFirstMore{\word}|                            & `\ucFirstMore{\word}' \\
  \verb|\ucFirstMore{  \space  foo  bar baz \space }|   & `\ucFirstMore{  \space  foo  bar baz \space }'\\
  \verb|\ucFirstMore{\sentence}|                        & `\ucFirstMore{\sentence}'\\
\end{tabular}

\end{document}

There are just a couple of minor glitches that I couldn't resolve (possibly because of my inexperience...):

• Why \TrimSpaces (or \tl_trim_spaces) doesn't remove explicit trailing \spaces?

• Why \seq_set_split cant't split by \space?