<?php
require_once('common.php');
/** @file gs-manual-export.php
* This script transforms the single XML manual file required by the rest of
* the Greenstone manual generation scripts into the series of dokuwiki pages
* that make up a certain manual (as specified by the 'm' argument) in a
* certain language ('l').
*/
if (!parseCLIArguments())
{
printError("Error! Failed to parse arguments...\nUsage: gs-manual-import.php -m [user|install|develop|paper]");
}
// 0. Initialization
if (!isset($_REQUEST['l']) || empty($_REQUEST['l']))
{
$_REQUEST['l'] = 'en';
}
if (!isset($_REQUEST['m']) || empty($_REQUEST['m']))
{
$_REQUEST['m'] = 'user';
//$_REQUEST['m'] = 'install';
//$_REQUEST['m'] = 'develop';
//$_REQUEST['m'] = 'paper';
}
// - validate arguments before we use them (security)
if (!preg_match('/^(develop|install|paper|user)$/',$_REQUEST['m']))
{
printError('Unknown manual type requested: ' . htmlspecialchars($_REQUEST['m']));
}
if (!preg_match('/^(ar|en|es|fr|pt-br|ru)$/',$_REQUEST['l']))
{
printError('Unknown language requested: ' . htmlspecialchars($_REQUEST['l']));
}
echo '<h2>Importing Greenstone Manual XML</h2>' . "\n";
echo '<p><b>XML Source Path:</b> ' . $xml_source_path . '<br/><b>Manual:</b> ' . $_REQUEST['m'] . '<br/><b>Language:</b> ' . $_REQUEST['l'] . "</p>\n<hr/>\n";
echo "<p><b>Frontmatter: </b><br/>\n";
// 1. By-and-large we're going to process all of this in a big state machine
// - the top level page, containing cover page and chapter order information,
// needs to be created last, so we have to store it's information
$manual_metadata = array();
$entity_replacements = array();
$footnotes = array();
$page_order = array();
$page_count = 2;
$looking_for_metadata = '';
$chapter_txt_out = false;
$frontmatter_text = '';
$in_section = false;
$sections_page_name = '';
$in_chapter = false;
$chapter_id = '';
$bullet_depth = 0;
$is_numbered_list = true;
$line_counter = 0;
$in_code = false;
$in_footnotes = false;
$in_numbered_item = 0;
$in_bullet_item = false;
$seen_code_in_item = false;
$in_indent = false;
$is_code_linenumbered = false;
// - construct the path using the information we've been provided as arguments
$xml_file_path = $xml_source_path . '/' . $_REQUEST['l'] . '/' . ucfirst($_REQUEST['m']) . '_' . $_REQUEST['l'] . '.xml';
$xml_in = fopen($xml_file_path, 'r');
if (!$xml_in)
{
printError('Failed to locate top level page for manual');
}
// - we also use this opportunity to read in any footnotes as we'll need to
// move them onto their appropriate page
while (($line = fgets($xml_in)) !== false)
{
if (preg_match('/<Footnote id="(\d+)">/', $line, $matches))
{
$footnote_id = $matches[1];
$text_line = fgets($xml_in);
if (preg_match('/<Text id="([^"]+)">(.+?)<\/Text>/', $text_line, $matches))
{
$footnotes[$footnote_id] = '<!-- id:' . $matches[1] . ' -->' . translateText($matches[2]);
}
// - throw away </Footnote>
fgets($xml_in);
}
}
fclose($xml_in);
// - now reopen to parse it
$xml_in = fopen($xml_file_path, 'r');
if (!$xml_in)
{
printError('Failed to locate top level page for manual');
}
while (($line = getLine($xml_in)) !== false)
{
// - Special Case: lingering code blocks, continue if next line also
// contains code, otherwise we need an extra newline
if ($in_code)
{
$code_text = "\n";
if (strpos($line, '<CodeLine') === false && strpos($line, 'type="code"') === false)
{
$code_text .= '</code>';
// - codeblocks that appear inside numbered lists do not get their own
// newlines as that would split the item. Instead newlines will be
// added when </NumberedItem> encountered.
if (!$is_numbered_list)
{
$code_text .= "\n\n";
}
else
{
$seen_code_in_item = true;
}
$in_code = false;
}
if ($in_chapter)
{
fwrite($chapter_txt_out, $code_text);
}
else
{
$frontmatter_text .= $code_text;
}
}
// - some system metadata to watch for
if (preg_match('/<!ENTITY\s+([^>]+)>/', $line, $matches))
{
$entity = $matches[1];
addMetadata('ENTITY',$entity);
if (preg_match('/([a-z]+)\s+"&#(\d+);"/', $entity, $matches))
{
$entity_replacements[$matches[1]] = $matches[2];
}
}
// - we have an explicit list of cover metadata to watch for
elseif (!$in_section && !$in_chapter && preg_match('/<(Author|Affiliation|Comment|Date|Heading|SupplementaryText|Title|Version)>/', $line, $matches))
{
$looking_for_metadata = $matches[1];
}
elseif (!$in_section && !$in_chapter && preg_match('/<\/(Author|Affiliation|Comment|Date|Heading|SupplementaryText|Title|Version)>/', $line, $matches))
{
$looking_for_metadata = '';
}
// - found metadata we have!
elseif (!empty($looking_for_metadata) && preg_match('/<Text id="([^"]+)">(.+?)<\/Text>/', $line, $matches))
{
$text_id = $matches[1];
$text = '<!-- id:' . $text_id . ' -->' . translateText($matches[2]);
addMetadata($looking_for_metadata, $text);
}
// - bogus metadata found in French version
elseif (!empty($looking_for_metadata) && preg_match('/<Text id="([^"]+)"\/>/', $line, $matches))
{
}
// - any text we encounter outside of both sections and chapters also
// belongs on the cover
elseif (!$in_section && !$in_chapter && !$in_footnotes && preg_match('/<Text id="([^"]+)">(.+?)<\/Text>/', $line, $matches))
{
// (for now I'll assume id's are persistent)
addMetadata('Text', '<!-- id:' . $matches[1] . ' -->' . translateText($matches[2]));
}
// - we will probably encounter the opening section (which is outside of a
// chapter) first, so we have a special case for it
elseif (!$in_chapter && preg_match('/<Section id="([^"]+)">/', $line, $matches))
{
$section_id = $matches[1];
// - if this is the first non-chapter section we have encountered then it
// gets the honor of having the page---that these sections will
// eventually be printed out on---named after it. Typically this should
// be "about_this_manual"
if (empty($frontmatter_text))
{
$sections_page_name = $section_id;
}
$in_section = true;
$title = getTitle($xml_in, 'section:' . $section_id);
if (empty($frontmatter_text))
{
array_unshift($page_order, $sections_page_name . '|' . noComments($title));
}
// - if the title, as is, wouldn't autogenerate the appropriate id, then
// we have to include the id explicitly (as another html comment block)
if ($section_id != generateID($title))
{
$title = '<!-- sid:' . $section_id . ' -->' . $title;
$seen_ids[$section_id] = 1;
}
$frontmatter_text .= '===== ' . $title . ' =====' . "\n\n";
// - whew. Chapter's going to be just as bad though.
}
elseif ($in_section && preg_match('/<\/Section>/', $line))
{
$in_section = false;
}
elseif (preg_match('/<Chapter id="([^"]+)">/', $line, $matches))
{
$chapter_id = $matches[1];
echo "</p>\n<p><b>Import Chapter:</b>" . $chapter_id . '<br/>' . "\n";
$chapter_page_name = $chapter_id;
// - create a new file to store this chapter
$chapter_file_dir = $dokuwiki_path . '/data/pages/' . $_REQUEST['l'] . '/manuals/' . $_REQUEST['m'];
if (!file_exists($chapter_file_dir))
{
mkAllDir($chapter_file_dir, 0755);
}
$chapter_file_path = $chapter_file_dir . '/' . $chapter_page_name . '.txt';
// - backup existing file
if (file_exists($chapter_file_path))
{
$chapter_backup_file_path = $chapter_file_path . '.bak';
rename($chapter_file_path, $chapter_backup_file_path);
}
// - open new file for writing
$chapter_txt_out = fopen($chapter_file_path, 'w');
if (!$chapter_txt_out)
{
printError('Failed to open page file for writing: ' . $chapter_page_name);
}
$in_chapter = true;
$title = getTitle($xml_in, 'chapter: ' . $chapter_id);
fwrite($chapter_txt_out, '====== ' . $title . ' ======' . "\n\n");
array_push($page_order, $chapter_page_name . '|' . noComments($title));
}
elseif ($in_chapter && preg_match('/<\/Chapter>/', $line))
{
fclose($chapter_txt_out);
$chapter_txt_out = false;
$in_chapter = false;
$page_count++;
}
// - section, subsection and part titles within chapter
elseif ($in_chapter && preg_match('/<(Section|Subsection|Part)\sid="([^"]*)">/', $line, $matches))
{
$title_type = $matches[1];
$section_id = $matches[2];
if (empty($section_id))
{
$section_id = generateID(strtolower($title_type));
}
echo '[adding ' . strtolower($title_type) . ': ' . $section_id . '] ';
$header_fix = '';
$title = getTitle($xml_in, 'heading: ' . $title_type);
if ($title_type == 'Section')
{
$header_fix = '=====';
}
if ($title_type == 'Subsection')
{
$header_fix = '====';
}
if ($title_type == 'Part')
{
$header_fix = '===';
// - remove b's and i's
$title = preg_replace('/<\/?(B|I)>/i', '', $title);
}
// - if the title, as is, wouldn't autogenerate the appropriate id, then
// we have to include the id explicitly (as another html comment block)
if ($section_id != generateID($title))
{
$title = '<!-- sid:' . $section_id . ' -->' . $title;
$seen_ids[$section_id] = 1;
}
fwrite($chapter_txt_out, $header_fix . ' ' . $title . ' ' . $header_fix . "\n\n");
}
elseif ($in_chapter && (strpos($line, '</Section') !== false || strpos($line, '</Subsection') !== false || strpos($line, '</Part') !== false))
{
// do nothing for now
}
// - figures (and their titles/captions)
elseif (preg_match('/<Figure id="([^"]+)"(.*?)>/', $line, $matches))
{
$figure_id = $matches[1];
$other_attributes = $matches[2];
echo '[adding figure: ' . $figure_id . "] \n";
// We need the title too
$caption = getTitle($xml_in, 'figure:' . $figure_id);
$caption = translateText(alternateComments($caption));
$txt = "<imgcaption figure_" . $figure_id . '|' . $caption . ' ';
// - we also check the other attributes to see if the XML has requested
// any following codeblock be linenumbered
if (strpos($other_attributes, 'withLineNumber') !== false)
{
$is_code_linenumbered = true;
$txt .= '%!-- withLineNumber --%';
}
$txt .= '>';
if ($in_chapter)
{
fwrite($chapter_txt_out, $txt);
}
else
{
$frontmatter_text .= $txt;
}
}
elseif (strpos($line, '</Figure>') !== false)
{
if ($in_chapter)
{
fwrite($chapter_txt_out, "</imgcaption>\n\n");
}
else
{
$frontmatter_text .= "</imgcaption>\n\n";
}
// - no longer required
$is_code_linenumbered = false;
}
elseif (preg_match('/<Table([^>]*).*?.*?>/', $line, $matches))
{
$attributes = $matches[1];
$table_txt = '';
$table_id = '';
if (preg_match('/id="([^"]+)"/', $attributes, $matches))
{
$table_id = $matches[1];
}
else
{
$table_id = generateID('table');
}
$hidden = false;
if (strpos($attributes, 'class="hidden"') !== false)
{
$hidden = true;
}
echo '[adding table: ' . $table_id . "] \n";
$table_caption = getTitle($xml_in, 'table: ' . $table_id);
if ($hidden)
{
$table_txt .= '<tblcaption table_' . $table_id . '|##HIDDEN##></tblcaption>' . "\n";
}
elseif (empty($table_caption))
{
$table_txt .= '<tblcaption table_' . $table_id . '|##NOCAPTION##></tblcaption>' . "\n";
}
else
{
$table_txt .= '<tblcaption table_' . $table_id . '|' . noComments($table_caption) . '></tblcaption>' . "\n";
}
// - in order to properly capture the table we're going to have to read in
// the whole thing here, and take note of column widths
$have_output_widths = false;
$column_widths = array();
while (strpos($line, '</Table>') === false)
{
// - find the start of a row
while(!empty($line) && strpos($line, '<tr>') === false && strpos($line, '</Table>') === false)
{
$line = getLine($xml_in);
}
if (strpos($line, '<tr>') !== false)
{
$row_txt = '|';
$line = getLine($xml_in);
// - now we read in multiple cells (line starting <th
while (strpos($line, '<th') === 0)
{
if (preg_match('/<th width="(\d+)"\/?>/', $line, $matches))
{
$cell_width = $matches[1];
if (!$have_output_widths)
{
array_push($column_widths, $cell_width);
}
}
// Ignore empty cells
// - adding another case for empty header cells (turned up in es
// version of "From Paper")
if (preg_match('/<th width="\d+"\/>/', $line) || preg_match('/<th width="\d+">.*<\/th>/', $line))
{
$row_txt .= ' |';
}
else
{
$line = getLine($xml_in);
$first = true;
while (strpos($line, '</th>') === false)
{
if (!$first)
{
$row_txt .= '\\\\';
}
// - we can have images or text in our tables
if (preg_match('/<File.*url="images\/([^"]+)".*\/>/', $line, $matches))
{
$payload = $matches[0];
$filename = $matches[1];
$width = 0;
if (preg_match('/width="(\d+)"/', $payload, $matches))
{
$width = $matches[1];
}
$height = 0;
if (preg_match('/height="(\d+)"/', $payload, $matches))
{
$height = $matches[1];
}
$image_txt = handleImage($filename, $width, $height);
$row_txt .= ' ' . $image_txt . ' ';
}
elseif (preg_match('/<Text id="([^"]+)">(.*)/', $line, $matches))
{
$tid = $matches[1];
$txt = $matches[2];
// - multiple line text block
while (strpos($txt, '</Text>') === false)
{
$txt .= getLine($xml_in);
}
$txt = str_replace('</Text>','',$txt);
$row_txt .= ' <!-- id:' . $tid . ' -->' . translateText($txt) . ' ';
}
elseif (preg_match('/<CodeLine>(.*?)<\/CodeLine>/',$line,$matches))
{
$row_txt .= ' \'\'' . translateTableCodeLine($matches[1]) . '\'\' ';
}
elseif (preg_match('/<CodeLine>(.*)/',$line,$matches))
{
$row_txt .= ' \'\'' . translateTableCodeLine($matches[1]) . '\'\' ';
}
elseif (preg_match('/(.*)<\/CodeLine>/',$line,$matches))
{
if (!empty($matches[1]))
{
$row_txt .= ' \'\'' . translateTableCodeLine($matches[1]). '\'\' ';
}
else
{
$row_txt .= ' ';
}
}
// we'll add (bogus) linebreaks
elseif (preg_match('/^\s*<br\s*\/?>\s*$/', $line))
{
$row_txt = ' ';
}
else
{
printError('Warning! Unrecognized element in table: ' . htmlspecialchars($line));
}
$first = false;
// - next line
$line = getLine($xml_in);
}
// - close the cell
$row_txt .= '|';
}
// next!
$line = getLine($xml_in);
}
// - if we haven't already, output the width command
if (!$have_output_widths)
{
$table_txt .= '|< - ' . implode(' ', $column_widths) . ' >|' . "\n";
$have_output_widths = true;
}
$table_txt .= $row_txt . "\n";
// - throw away the closing </tr>
$line = getLine($xml_in);
}
}
$table_txt .= "\n";
if ($in_chapter)
{
fwrite($chapter_txt_out, $table_txt);
}
else
{
$frontmatter_text .= $table_txt;
}
}
// - copy and insert images
elseif (preg_match('/<File width="(\d+)" height="(\d+)" url="images\/([^"]+)"\/>/', $line, $matches))
{
$image_txt = handleImage($matches[3], $matches[1], $matches[2]);
if ($in_chapter)
{
fwrite($chapter_txt_out, $image_txt);
}
else
{
$frontmatter_text .= $image_txt;
}
}
// - bullet lists
elseif (preg_match('/<BulletList>/', $line))
{
echo "[adding bulletlist] \n";
if ($in_bullet_item || $in_numbered_item)
{
if ($in_chapter)
{
fwrite($chapter_txt_out, "\n");
}
elseif ($in_section)
{
$frontmatter_text .= "\n";
}
}
$bullet_depth++;
$is_numbered_list = false;
}
// - numbered lists
elseif (preg_match('/<NumberedList>/', $line))
{
echo "[adding numbered list] \n";
$bullet_depth++;
$is_numbered_list = true;
// - reset this flag that keeps track of whether an item (numbered or
// otherwise) is legitimately split by a code block
$seen_code_in_item = false;
}
elseif (preg_match('/<\/BulletList>/', $line))
{
$bullet_depth--;
if ($bullet_depth == 0)
{
if ($in_chapter)
{
fwrite($chapter_txt_out, "\n");
}
elseif ($in_section)
{
$frontmatter_text .= "\n";
}
$is_numbered_list = false;
}
///cho "[finished bulletlist] ";
}
elseif (preg_match('/<\/NumberedList>/', $line))
{
$bullet_depth--;
if ($bullet_depth == 0)
{
if ($in_chapter)
{
fwrite($chapter_txt_out, "\n");
}
elseif ($in_section)
{
$frontmatter_text .= "\n";
}
$is_numbered_list = false;
}
///cho "[finished numbered list] ";
}
elseif (preg_match('/<NumberedItem>/', $line))
{
$in_numbered_item = 1;
}
elseif (preg_match('/<\/NumberedItem>/', $line))
{
$in_numbered_item = 0;
if ($in_chapter)
{
fwrite($chapter_txt_out, "\n");
}
else
{
$frontmatter_text .= "\n";
}
}
elseif (preg_match('/<Bullet>/', $line))
{
$in_bullet_item = true;
}
elseif (preg_match('/<\/Bullet>/', $line))
{
if ($in_chapter)
{
fwrite($chapter_txt_out, "\n");
}
else
{
$frontmatter_text .= "\n";
}
$in_bullet_item = false;
}
// TEXT HANDLING - this is the main case, but has disappeared into the mire
// of other cases.
elseif (!$in_footnotes && preg_match('/<Text id="([^"]+)">(.+?)$/', $line, $matches))
{
$id = $matches[1];
$str = $matches[2];
// - special case for those text elements split over multiple lines. We
// keep concatenating lines until we find the closing text element or we
// run out of lines!
$another_line = '';
while (strpos($str, '</Text>') === false && ($another_line = getLine($xml_in)) !== false)
{
$str .= ' ' . $another_line;
}
// - note that if we ran out of lines (eof) then we'll break out of this
// block anyway, it's just there won't be a <\Text> at the end of this
// block... despite this being a major validation issue in the XML it
// shouldn't result in this script being vladed
// - now remove the </Text> from the end (hopefully) of str
$str = preg_replace('/<\/Text>\s*/', '', $str);
// - and prepend the id while translating the str into Dokuwiki format
$str = '<!-- id:' . $id . ' -->' . translateText($str);
if ($bullet_depth > 0)
{
if ($is_numbered_list)
{
// - special case for those text elements legimately split in two by
// code blocks. They get no bullet of either type and are 'run-on'
// immediately to the end of the code element in order to prevent
// dokuwiki restarting numbering etc
if ($seen_code_in_item)
{
// - leave str as it is
// - reset flag just incase the item happens to contain another
// code block
$seen_code_in_item = false;
}
else if ($in_numbered_item == 1)
{
$str = '- ' . $str;
}
// - superspecial case for the poorly formatted numberlists that
// contain more than one text block per point. We'll nest them
// as a bullet list as that preserves order, formatting and (I
// hope) meaning.
else
{
if ($in_chapter)
{
fwrite($chapter_txt_out, "\n");
}
else
{
$frontmatter_text .= "\n";
}
$str = ' * ' . $str;
}
$in_numbered_item++;
}
else
{
$str = '* ' . $str;
}
for ($i = 0; $i < $bullet_depth; $i++)
{
$str = ' ' . $str;
}
}
else
{
// Indented text is preceeded by a >
if ($in_indent)
{
$str = '> ' . $str . "\n";
}
else
{
$str .= "\n";
}
}
if ($bullet_depth == 0)
{
$str .= "\n";
}
if ($in_chapter)
{
fwrite($chapter_txt_out, $str);
}
else
{
$frontmatter_text .= $str;
}
}
// - codified text blocks
elseif (preg_match('/<Text\s+type="code"\s+id="([^"]+)"\s*>(.+?)<\/Text>/', $line, $matches))
{
$code_id = $matches[1];
// - determine the appropriate code block prefix
$code_prefix = '';
if (!$in_code)
{
if ($is_code_linenumbered)
{
$code_prefix = '<code 1>';
}
else
{
$code_prefix = '<code>';
}
$in_code = true;
}
$code_txt = $code_prefix . '<!-- id:' . $matches[1] . ' -->' . translateText($matches[2], true);
if ($in_chapter)
{
fwrite($chapter_txt_out, $code_txt);
}
else
{
$frontmatter_text .= $code_txt;
}
}
elseif (preg_match('/<Text\s+id="([^"]+)"\s+type="code"\s*>(.+?)<\/Text>/', $line, $matches))
{
$code_id = $matches[1];
// - determine the appropriate code block prefix
$code_prefix = '';
if (!$in_code)
{
if ($is_code_linenumbered)
{
$code_prefix = '<code 1>';
}
else
{
$code_prefix = '<code>';
}
$in_code = true;
}
$code_txt = $code_prefix . '<!-- id:' . $matches[1] . ' -->' . translateText($matches[2], true);
if ($in_chapter)
{
fwrite($chapter_txt_out, $code_txt);
}
else
{
$frontmatter_text .= $code_txt;
}
}
elseif (preg_match('/<CodeLine>(.*?)$/', $line, $matches))
{
$code_txt = $matches[1];
// - determine the appropriate code block prefix
$code_prefix = '';
if (!$in_code)
{
if ($is_code_linenumbered)
{
$code_prefix = "<code 1>\n";
}
else
{
$code_prefix = "<code>\n";
}
$in_code = true;
}
// - arg. another special case for codelines that span more than one line
// (but I guess <CodeLineButSometimesMoreThanOneLine> is a bit cumbersome
// for an element name, eh?)
$another_line = '';
while (strpos($code_txt, '</CodeLine>') === false && ($another_line = getLine($xml_in)) !== false)
{
$code_txt .= ' ' . $another_line;
}
$code_txt = preg_replace('/<\/CodeLine>\s*/', '', $code_txt);
$code_txt = $code_prefix . translateText($code_txt, true);
if ($in_chapter)
{
fwrite($chapter_txt_out, $code_txt);
}
else
{
$frontmatter_text .= $code_txt;
}
}
// - there are also sometimes empty codelines - which indicate a newline in
// the code listing
elseif (preg_match('/<CodeLine\s*\/>/', $line, $matches))
{
$code_txt = '';
if (!$in_code)
{
$code_txt = "\n";
if ($is_code_linenumbered)
{
$code_txt = "<code 1>\n" . $code_txt;
}
else
{
$code_txt = "<code>\n" . $code_txt;
}
$in_code = true;
}
if ($in_chapter)
{
fwrite($chapter_txt_out, $code_txt);
}
else
{
$frontmatter_text .= $code_txt;
}
}
// - reference to an external XML file
elseif (preg_match('/^\s*&[a-z0-9_]+;\s+$/is', $line))
{
if ($in_chapter)
{
fwrite($chapter_txt_out, $line);
}
else
{
$frontmatter_text .= $line;
}
}
elseif (strpos($line, '<FootnoteList>') !== false)
{
$in_footnotes = true;
}
elseif ($in_footnotes && strpos($line, '</FootnoteList>') !== false)
{
$in_footnotes = false;
}
// Indentation - the closest thing we have is quoting, so we'll use that
elseif (strpos($line, '<Indented>') !== false)
{
$in_indent = true;
}
elseif (strpos($line, '</Indented>') !== false)
{
$in_indent = false;
}
// - pattern of lines to ignore
else if (preg_match('/^(<\?xml version="1.0" encoding="UTF-8"\?>|<\!DOCTYPE Manual \[|\]>|<Bullet>|<\/?Content>|<\/?Footnote|<Manual id=".+?" lang=".+?">|<\/Manual>)/', $line))
{
}
// - we ignore anything else in footnotes too, as they were handled in the
// preprocessing pass
else if ($in_footnotes)
{
}
// - ignore empty lines
else if (preg_match('/^\s*$/', $line))
{
}
// - meh. French versions have random, non-text element, linebreaks floating
// around. Guess I'll honor their formatting even though it's bogus
else if (preg_match('/^\s*<br\s*\/?>\s*$/', $line))
{
if ($in_chapter)
{
fwrite($chapter_txt_out, ' \\\\');
}
else
{
$frontmatter_text .= ' \\\\';
}
}
// - danger Will Robinson!
else
{
echo '<div style="background-color:yellow;"><hr /><b>Warning!</b> Failed to parse line ' . $line_counter . ': |' . htmlspecialchars($line) . "|<hr /></div>\n";
}
}
// 2. We should now have enough metadata to export the cover page
$top_page_path = $dokuwiki_path . '/data/pages/' . $_REQUEST['l'] . '/manuals/' . $_REQUEST['m'] . '.txt';
// - backup any existing file
if (file_exists($top_page_path))
{
$top_page_backup_path = $top_page_path . '.bak';
if(!rename($top_page_path, $top_page_backup_path))
{
printError('Failed to rename existing top page for backup');
}
}
// - and create a handle to the new file
$txt_out = fopen($top_page_path, 'w');
// - write the page (including the tables)
fwrite($txt_out, '====== ' . noComments(ucfirst(getFirstMetadata('Heading'))) . ': ' . noComments(ucfirst(getFirstMetadata('Title'))) . ' (' . strtoupper($_REQUEST['l']) . ') ======' . "\n");
fwrite($txt_out, "\n");
// - *NEW* ability to request imports and exports from within the page
fwrite($txt_out, "<ifauth @admin>\n\n");
fwrite($txt_out, '**Administrator Commands:**' . "\n");
// On second thoughts we probably never want to do this casually, as it boguses
// all history/approval/edit information. Instead I'll leave this as a manual
// process.
fwrite($txt_out, '<!-- Import available at this link - but be warned all current wiki data for this manual will become bogus: http://~~baseurl~~/../../php/gs-manual-import.php?m=' . $_REQUEST['m'] . '&l=' . $_REQUEST['l'] . " -->\n");
fwrite($txt_out, ' * Export manual: [[http://~~baseurl~~/../../php/gs-manual-export.php?m=' . $_REQUEST['m'] . '&l=' . $_REQUEST['l'] . '&v=draft&a=download|draft version]] [[http://~~baseurl~~/../../php/gs-manual-export.php?m=' . $_REQUEST['m'] . '&l=' . $_REQUEST['l'] . '&a=download|approved version]]' . "\n");
fwrite($txt_out, "</ifauth>\n\n");
// - regular metadata
fwrite($txt_out, '<!-- Note: cover page information -->' . "\n");
fwrite($txt_out, '^ Metadata ^ Value ^' . "\n");
$fields = array('Heading','Title','Author','Affiliation','Text','Comment','Version','Date');
foreach ($fields as $field)
{
$values = getMetadata($field);
foreach ($values as $value)
{
fwrite($txt_out, '^ ' . $field . ' | ' . $value . ' |' . "\n");
}
}
fwrite($txt_out, "\n");
// - contents (which also provides order information for exporting)
fwrite($txt_out, '===== Contents =====' . "\n");
fwrite($txt_out, "\n");
fwrite($txt_out, '<!-- Note: The ordering of pages here is used when creating the HTML and PDF versions of the manual -->' . "\n");
foreach ($page_order as $page_info)
{
fwrite($txt_out, ' * [[.:' . $_REQUEST['m'] . ':' . $page_info . ']]' . "\n");
}
// - system metadata
fwrite($txt_out, "<ifauth @admin>\n\n");
fwrite($txt_out, '===== System Metadata =====' . "\n");
fwrite($txt_out, '<!-- Note: configuration options for the manual -->' . "\n");
fwrite($txt_out, '^ Metadata ^ Value ^' . "\n");
$fields = array('ENTITY','SupplementaryText');
foreach ($fields as $field)
{
$values = getMetadata($field);
foreach ($values as $value)
{
fwrite($txt_out, '^ ' . $field . ' | ' . $value . ' |' . "\n");
}
}
fwrite($txt_out, "</ifauth>\n\n");
fwrite($txt_out, "\n");
// - done!
fclose($txt_out);
$page_count++;
// 3. And the 'sections' page, grouping together all the loose sections as
// frontmatter
$frontmatter_page_path = $dokuwiki_path . '/data/pages/' . $_REQUEST['l'] . '/manuals/' . $_REQUEST['m'] . '/' . $sections_page_name . '.txt';
// - backup any existing file
if (file_exists($frontmatter_page_path))
{
$frontmatter_page_backup_path = $frontmatter_page_path . '.bak';
if(!rename($frontmatter_page_path, $frontmatter_page_backup_path))
{
printError('Failed to rename existing frontmatter page for backup');
}
}
// - populate the new frontmatter file
file_put_contents($frontmatter_page_path, $frontmatter_text);
echo "</p>\n<p><b>Complete!</b> Imported " . $page_count . " pages</p><hr/>\n";
echo '<p>Click <a href="' . $dokuwiki_url . '/doku.php?id=' . $_REQUEST['l'] . ':manuals:' . $_REQUEST['m'] . '">here</a> to return to wiki page</p>' . "\n";
exit(0);
/**
*/
function addMetadata($field, $value)
{
global $manual_metadata;
echo '[adding metadata: ' . $field . "] \n";
$values = array();
if (isset($manual_metadata[$field]))
{
$values = $manual_metadata[$field];
}
array_push($values, $value);
$manual_metadata[$field] = $values;
}
/** addMetadata() **/
function getFirstMetadata($field)
{
global $manual_metadata;
$value = '';
if (isset($manual_metadata[$field]))
{
$values = $manual_metadata[$field];
if (!empty($values))
{
$value = $values[0];
}
}
return $value;
}
/** getFirstMetadata() **/
function getMetadata($field)
{
global $manual_metadata;
$values = array();
if (isset($manual_metadata[$field]))
{
$values = $manual_metadata[$field];
}
return $values;
}
/** getMetadata() **/
/**
* Read in the next title element with nested text element and extract the
* title.
*/
function getTitle($xml_in, $element)
{
$title = '';
$in_title_element = false;
// - the first thing in a chapter will be it's title
$title_line = getLine($xml_in);
// - super special case: some language versions don't wrap titles in title
// element, so if the first thing we see is a text, we treat that as the
// title
if (strpos($title_line, '<text') !== false)
{
}
// - super special case: a table with an empty title
if (strpos($title_line, '<Title/>') !== false)
{
return '';
}
if (strpos($title_line, '<Title>') !== false)
{
$in_title_element = true;
$title_line = getLine($xml_in);
}
// - some horribly formed entries have the subtitle first within the title
// element
if (strpos($title_line, '<SubTitle>') !== false)
{
$title_line = getLine($xml_in);
if (preg_match('/<Text id="([^"]+)">(.+?)<\/Text>/', $title_line, $matches))
{
$title = '<!-- id:' . $matches[1] . ' -->' . $matches[2] . ' ' . $title;
}
$title_line = getLine($xml_in);
if (strpos($title_line, '</SubTitle>') === false)
{
printError('Failed to find closing title for: ' . $element);
}
$title_line = getLine($xml_in);
}
// - grab the chapter title now so we can store it in the page ordering
if (preg_match('/<Text id="([^"]+)">(.*?)$/', $title_line, $matches))
{
$id = $matches[1];
$str = $matches[2];
// - special case for text blocks that span multiple lines (as discovered
// in the russian "From Paper"
$another_line = '';
while (strpos($str, '</Text>') === false && ($another_line = getLine($xml_in)) !== false)
{
$str .= ' ' . $another_line;
}
// - now remove </Text>
$str = preg_replace('/<\/Text>\s*/', '', $str);
$title = '<!-- id:' . $id . ' -->' . $str . $title;
}
// - special case for (stoopid) empty titles that use up a text id
elseif (preg_match('/<Text id="([^"]+)"\s*\/>/', $title_line, $matches))
{
$title = '<!-- id:' . $matches[1] . ' -->' . $title;
}
else
{
printError('Failed to find title text for: ' . $element);
}
// - watch for subtitle elements
if ($in_title_element)
{
$title_line = getLine($xml_in);
if (strpos($title_line, '<SubTitle>') !== false)
{
$title_line = getLine($xml_in);
if (preg_match('/<Text id="([^"]+)">(.+?)<\/Text>/', $title_line, $matches))
{
$title = '<!-- id:' . $matches[1] . ' -->' . $matches[2] . ' ' . $title;
}
$title_line = getLine($xml_in);
if (strpos($title_line, '</SubTitle>') === false)
{
printError('Failed to find closing title for: ' . $element);
}
$title_line = getLine($xml_in);
}
if (strpos($title_line, '</Title>') === false)
{
printError('Failed to find closing title for: ' . $element);
}
}
return $title;
}
/** getTitle() **/
function alternateComments($text)
{
$text = str_replace('<!--', '%!--', $text);
$text = str_replace('-->', '--%', $text);
// remove any lurking crossrefs while we are at it
$text = preg_replace('/<CrossRef.*?ref="([^"]+)".*?>/', '\\1', $text);
return $text;
}
function noComments($text)
{
$text = preg_replace('/<i>(.*?)<\/i>/','//\1//',$text);
return preg_replace('/<!--[^>]+-->/', '', $text);
}
function translateTableCodeline($text)
{
///cho "<b>Debug:</b> translateTableCodeLine('" . htmlspecialchars($text) . "')<br />\n";
// Escape the current italics tags to prevent the translate destroying them
$text = str_replace('<i>', '%!--i--%', $text);
$text = str_replace('</i>', '%/i%', $text);
// Translate the text, just decoding the entities
$text = translateText($text, true);
// Now turn the italic tags (escaped) into HTML comments so we remember them
// but they are hidden in the text
$text = str_replace('%!--i--%', '<!--i-->', $text);
$text = str_replace('%/i%', '<!--/i-->', $text);
///cho " => '" . htmlspecialchars($text) . "<br />\n";
return $text;
}
function translateText($text, $entities_only=false)
{
global $entity_replacements;
global $footnotes;
global $in_code;
// - immediate find and protect any legitimate HTML comments in the text
// (so already using encoded entities), otherwise they'll be throughly
// vladed during the following tranforms. This has to be matched with
// changes to the HTMLComments plugin in Dokuwiki to allow the correct
// thing to be displayed to the user.
if (!$in_code)
{
$text = str_replace('<!--', '%!--', $text);
$text = str_replace('-->', '--%', $text);
}
if (!$entities_only)
{
// - replace linking constructs with dokuwiki ones
// - external chapter section crossrefs are easily the worst of all...
while (preg_match('/<CrossRef\s[^>]*external[^>]*\/>/', $text) && preg_match('/<CrossRef\s[^>]*target="Chapter"[^>]*\/>/', $text) && preg_match('/<CrossRef\s+(.*?)\/>/', $text, $matches))
{
$pattern = $matches[0];
$attributes = $matches[1];
$manual_name = '';
if (preg_match('/external="([^"]+)"/', $attributes, $matches))
{
$manual_name = $matches[1];
}
$language = '';
if (preg_match('/lang="([^"]+)"/', $attributes, $matches))
{
$language = $matches[1];
}
$page_id = '';
if (preg_match('/ref="([^"]+)"/', $attributes, $matches))
{
$page_id = $matches[1];
}
if (empty($manual_name) || empty($language) || empty($page_id))
{
printError('Failed to parse external reference: ' . $pattern);
}
// - best we can do is a search within a restricted namespace
$reference = '[[?do=search&id=' . $page_id . ' @' . $language . ':manuals:' . $manual_name . '|' . $page_id . ']]';
$text = str_replace($pattern, $reference, $text);
}
// - chapter crossrefs are tricksie due to needing to know ordering numbers
while (preg_match('/<CrossRef target="Chapter" ref="([^"]+)"\/>/', $text, $matches))
{
$chapter_id = $matches[1];
$page_name = $chapter_id;
$text = preg_replace('/<CrossRef target="Chapter" ref="' . $chapter_id . '"\/>/', '[[.:' . $page_name . '|' . $chapter_id . ']]', $text);
}
// - internal figure and table references
$text = preg_replace('/<CrossRef target="Figure" ref="([^"]+)"\/>/','<imgref figure_\1>', $text);
$text = preg_replace('/<CrossRef target="Table" ref="([^"]+)"\/>/','<tblref table_\1>', $text);
// - simple internal reference
$text = preg_replace('/<CrossRef target="Section" ref="([^"]+)"\/>/', '[[#\1|\1]]', $text);
$text = preg_replace('/<CrossRef target="Subsection" ref="([^"]+)"\/>/', '[[##\1|\1]]', $text);
$text = preg_replace('/<CrossRef target="Part" ref="([^"]+)"\/>/', '[[###\1|\1]]', $text);
// - simple external url
$text = preg_replace('/<Link url="([^"]+)">(.+?)<\/Link>/', '[[\1|\2]]', $text);
// - footnote references are also tricksie as we've had to extract the
// footnotes earlier (during chapter counting)
while (preg_match('/<FootnoteRef id="(\d+)"\/>/', $text, $matches))
{
$footnote_id = $matches[1];
if (!isset($footnotes[$footnote_id]))
{
printError('Unknown footnote referenced: ' . $footnote_id);
}
$footnote = $footnotes[$footnote_id];
$text = preg_replace('/<FootnoteRef id="' . $footnote_id . '"\/>/', '((' . $footnote . '))', $text);
}
// - detect and handle URLs surrounded by <i> tags very carefully (as doku
// will less than helpfully turn them into an external link and screw up
// everything that follows them on the page).
// example: <i>www.microsoft.com</i>
// example: <i>http://nzdl.org/cgi-bin/library</i>
// example: <i>www.yourserver.com</i>
// example: <i>http://www.yourserver.com</i>
// example: <i>http://www.yourserver.com/greenstone</i>
$text = preg_replace('/<i>((?:http\:\/\/)?[a-z]+\.[a-z0-9\-]+\.[a-z0-9\.\-]+(?:\/.*?)?)<\/i>/i','// \1 //', $text);
// - superspecial case for two-part URLs ending in .org (like nzdl.org)
$text = preg_replace('/<i>((?:http\:\/\/)?[a-z0-9\-]+\.org(?:\/.*?)?)<\/i>/i','// \1 //', $text);
// - another superspecial case, this time for URLs on localhost
$text = preg_replace('/<i>((?:http\:\/\/)?localhost(?:\/.*?)?)<\/i>/i','// \1 //', $text);
// - p00p, underlines have the same issue around URLs.
$text = preg_replace('/<u>((?:http\:\/\/)?[a-z]+\.[a-z0-9\-]+\.[a-z0-9\.\-]+(?:\/.*?)?)<\/u>/i','__ \1 __', $text);
$text = preg_replace('/<u>((?:http\:\/\/)?[a-z0-9\-]+\.org(?:\/.*?)?)<\/u>/i','__ \1 __', $text);
$text = preg_replace('/<u>((?:http\:\/\/)?localhost(?:\/.*?)?)<\/u>/i','__ \1 __', $text);
// - replace HTML elements with the dokuwiki style equivilents
$text = preg_replace('/(file|ftp|http):\/\//','\1:%%//%%', $text);
// - restore the double slashes in dokuwiki links
while (preg_match('/\[\[[^\]]*%%[^\]]*\]\]/', $text))
{
$text = preg_replace('/(\[\[[^\]]*)%%([^\]]*\]\])/', '\1\2', $text);
}
$text = str_replace('<b>', '**', $text);
$text = str_replace('</b>', '**', $text);
//$text = str_replace(' <br/>', '\\\\ ', $text);
//$text = str_replace(' <br/>', '\\\\ ', $text);
//$text = str_replace('<br/>', '\\\\ ', $text);
$text = str_replace('<i>', '//', $text);
$text = str_replace('</i>', '//', $text);
$text = str_replace('<u>', '__', $text);
$text = str_replace('</u>', '__', $text);
}
// Decode entities
// - user defined entities (in the manual metadata)
foreach ($entity_replacements as $entity=>$code)
{
$text = str_replace('&' . $entity . ';', html_entity_decode('&#'.$code.';',ENT_NOQUOTES,'UTF-8'), $text);
}
// - standard entities
$text = str_replace('>','>', $text);
$text = str_replace('<','<', $text);
$text = str_replace('&','&', $text);
return $text;
}
/** translateText() **/
function handleImage($filename, $width, $height)
{
global $dokuwiki_path;
global $xml_source_path;
echo '[copying image: ' . $filename . "] \n";
// - copy file into place
$source_path = $xml_source_path . '/' . $_REQUEST['l'] . '/images/' . $filename;
$destination_dir = $dokuwiki_path . '/data/media/' . $_REQUEST['l'] . '/manuals/images/';
if (!file_exists($destination_dir))
{
mkAllDir($destination_dir, 0755);
}
$destination_path = $destination_dir . strtolower($filename);
copy($source_path, $destination_path);
if (!file_exists($destination_path))
{
printError('Failed to copy image file: ' . $filename);
}
// - create the string
$image_txt = '{{..:images:' . strtolower($filename) . '?' . $width . 'x' . $height . '&direct}}';
return $image_txt;
}
function getLine($in)
{
global $line_counter;
$line_counter++;
return fgets($in);
}
function noFormatting($text)
{
$text = str_replace( '<b>', '', $text);
$text = str_replace('</b>', '', $text);
$text = str_replace( '<i>', '', $text);
$text = str_replace('</i>', '', $text);
$text = str_replace('<br/>', '', $text);
return $text;
}
?>