====== HTMLDOC ======
[[http://www.htmldoc.org/|HTMLDOC]] is a free, high quality  HTML to PDF converter. The only drawback is, that it doesn't support CSS in its current version. (You can still achieve good results!) The big advantage is, that you don't need to install something else. (no ghostscript e.g.) \\ To use htmldoc in your wiki do the following:
  * Install htmldoc (pretty easy)
  * Add the //Export to PDF// button as described in [[tips:pdfexport#common_changes]].
  * Create a temporary directory that the webserver can write to for the intermediate step.
  * In the function ''act_export'', in ''inc/actions.php'', add this (just after the "global" lines):<code php>
  if($act == 'export_pdf'){
    pdfmake(  p_wiki_xhtml($ID,$REV,false)  );
    exit;
  }
</code>
  * Modify inc/common.php: Append: <code php>
function pdfmake($text){
  $dir="/home/wikiroot/files/pdfexporttmp/"; 
  $filenameInput=$dir."input.html";
  $filenameOutput=$dir."output.pdf";

#clean up the string (so we don't have any artifacts)
  $text = umlaute($text); #this is for german umlaute
  $text = preg_replace("'<div class=\"toc\">\n<div class=\"tocheader\">.*?</div>\n</div>'si",'',$text );
  $text = preg_replace("'<a[^>]*?></a>'si", '', $text );

  $text = str_replace('&ldquo;','&quot;',$text );
  $text = str_replace('&rdquo;','&quot;',$text );
 
  $text = str_replace('<table', '<table border="1" ', $text);

#write the string to temporary html-file
  $fp = fopen ($filenameInput, "w") or die ("cant create file");
  fwrite($fp,$text);
  fclose($fp);

#convert using htmldoc
  $command = "/usr/bin/htmldoc --no-title -f " . $filenameOutput . " " . $filenameInput;
  system($command);
  system("exit(0)");

#send to browser
  $filenameOutput=trim($filenameOutput);
  header("Content-type: application/pdf");
  header("Content-Disposition: attachment; filename=wikiexport_" . str_replace(':','_',$_GET["id"]) . ".pdf");
  header("Pragma: public"); // This makes the PDF-Export via https:// work with IE - it enables the IE to save the PDF file.
  $fd = @fopen($filenameOutput,"r");
  while(!feof($fd)){
    echo fread($fd,2048);
  }
  fclose($fd);

#clean up temporary files
  system("rm " . $filenameInput);
  system("rm " . $filenameOutput);
}
function umlaute($text){
  return strtr($text,array(
      "ß"=>"&szlig;",
      "ä"=>"&auml;",
      "ü"=>"&uuml;",
      "ö"=>"&ouml;",
      "Ä"=>"&Auml;",
      "Ü"=>"&Uuml;",
      "Ö"=>"&Ouml;"));}
</code>

With IE6 you get a "file not found" error if you try to open the pdf generated by dokuwiki with the default pdf reader ( although you may save the file to your local disk). You can fix this by appending this line in the code above :
<code>
header('Cache-control: private, must-revalidate');
</code>

Da DokuWiki auf UTF-8 basiert und die PDFs die Umlaute nicht richtig anzeigen, kann die letzte Funktion durch diese ersetzt werden:
<code php>
function umlaute($text){
  return strtr($text,array(
      utf8_encode("ß") => "&szlig;",
      utf8_encode("ä") => "&auml;",
      utf8_encode("ü") => "&uuml;",
      utf8_encode("ö") => "&ouml;",
      utf8_encode("Ä") => "&Auml;",
      utf8_encode("Ü") => "&Uuml;",
      utf8_encode("Ö") => "&Ouml;"));
}
</code>

====== HTML->PS->PDF ======
This is some of the steps and modifications that were previously at [[http://www.bobbaddeley.com|bobbaddeley.com]]. 

  * Install the appropriate ghostscript and pdf packages: html2ps and Ghostscript's ps2pdfwr are both called. ((On Debian: These are included in packages html2ps and gs-common. All other needed packages (gs for example) are installed using apt-get automatically.))
  * Add the //Export to PDF// button as described above.  
  * Create a temporary directory that the webserver can write to for the intermediate step. (I used pdftmp)
  * In the function ''act_export'', in ''inc/action.php'', add this:<code php>
  }elseif($act == 'export_pdf'){
    $data = p_wiki_xhtml($ID,$REV,false);
    $data = preg_replace("'<div class=\"toc\"><div class=\"tocheader\">.*?</div></div>'si",'',$data);
    $data = preg_replace("'<a[^>]*?></a>'si", '', $data); #to prevent </a> after each header
    pdfmake($data);
    exit;
  }</code> 
  * Modify inc/common.php: Append: <code php>
#PDF Exporting code added to dokuwiki from bobbaddeley.com 
function pdfmake($text){
  $file="pdfout";
  $dir="pdftmp/";
  $fp = fopen ($dir.$file.".html", "w") or die ("cant create file");
  fwrite($fp,$text);
  fclose($fp);
  $filenameInput=$dir.$file.".html";
  $filenameOutput=$dir.$file.".pdf";
  $filenameTemp=$dir.$file.".ps";
  $filenameCapture=$dir.$file.".txt";
  $command1="/usr/bin/html2ps -o " . $filenameTemp . " " . $filenameInput;
  exec($command1);
  system("exit(0)");
  # MAKE SURE THAT YOU DO EXIT(0) OR UNIX REDIRECTION FAILS!! at least for me
  # now we use Ghostscript's ps2pdf perl script to get it into pdf format
  # assumes Ghoscript is installed
  $syscall2="/usr/bin/ps2pdfwr " . $filenameTemp . " " . $filenameOutput . " ";
  system($syscall2);
  system("exit(0)");
  # clean up temporary files
  system("rm " . $filenameInput);
  system("rm " . $filenameTemp);
  $filenameOutput=trim($filenameOutput);
  /**
  * output the pdf code, streaming it to the browser
  * the relevant headers are set so that hopefully the browser will recognise it
  */
  header("Content-type: application/pdf");
  header("Content-Disposition: inline; filename=".$filenameOutput);
  @readfile($filenameOutput);
} </code>

----

For inline image support and internal dokuwiki refs to work in the pdf, modify the html2ps command:

<code bash>
$docbase="/var/www/html/";
$urlbase="http://my.dokuwiki.host/dokuwiki";
$command1="/usr/bin/html2ps -b " . $urlbase . " -r " . $docbase . " -o " . $filenameTemp . " " . $filenameInput;
</code>

Probably the $docbase and $urlbase path could be read from some variable which someone else knows...

====== Discussion ======
Great work thanks. But in order to get filenames based on page name with HTMLDoc, I had to change the line inserted in ''lib/tpl/template/main.php''  
<code php>
<?php print html_btn('exportpdf',$ID,'',array('do' => 'export_pdf', 'id' => $ID)) ?>  <!-- inserted line -->
</code>
I also had to remove space after //wikiexport// in ''inc/common.php''
<code php>
header("Content-Disposition: attachment; filename=wikiexport" . str_replace(':','_',$_GET["id"]) . ".pdf");
</code>
To retrieve images from the wiki server (relative links, hope that it won't cause security issues) (I had problems with PNG files, so I converted them into JPEG format).
<code xml>
$text = preg_replace("'<img src=\"/(.*?)/lib/exe/fetch.php(.*?)media=(.*?)\"(.*?)>'si","<img src=\"http://" . $_SERVER['SERVER_NAME'] . "/\\1/data/media/\\3\">", $text); # for uploaded images
$text = preg_replace("'<img src=\"/(.*?)\"'si","<img src=\"http://" . $_SERVER['SERVER_NAME'] . "/\\1\"", $text); # for built-in images, smileys for example
</code>
The generated code for the table of contents contains endlines so
<code xml>
$text = preg_replace("'<div class=\"toc\">.<div class=\"tocheader\">.*?</div>.</div>'si",'',$text );
</code>
For the //umlaute// function (french support and images link support)
<code xml>
function umlaute($text){
  return strtr($text,array(
      "ß"=>"&szlig;",
      "ä"=>"&auml;",
      "ë"=>"&euml;",
      "ï"=>"&iuml;",
      "ü"=>"&uuml;",
      "ö"=>"&ouml;",
      "Ä"=>"&Auml;",
      "Ë"=>"&Euml;",
      "Ë"=>"&Iuml;",
      "Ü"=>"&Uuml;",
      "Ö"=>"&Ouml;",
      "â"=>"&acirc;",
      "ê"=>"&ecirc;",
      "î"=>"&icirc;",
      "ô"=>"&ocirc;",
      "û"=>"&ucirc;",
      "Â"=>"&Acirc;",
      "Ê"=>"&Ecirc;",
      "Î"=>"&Icirc;",
      "Ô"=>"&Ocirc;",
      "Û"=>"&Ucirc;",
      "à"=>"&agrave;",
      "è"=>"&egrave;",
      "ù"=>"&ugrave;",
      "é"=>"&eacute;",
      "À"=>"&Agrave;",
      "È"=>"&Egrave;",
      "Ù"=>"&Ugrave;",
      "É"=>"&Eacute;",
      "ç"=>"&ccedil;",
      "%3A"=>"/"
));
}
</code>
--- //[[virgile.gerecke@gmail.com|Virgile Gerecke]] 2005-08-19 15:00//

Why does the background not get right to de pdf? I mean for ''<code>'' sections?
  *  What about using something like http://www.fpdf.org/ or http://ros.co.nz/pdf/ ?

After a bunch of experimentation with the various techniques noted above, I found that the best way to get PDF versions of DokuWiki pages is simply to print to PDF.  On Linux and Mac that should be included with your operating system, and Windows users can use [[http://www.primopdf.com/|PrimoPDF]].

Hi, \\
with htmldoc-1.8.24, I have to  modify inc/common.php: 
<code php>
 $command = "/usr/bin/htmldoc  --webpage --no-title -f " . $filenameOutput . " " . $filenameInput;
</code>

====== An HTMLDOC variant ======

Hi, my main problem was the following:
  * HTMLDOC can't handle the utf-8 code-page (I hope it will change at the future)
  * The pictures missed in the PDF document
  * etc.
So, I expanded the first HTMLDOC conversion code with some usable features:\\ 
  * At the first you need to make the [[pdfexport#common_changes|common changes]]
  * You need to change the conf/dokuwiki.php -> add the followings: <code php>
/* PDF Export settings  */
$conf['customtoc']   = '';               #''=use default TOC, any other=always use this TOC e.g. 'Contents'
$conf['uselangtoc']  = 1;                #Use language dependent TOC (1=yes,0=no) - it will be ignored when use customtoc
$conf['embedfonts']  = 1;                #1=use embedded fonts; 0=don't use /smaller pdf, but platform dependent/
$conf['jpgrate']     = 0;                #Jpeg compression rate 0=off jpeg compression (smaller value -> smaller size, but worst quality)
$conf['usecustomreplace'] = 1;           #Use custom replace table (replaces based on conf/replace.conf)
$conf['pdfcp']       = 'windows-1250';   #Which code-page will be used for conversion to pdf
                                         #Possible values are:
                                         # 'windows-874','windows-1250','windows-1251','windows-1252',
                                         # 'windows-1253','windows-1254','windows-1255','windows-1256',
                                         # 'windows-1257','windows-1258','iso-8859-1','iso-8859-2','iso-8859-3',
                                         # 'iso-8859-4','iso-8859-5','iso-8859-6','iso-8859-7','iso-8859-8',
                                         # 'iso-8859-9','iso-8859-14','iso-8859-15'
$conf['htmldocdir']  = '/usr/bin/';      #Folder of htmldoc binary
$conf['pdfversion']  = 'pdf14';          #PDF version compatibility
$conf['browserwidth']= 1280;             #Size of pictures is optimized for this size (your browser width) in your documents
</code>
  * You need to create ''conf/replace.conf'' file for your alternative changes. This is my custom conf: <code>
# Patterns configured here will be replaced by the following patterns.

// Please create a new section for the code-page with @code-page tag

# Replaces for Windows-1250 code-page
#
# If a tag thereisn't in the code-page you can create a missing-like
# tag with multiple tags:

@windows-1250
&ldquo; &#132;                  # Left double quote
&rdquo; &#148;                  # Right double quote
&trade; &#153;                  # TradeMark
&ndash; &#150;                  # Short dash --
&mdash; &#151;                  # Long dash ---
&larr;  &#139;&#150;            # Left single arrow <-
&rarr;  &#150;&#155;            # Right single arrow ->
&harr;  &#139;&#150;&#155;      # Horizontal single arrow <->
&lArr;  &laquo;-                # Left double arrow <=
&rArr;  -&raquo;                # Right double arrow =>
&hArr;  &laquo;-&raquo;         # Horizontal double arrow <=>
</code>

:!: //You need to use same 'pdfcp' code-page-string and @code-page string.// You can use comments as two ways: with ''/''''/'' or # delimiters.\\
You can use more @code-page sections in this file.
  * After these steps you need to append the following (modification of above version) PHP code to inc/common.php: <code php>
function pdfmake($text){
  global $lang;
  global $conf;

  $dir=DOKU_INC."tmp/";
  $filenameInput=$dir."input.html";
  $filenameOutput=$dir."output.pdf";

# Convert text and toctitle to destination code-page
  $text=iconv("utf-8",$conf['pdfcp'],$text);
# Change toctitle if needed
  if ($conf['customtoc']) {
    $toctitle=$conf['customtoc'];
    }
  elseif ($conf['uselangtoc']) {
    $toctitle=$lang['toc'];
    }
  else {
    $toctitle="Table of contents";
  }
  $toctitle=iconv("utf-8",$conf['pdfcp'],$toctitle);

# htmldoc compatible name-conversion
  $pdfcp=preg_replace("/windows/i","cp",$conf['pdfcp']);
  $text = preg_replace("'<div class=\"toc\"><div class=\"tocheader\">.*?</div></div>'si",'',$text );
  $text = preg_replace("'<a[^>]*?></a>'si", '', $text );

# Execute changes based on replaces.conf
  $replacesf=DOKU_INC . "conf/replaces.conf";
  if ($conf['usecustomreplace'] && file_exists($replacesf)) {
    $allreplaces=file_get_contents($replacesf);

# Delete comments from file
    $allreplaces=preg_replace("'(//.*|\s+#.*|^#.*)'",'',$allreplaces);

# Legalize multiple white-spaces
    $allreplaces=preg_replace("'(\t+| +)'",' ',$allreplaces);

# Delete unwanted spaces
    $allreplaces=preg_replace("'(^ +| +$)'",'',$allreplaces);

# Delete multiple empty lines
    $allreplaces=preg_replace("'\n+'","\n",$allreplaces);

# Split codepage sections
    $codepages=preg_split("'\n@'",$allreplaces,-1, PREG_SPLIT_NO_EMPTY);
    $cpreg=preg_quote($conf['pdfcp']);

# Find the used codepage
    foreach ($codepages as $codepage) {
      if (preg_match("'" . $cpreg . "'si",$codepage)) {
        $replaces=preg_replace("'" . $cpreg . "\n'si",'',$codepage);
        break;
      }
    }

# Split patterns
    $patterns=preg_split("'\n'",$replaces,-1, PREG_SPLIT_NO_EMPTY);
    foreach ($patterns as $onepair) {
# Split pairs
      $pairarray=preg_split("' '",$onepair);
# Make changes
      $text=str_replace($pairarray[0],$pairarray[1],$text);
    }
  }

  $text = preg_replace("'src=\"http://.*media='i",'src="'.DOKU_INC.'data/media/',$text); # Change links to path of images
  $text = preg_replace("'href=\"http://.*media=.*class=\"media\"'i",'href="" class="media"',$text); # remove picture link address
  $textarr = preg_split("/\n/",$text);

# Find and change linked images
  $linkeds = preg_grep("'<a href=.*<img src=.* /></a>'i",$textarr);
  foreach ( $linkeds as $linked ) {
    $picture = preg_replace("/<a href=.*\">/i",'',$linked);
    $picture = preg_replace("'</a>'i",'',$picture);
    $found = "'".preg_quote($linked)."'";
    $text = preg_replace($found,$picture,$text);
  }
# HTML compatibility -> htmldoc can use <br> instead of <br/>
  $text = str_replace('/>','>',$text);
  $text = str_replace('<table', '<table border="1" ', $text);

#write the string to temporary html-file
  $fp = fopen ($filenameInput, "w") or die ("cant create file");
  fwrite($fp,$text);
  fclose($fp);

#Use embedded fonts if needed
  if ($conf['embedfonts']) {
    $fontparam='--embedfonts';
  } else {
    $fontparam='';
  }

#JPEG compression rate settings
  $jpeg=" --jpeg=" . $conf['jpgrate'];

#PDF compatibility
  $pdf="-t " . $conf['pdfversion'];

#Documentwidth
  $width=" --browserwidth " . $conf['browserwidth'];

#convert using htmldoc
  $command = $conf['htmldocdir'] . "htmldoc " . $pdf . $width . $jpeg . " --charset " . $pdfcp . " --no-title " . $fontparam . " --toctitle \"" . $toctitle . "\" -f " . $filenameOutput . " " . $filenameInput;
  system($command);
  system("exit(0)");

#send to browser
  $filenameOutput=trim($filenameOutput);
  header("Content-type: application/pdf");
  header("Content-Disposition: attachment; filename=dokuwikiexport_" . str_replace(':','_',$_GET["id"]) . ".pdf");
  $fd = @fopen($filenameOutput,"r");
  while(!feof($fd)){
    echo fread($fd,2048);
  }
  fclose($fd);

#clean up temporary files
  system("rm " . $filenameInput);
  system("rm " . $filenameOutput);
}
</code>

So, you can get better, nicer, more varied and more portable documents with these modifications.\\
Cheers, --- //[[peti@szladovics.hu|Peter Szládovics]] 2005-11-17 20:09//

----

====== HTMLDOC Variant Modifications ======

I couldn't get PDF export to handle images and links properly until I replaced the following lines in the inc/common.php (from the Peter's code shown immediately above):

<code php>
  $text = preg_replace("'src=\"http://.*media='i",'src="'.DOKU_INC.'data/media/',$text); # Change links to path of images
  $text = preg_replace("'href=\"http://.*media=.*class=\"media\"'i",'href="" class="media"',$text); # remove picture link address
</code>

with these lines:
<code php>
  $text = preg_replace("'<img src=\"/(.*?)/lib/exe/fetch.php(.*?)media=(.*?)\"(.*?)>'si","<img src=\"http://" . $_SERVER['SERVER_NAME'] . "/\\1/lib/exe/fetch.php?media=\\3\">", $text); # for uploaded images
  $text = preg_replace("'<img src=\"/(.*?)>'si","<img src=\"http://" . $_SERVER['SERVER_NAME'] . "/\\1>", $text); # for built-in images, smileys for example
  $text = preg_replace("'href=\"/(.*?)>'si","href=\"http://" . $_SERVER['SERVER_NAME'] . "/\\1\">", $text); # for internal links
</code>

I also changed the temporary filenames so that static names were not used by
changing (also in Peter's modifications to inc/common.php)
<code php>
  $filenameInput=$dir."input.html";
  $filenameOutput=$dir."output.pdf";
</code>
to this:
<code php>
  $filenameInput=tempnam($dir,"input_");
  $filenameOutput=tempnam($dir,"output_");
</code>

The PDF export is a great and very essential feature.  Thanks! :-)
 --- //[[brian@homelan.org|Brian Dundon]] 2006-3-18 22:44 //

===== Remark on HTMLDOC Variant Modifications =====

I use the code and I get a PDF file. But when I try to open this file, I get "The file is damaged and could not be repaired". The command line parameters are 
  htmldoc -t pdf12 --browserwidth 1280 --jpeg=0 --charset iso-8859-15 --no-title --embedfonts --toctitle "Inhaltsverzeichnis" -f /tmp/output_FSJlGa /tmp/input_GM1H6V
I tried with ''-t pdf14'' and ''%%--%%charset windows-1252'' too, but the same result. I use //htmldoc-1.9.x-r1484//. -- [[w.flamme@web.de|Werner]] 2006-05-13 -- P.S.: with HTMLDOC version 1.8.26 everything works fine. It may be a HTMLDOC issue since even the htmldoc.pdf generated via ''make'' is unreadable.

===== Remark 2 on HTMLDOC Variant Modifications =====

The updated code (from Brian Dundon) used above uses urls to find the images. This variant uses full paths on the unix/linux system to find the images. Urls won't work when logins are required for dokuwiki.

Note: still have to have a look at the line for changing path of built-in images
<code php>
  $text = preg_replace("'src=\"http://.*media='i",'src="'.DOKU_INC.'data/media/',$text); # Change links to path of images
  $text = preg_replace("'href=\"http://.*media=.*class=\"media\"'mi",'href="" class="media"',$text); # remove picture link address
  $textarr = preg_split("/\n/",$text);
 
# Find and change linked images
  $linkeds = preg_grep("'<a href=.*<img src=.* /></a>'i",$textarr);
  foreach ( $linkeds as $linked ) {
    $picture = preg_replace("/<a href=.*\">/i",'',$linked);
    $picture = preg_replace("'</a>'i",'',$picture);
    $found = "'".preg_quote($linked)."'";
    $text = preg_replace($found,$picture,$text);
  }
</code>
with

<code php>
  # find all user uploaded images (and make sure this code doesn't get treated as an image!
  preg_match_all ( '/media=([\./a-z].*?)"/mi', $text, $matches);
  # only use unique elements from array with matches following first parenthesis
  foreach (array_unique($matches[1]) as $match){
    # change namespace into directory syntax
    $newimg = preg_replace ('/:/', '/', $match );
    $text = preg_replace ( "/media=$match/m", "media=$newimg", $text );
  }

  $text = preg_replace('|<img src=".*?/lib/exe/fetch.php.*?media=(.*?)".*?>|mi',"<img src=\"" . DOKU_INC . "data/media/\\1\" />",$text); # for images
  $text = preg_replace('|<img src=".*?/(lib/images/.*?)".*?>|mi',"<img src=\"" . DOKU_INC . "\\1\" />",$text); # for internal images
  $text = preg_replace("|href=\"/(.*?)>|mi","href=\"http://" . $_SERVER['SERVER_NAME'] . "/\\1\">", $text); # for internal links
#you can put this in or leave it out...  $text = preg_replace("|href=\".*?media=.*?class=\"media\"|i",'href="" class="media"',$text); # remove picture link address
  $textarr = preg_split("/\n/",$text);
 
# Find and change linked images
  $linkeds = preg_grep("|<a href=.*?<img src=.*? /></a>|i",$textarr);
  foreach ( $linkeds as $linked ) {
    $picture = preg_replace("|<a href=.*?\">|i",'',$linked);
    $picture = preg_replace("|</a>|i",'',$picture);
   $found = "'".preg_quote($linked)."'";
    $text = preg_replace($found,$picture,$text);
  }
</code>

Little remark: \\
Make sure ''conf/replace.conf'' is called ''conf/replaces.conf'' (with an 's').

Other little update: \\
swap 
<code php>
system("exit(0)");
</code>
with
<code php>
system("exit 0");
</code>
The exit(0) generated an error in the apache error.log file.

PDFexport is very handy for dokuwiki. Thanx for the code.
Hope this remark helps others as well.
 --- //[[fmasselink@gmail.com| F. Masselink]] 2006-10-24 11:33 //


==== Remark of Remark 2 on HTMLDOC Variant Modifications ====
i had a problem with //preg_match_all//, solved by changing the fallowing lines: 
replace in file //inc/common.php// 
<code php>
preg_match_all ( '/media=([\./a-z].*?)"/mi', $text, $matches);
</code>
with
<code php>
preg_match_all ( '/media=([\.\/a-z].*?)"/mi', $text, $matches);
</code>
 --- //[[webmaster@doku.snix.ch| A.Chiapparini]] 2007-01-22 14:47 //

===== Modification for nice URLs =====

I'm using nice URLs with rewrite method and I must add this line for internal pictures to work. It goes to the other ''preg_replace'' lines.

<code php>
$text = preg_replace('|<img src=".*?/_media/(.*?)\?.*?".*?>|mi',"<img src=\"" . DOKU_INC . "data/media/\\1\" />",$text);
</code>

Also I have troubles with filename of downloaded PDF, there is my modification of header line at the end of the pdfmake function. This sends filename containing wiki name and only the name of the wikipage, not the namespace location.

<code php>
header("Content-Disposition: attachment; filename=".str_replace(' ','_',$conf['title']).'-'.end(split('/',$_GET["id"])).".pdf");
</code>
====== HTMLDOC recursive variant ======
My problem was that i needed support for child page export. It therefore choose to modify / hack [[#An_HTMLDOC_variant]] found on this page. Some of the remarks / improvements to [[#An_HTMLDOC_variant]] have also been included.

It will thus perform a recursive export of your current page. This means that any internal links will be followed and converted to PDF too. The internal links should copied to the PDF - meaning that they are click-able like they are in dokuwiki.

  * Follow the first steps of [[#An_HTMLDOC_variant]] (on this page)
  * Then insert this into "inc/common.php":<code php>
function pdfmake($text)
{
  //Variables used to stop the search for child pages
  global $pdfmake_recursion_level;
  global $pdfmake_recursion_current;
  global $pdfmake_links;
  $pdfmake_links = array();
  
  $pdfmake_recursion_level = 30;
  $pdfmake_recursion_current = 0;
 
	// Now search for children.
	$text = pdfmake_children($text);
	
	// And create the pdf
	pdfmake_inner($text);
}
function pdfmake_inner($text){
  global $lang;
  global $conf;
 
  $dir=DOKU_INC."tmp/";
  $filenameInput=$dir."input.html";
  $filenameOutput=$dir."output.pdf";
 
# Convert text and toctitle to destination code-page
  $text=iconv("utf-8",$conf['pdfcp'].'//TRANSLIT',$text);
# Change toctitle if needed
  if ($conf['customtoc']) {
    $toctitle=$conf['customtoc'];
    }
  elseif ($conf['uselangtoc']) {
    $toctitle=$lang['toc'];
    }
  else {
    $toctitle="Table of contents";
  }
  $toctitle=iconv("utf-8",$conf['pdfcp'],$toctitle);
 
# htmldoc compatible name-conversion
  $pdfcp=preg_replace("/windows/i","cp",$conf['pdfcp']);
  $text = preg_replace("'<div class=\"toc\"><div class=\"tocheader\">.*?</div></div>'si",'',$text );
  $text = preg_replace("'<a[^>]*?></a>'si", '', $text );
 
# Execute changes based on replaces.conf
  $replacesf=DOKU_INC . "conf/replaces.conf";
  if ($conf['usecustomreplace'] && file_exists($replacesf)) {
    $allreplaces=file_get_contents($replacesf);
 
# Delete comments from file
    $allreplaces=preg_replace("'(//.*|\s+#.*|^#.*)'",'',$allreplaces);
 
# Legalize multiple white-spaces
    $allreplaces=preg_replace("'(\t+| +)'",' ',$allreplaces);
 
# Delete unwanted spaces
    $allreplaces=preg_replace("'(^ +| +$)'",'',$allreplaces);
 
# Delete multiple empty lines
    $allreplaces=preg_replace("'\n+'","\n",$allreplaces);
 
# Split codepage sections
    $codepages=preg_split("'\n@'",$allreplaces,-1, PREG_SPLIT_NO_EMPTY);
    $cpreg=preg_quote($conf['pdfcp']);
 
# Find the used codepage
    foreach ($codepages as $codepage) {
      if (preg_match("'" . $cpreg . "'si",$codepage)) {
        $replaces=preg_replace("'" . $cpreg . "\n'si",'',$codepage);
        break;
      }
    }
 
# Split patterns
    $patterns=preg_split("'\n'",$replaces,-1, PREG_SPLIT_NO_EMPTY);
    foreach ($patterns as $onepair) {
# Split pairs
      $pairarray=preg_split("' '",$onepair);
# Make changes
      $text=str_replace($pairarray[0],$pairarray[1],$text);
    }
  }
 
  $text = preg_replace("'<img src=\"/(.*?)/lib/exe/fetch.php(.*?)media=(.*?)\"(.*?)>'si","<img src=\"http://" . $_SERVER['SERVER_NAME'] . "/\\1/lib/exe/fetch.php?media=\\3\">", $text); # for uploaded images
  $text = preg_replace("'<img src=\"/(.*?)>'si","<img src=\"http://" . $_SERVER['SERVER_NAME'] . "/\\1>", $text); # for built-in images, smileys for example
  $text = str_replace("href=\"/doku.php?id=", "href=\"#", $text); //correct internal links  
  
  $textarr = preg_split("/\n/",$text);
 
# Find and change linked images
  $linkeds = preg_grep("'<a href=.*<img src=.* /></a>'i",$textarr);
  foreach ( $linkeds as $linked ) {
    $picture = preg_replace("/<a href=.*\">/i",'',$linked);
    $picture = preg_replace("'</a>'i",'',$picture);
    $found = "'".preg_quote($linked)."'";
    $text = preg_replace($found,$picture,$text);
  }
# HTML compatibility -> htmldoc can use <br> instead of <br/>
  $text = str_replace('/>','>',$text);
  $text = str_replace('<table', '<table border="1" ', $text);
 
#write the string to temporary html-file
  $fp = fopen ($filenameInput, "w") or die ("cant create file");
  fwrite($fp,$text);
  fclose($fp);
 
#Use embedded fonts if needed
  if ($conf['embedfonts']) {
    $fontparam='--embedfonts';
  } else {
    $fontparam='';
  }
 
#JPEG compression rate settings
  $jpeg=" --jpeg=" . $conf['jpgrate'];
 
#PDF compatibility
  $pdf="-t " . $conf['pdfversion'];
 
#Documentwidth
  $width=" --browserwidth " . $conf['browserwidth'];
 
#convert using htmldoc
  $command = $conf['htmldocdir'] . "htmldoc " . $pdf . $width . $jpeg . " --charset ". $pdfcp  . " --no-title " . $fontparam . " --toctitle \"" . $toctitle . "\" -f " . $filenameOutput . " " . $filenameInput;

  system($command);
  system("exit(0)");
 
#send to browser
  $filenameOutput=trim($filenameOutput);
  header("Content-type: application/pdf");
  header("Content-Disposition: attachment; filename=dokuwikiexport_" . str_replace(':','_',$_GET["id"]) . ".pdf");
  $fd = @fopen($filenameOutput,"r");
  //Puke on error
  if($fd == false)
  {
    print 'Output file cannot be opened';
    exit;
  }

  while(!feof($fd)){
    echo fread($fd,2048);
  }
  fclose($fd);
 
#clean up temporary files
  system("rm " . $filenameInput);
  system("rm " . $filenameOutput);
}

//search for child pages and render their html
function pdfmake_children($text)
{
  //Extract recusion levels
  global $pdfmake_recursion_level;
  global $pdfmake_recursion_current;
  global $pdfmake_links;
  
  $links = array();
  
 	$pdfmake_recursion_current += 1;
 	//echo 'Current recursion level: ', $pdfmake_recursion_current, '<br>';
 	//will contain all subpages at the end.
 	$innerText = '';	
 		
	//find all links on page
	$regex_pattern = "/<a href=\"(.*)\">(.*)<\/a>/";
	preg_match_all($regex_pattern,$text,$matches);
	
  //The matching pairs will be listed in matches[1]. Sort these matches, so that subnamspaces comes before their parent namespaces.
  //sort($matches[1]);
	for($i=0; $i< count($matches[1]); $i++) {
    //extract the internal dokuwiki id of the subpage. This is needed to perform the rendering
    $link = substr($matches[1][$i], stripos($matches[1][$i],'title=')+7);
   // echo $link, '<br>';
    //Dont add a page which has already been included
    if(!in_array($link, $pdfmake_links)) {
   	  // Call the dokuwiki renderer, if the link does not start with http (then it is not an internal link)
   	  if(substr($link, 0, 4) != 'http') {
	      $innerText .= p_wiki_xhtml($link,'',false);
        
        //Add the link to the collection so it can be sanitized later.
        $pdfmake_links[] = $link;
        $links[] = $link;
	    }
	  }
	}
	//Recurse into the next level of internal links
  if($pdfmake_recursion_current < $pdfmake_recursion_level) {
      //echo "inside recursion<br>";
    $innerText = pdfmake_children($innerText);
  }
  $text = pdfmake_correctlinks($text, $links);
  $innerText = pdfmake_correctlinks($innerText, $links);
  
  // return all the text to caller.
  return $text.$innerText;
}
function pdfmake_correctlinks($text, $links)
{ 
  for($i = 0; $i < count($links); $i++) {
    $link = $links[$i];
    // this $link is the full path to the dokuwiki page. However, in the HTML output, it is only the name after the last ":" which is inserted as id for a heading.
    $text = str_replace($link, substr($link, strrpos($link, ':')+1), $text);
  }
   
  return $text;
}
</code>

Remember that I only tested this on my own sever (on which it works). So expect bugs and / or strange behavior.

===== Bug fixes =====
Here follows a list of fixed bugs
  * 2009-10-31:
    * Fixed a bug in the command line which, on some pages, caused the PDF generation to fail.
    * Fixed a bug with unconvertable UTF8 chars breaking pdf generation (chars like -> and <-)

 --- //[[nicklas.overgaard@gmail.com| Nicklas Overgaard]] 2009-10-31 16:45 GMT+1 //

====== HTMLDOC and OS X ======

The official HTMLDOC packages for OS X are not free. I did find another package at [[http://www.bluem.net/downloads/htmldoc_en/]]

In //inc/common.php// I had to change

<code php>$command = "/usr/bin/htmldoc --no-title -f " . $filenameOutput . " " . $filenameInput;</code>

to 

<code php>$command = "/usr/local/bin/htmldoc --webpage --outfile " . $filenameOutput . " " . $filenameInput;</code>

--- //[[d@mentalfloss.ca|David McCallum]] 2006-06-30 16:09//

----

====== Installing HTMLDOC ======
Installing HTML should be pretty easy, as written above.\\   
But sadly I don't know how to install it on the server (only got uploadpermission by ftp-client).\\
I downloaded htmldoc-1.9.x-r1474 and uploaded the folder but can't find an installer package.\\ 

Does anyone has a simple step by step instruction for beginners?\\
TANK YU

if you are using a Debian or derived operative system than you can use the following command:
  # apt-get install htmldoc

===== Remark of Remark 3 on HTMLDOC Variant Modifications =====
Improvements:
  - not need of a writable directory: use the system temporary directory (automatic)
  - use of unique filename so more people can export file without problems

if you replace
<code php>
  $dir=DOKU_INC."tmp/";
  $filenameInput=$dir."input_";
  $filenameOutput=$dir."output_";
</code>
with
<code php>
  $filenameInput=tempnam('','html');
  $filenameOutput=tempnam('','pdf');
</code>
====== HTMLDOC request ======
I think that will be very useful if you can create a page with the list of wiki page to export and HTMLDOC export all these pages into a PDF file.\\
For example if the wiki page start with a special string (example: HTMLDOC_EXTRACT) then he must extract all pages listed below.\\
Example:
<code>
HTMLDOC_EXTRACT

  * [[tips]]
    * [[tips:pdfexport]]
    * [[tips:browserlanguagedetection]]
</code>

So you can create pages from which you can extract a PDF file based on more wiki pages

**Check the** [[#HTMLDOC_recursive_variant]] it should support the requested feature.

===== Config problem with HTMLDOC variant =====
If you  had added all the ''$conf['abc']'' lines to ''dokuwiki.php'' file and you get an 'Undefined Settings' section on your config page like

|$conf['browserwidth']|No setting metadata.|
|$conf['customtoc'] |No setting metadata.|
|...|...|

you have to declare all value in your ''config.metadata.php''