#!/usr/bin/perl
# Most of the files of project Gutenbers are written
# in plain ascii. They are easy to read on a screen, but
# the quality of a print--out is usually sub--optimal.
# The aim of this script is to convert automatically
# such a text into a LaTeX file for high quality
# typesetting.
#
# There are some drawbacks, though:
# namely, there is little consistency in the way 
# pg texts are marked: even to guess author and title of
# a book some work is required.
# Furthermore, sometimes it is possible to find mistakes in
# the form of inverted commas which have never been closed,
# marking which do not correspond to each other, etc.


$quo=0;
$ital=0;
$bold=0;
$text=0;
$chapter=0;
$gottitle=0;

while(<>)
{
    # STANDARD REPLACEMENTS
    s/\r//g;                     # Remove CR
    s/--/ --- /g;                # TeX uses --- as punctuation
    s/(\w)-(\w)/$1--$2/g;        #     and --  as inter-word hyphen
    s/\. \. \./\\dots /g;        #     dots
    s/\.\.\./\\dots /g;          #     dots (sometimes they are like this)
    s/\&/\\\&/g;                 #     &
    s/\%/\\\%/g;                 #     %
    s/\$/\\\$/g;                 #     $ 
    s/([A-Z][a-z]*)\.\s([A-Z])/$1.\\ $2/mg; # Abbreviations
                                            # like Mr., Dr., etc.

    # Deduce title and author from the heading (if present)
    if ((/Etext of (.*) by (.*)/) ||
	(/Gutenberg Etext, (.*) by (.*)/))
    {
	print "% Title algorithm 1\n";
	$title=$1; 
	$author=$2;
	$title=~s/\W*([ [:alpha:]*])\W$/$1/g; 
	$author=~s/\W*$//g;
    }; 
    # Sometimes we are given title and author in a nice table...
    if (!($title) && (/Title: /))
    { $title= $';
    printf "%Title algorithm 2\n";
    }
    if (!($author) && (/Author: /))
    { $author=$'; }
    # ... but this is not always the case...
    if (!($title || $author || $text))
    { if (/(.*), by (.*)/)
      {
	  printf "%Title algorithm 3\n";
	  $title=$1;
	  $author=$2;
	  $title=~s/\W*(.*)/$1/g;
	  #$title=~s/\W*([ [:alpha:]*])\W$/$1/g; 
	  $author=~s/\W*$//g;
      }
  }
# Begin a new chapter 
if ($chcg)
{
    if (/^\n/)
    {
	next;
    }
    else
    {
	print "\\chapter{}\n";
	if (/^(\s*)\w/)
	{
       	print "\\par\\cappar ";
	}
	else
	{
	print "\\par";
	s/^(\s*)([^\w]*\w)/\\def\\a\{$2\}\\docappar /;
	}
	$chcg=0;
   }
};
# Replacements for the text
if ($text)
{
    while (/\"/) 
    {
	if ($quo)
	{ s/\"/\'\'/;
          s/\'\'\'/\'\\,\'\'/;
	 $quo=0; next; }
	s/\"/\`\`/; $quo=1; 
    };
    while (/_/)
    {
	if ($ital)
	{
          #Behold the italic correction 
          (s/_\s/\}\\, /) || (s/_/\}/);
          $ital=0; next;
        }
	s/_/\{\\it /; $ital=1;
    }
    # Chapter marking is a mess: sometimes there is a
    #                number, sometimes there is not; the
    #                capitalisation might vary as well.
    if ( 
         (/^CHAPTER/)||
         (/^(\s*)Chapter[IVXCM [0-9]]*(\s*)/)
       )
    {   if($chapter==0) 
	{ print "\\tableofcontents\n"; };
	$chapter++; $chcg=1;
	next; }
    # Close quotes when paragraph ends
    #  (I would advise to kill the editor, but this is
    #   a voluntary effort, hence...)
    if ((/^$/) && ($quo))
    { print "\,''\n\n"; $quo=0; } 
                                
    if (/End (.*) Gutenberg/)   
    { $text = 0;
      #standard Epilog
      print <<ENDEPILOG;
      \\vskip2cm
      \\centerline{\\sc THE END}
      \\end{document}
ENDEPILOG
      }
    print; 
}
if (/.END.THE SMALL/)
{
  #Standard Prologue
  print <<ENDPROLOG;
  \\documentclass[A4,12pt]{book}
  \\usepackage{fontenc}[t1]
  \\usepackage{times}
  \\usepackage[verbose]{geometry}
  \\title{$title}
  \\author{$author}
  \\ifx\\pdfoutput\\undefined \\newcount\\pdfoutput \\fi
  \\ifcase\\pdfoutput
  \\special{pdf: docinfo
      <<  /Author ($author)
          /Title ($title)
      >> }
  \\else
  %\\usepackage{times}
  %% Document info
  \\pdfinfo{
          /title        ($title)
	  /Author       ($author)
           }
  \\pdfcompresslevel 9
  \\geometry{pdftex}
  \\fi
  \\usepackage[british]{babel}
  \\usepackage{fancyhdr}
  \\usepackage{calc}
  %% SIZES
  \\geometry{a4paper,twoside}
  \\geometry{tmargin=2cm,bmargin=4cm}
  \\geometry{lmargin=3cm,rmargin=4cm}
  %\\baselineskip=.8cm
  \\setcounter{tocdepth}{4}
  %%% DEFINE HEADINGS\n
  \\pagestyle{fancy}
  \\fancyhead[LE]{\\slshape \\thepage}
  \\fancyhead[RE]{\\slshape \\MakeUppercase{$title}}
  \\fancyhead[LO]{\\slshape \\leftmark}
  \\fancyhead[RO]{\\slshape \\thepage}
  \\fancyfoot[C]{}
  \\makeatletter
  % Fix-up Chapter (use empty pagestyle)
  \\renewcommand\\\@chapapp{}
  \\renewcommand\\chapter{\\if\@openright\\cleardoublepage%
                          \\else\\clearpage\\fi
			  \\thispagestyle{empty}%
			  \\global\\\@topnum\\z\@
			  \\\@afterindentfalse
			  \\secdef\\\@chapter\\\@schapter}
  \\renewcommand\\thechapter{Chapter \\\@arabic\\c\@chapter}
  \\makeatother
  %% DROPPED CAPITALS (From TeX FAQ)
  \\newcommand\\capfont{\\fontsize{55}{60}\\selectfont} %FONT (T1)
  \\newbox\\capbox \\newcount\\capl \\def\\a{A} %BOX
  \\def\\docappar{\\medbreak\\noindent\\setbox\\capbox\\hbox{%
		  \\capfont\\a}\\hskip-0.18em%
  \\hangindent=\\wd\\capbox%
  \\capl=\\ht\\capbox\\divide\\capl by\\baselineskip\\advance\\capl by1%
  \\hangafter=-\\capl%
  \\raise2pt\\hbox{\\vbox to6pt{\\hbox to0pt{\\hss\\box\\capbox}\\vss}}}
  \\def\\cappar{\\afterassignment\\docappar\\noexpand\\let\\a }
  %% End DROPPED CAPITALS
  %% Title support
  \\makeatletter
  \\newcommand{\\HRule}{\\rule{\\linewidth}{1mm}}
  \\renewcommand\\maketitle{\\begin{titlepage}%
  \\let\\footnotesize\\small
  \\let\\footnoterule\\relax
  \\let \\footnote \\thanks
  \\vspace*{\\stretch{1}}
  \\noindent\\HRule
  \\vskip.6cm
  \\begin{flushright}
    {\\Huge \\\@author} \\\\[.5cm]
    {\\Huge \\\@title}  \\\\[.6cm]
  \\end{flushright}
  \\noindent\\HRule
  \\vspace*{\\stretch{2}}
  \\begin{center}
    {\\large Project Gutenberg e--text \\par}%
  \\end{center}\\par
  \\end{titlepage}%
  \\setcounter{footnote}{0}%
  \\global\\let\\thanks\\relax
  \\global\\let\\maketitle\\relax
  \\global\\let\\\@thanks\\\@empty
  \\global\\let\\\@author\\\@empty
  \\global\\let\\\@date\\\@empty
  \\global\\let\\\@title\\\@empty
  \\global\\let\\title\\relax
  \\global\\let\\author\\relax
  \\global\\let\\date\\relax
  \\global\\let\\and\\relax
  }
  \\makeatother
  \\begin{document}
  \\maketitle
ENDPROLOG
    $text=1;
 };
}