.
*/
// FIXME: TODO: Change path above to production installation location of perl
// when we deploy, Brad Neuberg, bkn3@columbia.edu
//
// From:
// search.cgi v0.4
// by Ralf Muehlen
//
// slight alterations by Brad Neuberg, bkn3@columbia.edu
//
// ported from perl to php by tracey, Oct 2005
//fixxx require_once '/petabox/setup.inc';
//I think this fixxx refers to the need to set DOCUMENT_ROOT below -- mang
if (strpos($_SERVER["REQUEST_URI"], "/~mang") === 0) { // Serving out of home dir
$_SERVER['DOCUMENT_ROOT']='/home/mang/petabox/www/sf';
require_once '/home/mang/petabox/setup.inc';
} else if (strpos($_SERVER["REQUEST_URI"], "/~testflip") === 0) { // Serving out of home dir
$_SERVER['DOCUMENT_ROOT']='/home/testflip/petabox/www/sf';
require_once '/home/testflip/petabox/setup.inc';
} else {
$_SERVER['DOCUMENT_ROOT']='/petabox/www/sf';
require_once '/petabox/setup.inc';
}
ini_set("memory_limit","200M"); // XML can be big, esp. brittanica (100MB)
/////// SETUP /////////////////////////////////////////////////////////////
$debug_level = 0; // 0=least, 3=most debugging info
$num_pre = 3; // context words before search term
$num_post = 9; // context words after search term
// defaults for testing (when no args given)
$url ='http://ia300202.us.archive.org/0/items/englishbookbindings00davenuoft/englishbookbindings00davenuoft_djvuxml.xml';
$term = "history";
$format = "XML";
$callback = false;
/////// SETUP /////////////////////////////////////////////////////////////
// fixxx prolly should escapesystemcall() these...
if (isset($_GET['url']))
$url = $_GET['url'];
if (isset($_GET['term']))
$term = $_GET['term'];
if (isset($_GET['format']))
$format = $_GET['format'];
if (isset($_GET['callback']))
$callback = $_GET['callback'];
//$url='http://homeserver.hq.archive.org/metafetch/thespy00cooparch_djvu.xml';
//$url='http://homeserver.hq.archive.org/metafetch/oldchristmas00irviarch_djvu.xml';
//$url = 'http://homeserver.hq.archive.org/metafetch/intlepisode00jamearch_djvu.xml';
if ($format == "XML")
{
// This is kinda weird (confession!) but allows existing calls to "fatal()"
// to throw an exception instead of dumping HTML to stdout/browser!
$GLOBALS['fatal-exceptions']=1;
}
try
{
// pageFiles was added to keep track of on what page each search match was
// found, Brad Neuberg, bkn3@columbia.edu
// pageHeights and pageWidths was added to track the size of each page so that
// we can send it over to the client; this is necessary for scaling the images
// for search, Brad Neuberg, bkn3@columbia.edu
$pages = array();
$pageFiles = array();
$pageHeights = array();
$pageWidths = array();
$time0 = microtime(true);
$timestamp = date('Y-m-d H:i:s');
$pid = posix_getpid();
debug_msg("Invoked at ".$time0."=$timestamp under UID ".posix_getuid(),2);
//////////////////////////////////////
debug_msg("query: ".$_SERVER['QUERY_STRING'],3);
$term = preg_replace('/[^A-Za-z0-9 ]/', ' ', $term); // legal characters
$terms = explode(' ',$term);
debug_msg("url,term,format: $url,".var_export($terms,true).",$format",3);
if ($format == "HTML")
{
echo "Search Searching
";
$tag_pre = '';
$tag_post = '';
}
else if ($format == "XML")
{
if (false === $callback) {
header('Content-type: text/xml');
} else {
header('application/x-javascript');
}
$tag_pre = '';
$tag_post = '';
}
else
{
fatal("Unknown format request. ");
}
// Ensure file is readable
checkPrivs($url);
// This looks like where we load the djvu.xml - $$$ and rapidly exhaust memory for large books such as OED
if (!($document = file_get_contents($url)))
fatal("could not load $url");
$time1 = microtime(true) - $time0;
//// Pass 1 - build up page* arrays with xml fragments corresponding to matches
$pagenumber=0;
foreach (explode('', $document) as $page)
{
$pagenumber++;
if (matches_terms($page, $terms) &&
// 2nd clause here is to ensure that we aren't matching in the end
// of the overall XML document -- thus we ensure that OBJECT tag starts
// in the chunk we just were handed. (traceyj)
strstr($page, '|', $page, $match))
fatal("page value not set on page number $pagenumber in $page!");
$pageFile = $match[1];
// extract the page width and height, Brad Neuberg, bkn3@columbia.edu
if (!preg_match('/width="([^"]*)"/', $page, $match))
fatal("page width not set!");
$pageWidth = $match[1];
if (!preg_match('/height="([^"]*)"/', $page, $match))
fatal("page height not set!");
$pageHeight = $match[1];
$page_new='';
foreach (explode('',$page) as $token)
{
if (matches_terms($token, $terms))
{
list($junk, $keep) = explode('/','', $token); //mark-up
$token = preg_replace('/[\&\#\d+;]/', ' ', $token);//non-ascii chars
$token = preg_replace('/\s+/', ' ', $token); //white space
}
$page_new .= $token;
}
$page_new =
preg_replace('|.*((\W\w*){'.$num_pre.'}'.$tag_pre.')|',"$1",$page_new);
$page_new =
preg_replace('/('.$tag_post.'(\w*\W){'.$num_post.'}).*/',"$1",$page_new);
// added to keep track of the page we are on
// Brad Neuberg, bkn3@columbia.edu
$pageFiles [$pagenumber] = $pageFile;
$pages [$pagenumber] = $page_new;
// added to keep track of page widths and heights
$pageWidths [$pagenumber] = $pageWidth;
$pageHeights[$pagenumber] = $pageHeight;
}
}
$time2 = microtime(true) - $time1;
//// Pass 2 - generate output from previously built arrays
if ($format == "HTML")
{
echo "Found ".count($pages)." pages containing $tag_pre";
print_r($terms);
echo "$tag_post. \n";
foreach ($pages as $index => $page)
{
echo "
\n";
echo $tag_pre . "Processed document in $time2 ms.$tag_post
\n";
echo $tag_pre . "Printed document in $time3 ms.$tag_post
\n";
echo "\n";
}
else if ($format == "XML")
{
$xml = "";
$xml .= ''."\n";
// Added to prevent Internet Explorer from adding default XML stylesheet,
// which messes up processing, Brad Neuberg, bkn3@columbia.edu
$xml .= ''."\n";//fixxx
$xml .= '';
foreach ($pages as $index => $page)
{
$xml .= "\n";
$xml .= "\n";
$xml .= $page;
$xml .= "\n";
$xml .= "\n";
}
$xml .= "\n";
if (false === $callback) {
// The XML contains the page numbers from the DJVU XML. We must remap them to flipbook indices
// since the flipbook indices are monotonically increasing generated from the pages with
// addToAccessFormats true (maybe)
$fsm = FlipSearchMap::buildSearchMap($url);
echo $fsm->remapSearch($xml);
} else {
$patterns[0] = '/\n/';
$patterns[1] = "/\'/";
$replac[0] = '';
$replac[1] = ''';
// We don't have FlipSearchMap remap since we have our own mapping between
// scandata.xml leaf numbers and BR indices that happens in BRSearchCallback
echo "$callback('". preg_replace($patterns, $replac, $xml)."');";
}
//echo $xml;
}
//////
debug_msg("Done and exiting!",2);
exit;
//////
}
catch (Exception $e)
{
// an internal method call invoked "fatal()"...
XML::resultMessage('error','internal_error', $e->getMessage());
}
function matches_terms(&$text, // search space
&$terms)// array of search terms
{
foreach ($terms as $term)
{
if (preg_match("/$term/i", $text))
return true;
}
return false;
}
function debug_msg($msg, $level)
{
global $debug_level;
global $pid;
global $format;
if ($level <= $debug_level)
{
if ($format == "XML")
echo "\n";
else
echo "FILL ($pid):$level: $msg \n";
}
}
function checkPrivs($filename) {
if (!is_readable($filename)) {
header('HTTP/1.1 403 Forbidden');
exit(0);
}
}
?>