die();
mb_internal_encoding('utf-8');
/* =========================================================
========================================================= */
# スクリプトを置くファイルの相対PATH
#$dir = '/home/webstats/logs/';
$dir = 'logs/';
$bad_ref = '^http://news\.scenecritique\.com';
$files = get_ref_files();
$refs = get_ref_line($files);
ob_start();
header('Content-Type: text/html; charset=Shift_JIS');
$html = render_html($refs);
$html = mb_convert_encoding($html, 'sjis-win', 'utf-8');
if ( header_gzip() ) {
$html = gzencode($html, 3);
}
echo $html;
header('Content-Length: '. ob_get_length());
ob_end_flush();
exit();
/* =========================================================
render_html
========================================================= */
function render_html(&$refs) {
$pre_month = date('Ym', strtotime('-1 month'));
$now_month = date('Ym', time());
$file = basename(__FILE__);
$html = <<<_EOF_
iRCニュース Referer
{$pre_month}～{$now_month}のりふぁら～
[リファラでソート]
[Hit数でソート]
[時間でソート]
_EOF_;
$i = 0;
foreach ( $refs as $ref ) {
if ( 0 === $i % 2 ) {
$html .= '- ';
}
else {
$html .= '
- ';
}
$html .= ''. htmlspecialchars($ref['ref']).'';
if ( (time() - 60 * 60 * 24 * 1) < $ref['time'] ) {
$html .= ' NEW! ';
}
$html .= ' ('.$ref['cnt'].') ';
$html .= '
'."\n";
$i ++;
}
$html .= <<<_EOF_
_EOF_;
return $html;
}
/* =========================================================
get_ref_line
========================================================= */
function get_ref_line(&$files) {
global $bad_ref;
$refs = array();
$cnt = array();
$url_regex = 'https?://[\w.!~*\'();/?:@&=+$,%#-]+';
foreach ( $files as $file ) {
$fh = fopen($file, 'rb') or die($file.' not found');
while ( !feof($fh) ) {
$line = fgets($fh, 4096);
if ( preg_match('<\[([^]]+)].+"([^"]*)" "[^"]*"$>', $line, $m) ) {
$time = $m[1];
$ref = $m[2];
if ( ! preg_match('<^'.$url_regex.'$>', $ref) ) {
continue;
}
if ( preg_match('<'. $bad_ref. '>i', $ref) ) {
continue;
}
if ( ! isset($cnt[$ref]) ) {
$cnt[$ref] = array();
$cnt[$ref]['cnt'] = 1;
$cnt[$ref]['time'] = str2time($time);
}
else {
$cnt[$ref]['cnt']++;
$time = str2time($time);
if ( $time > $cnt[$ref]['time']) {
$cnt[$ref]['time'] = $time;
}
}
}
}
fclose($fh);
}
$new = array();
foreach ( $cnt as $ref => $val ) {
$new[] = array(
'ref' => $ref,
'cnt' => $val['cnt'],
'time' => $val['time'],
);
}
unset($cnt);
foreach ( $new as $key => $row ) {
$ref_arr[$key] = $row['ref'];
$cnt_arr[$key] = $row['cnt'];
$time_arr[$key] = $row['time'];
}
if ( isset($_GET['sort']) && 'cnt' == $_GET['sort'] ) {
array_multisort($cnt_arr, SORT_NUMERIC, SORT_DESC, $new);
return $new;
}
elseif ( isset($_GET['sort']) && 'time' == $_GET['sort'] ) {
array_multisort($time_arr, SORT_NUMERIC, SORT_DESC, $new);
return $new;
}
else {
array_multisort($ref_arr, SORT_STRING, SORT_ASC, $new);
return $new;
}
}
/* =========================================================
get_ref_files
========================================================= */
function get_ref_files() {
global $dir;
$pre_month = date('Ym', strtotime('-1 month'));
$now_month = date('Ym', time());
$prefix = 'news.scenecritique.com-access_log.';
if (! is_dir($dir)) {
die($dir. ' is not directory!');
}
$files = array();
$now_month_file = $prefix. $now_month;
$pre_month_file = $prefix. $pre_month;
$dh = opendir($dir);
while ( FALSE !== ($file = readdir($dh)) ) {
if ( FALSE !== strpos($file, $now_month_file) ) {
array_push($files, $dir.$file);
}
elseif ( FALSE !== strpos($file, $pre_month_file) ) {
array_push($files, $dir.$file);
}
}
closedir($dh);
if ( empty($files) ) {
die('no data file!');
}
return $files;
}
/* =========================================================
get_ref_files
========================================================= */
function str2time($s) {
$month_arr = array(
'Jan' => '01',
'Feb' => '02',
'Mar' => '03',
'Apr' => '04',
'May' => '05',
'Jun' => '06',
'Jul' => '07',
'Aug' => '08',
'Sep' => '09',
'Oct' => '10',
'Nov' => '11',
'Dec' => '12',
);
preg_match('#^(\d\d)/(\w{3})/(\d{4}):(\d\d):(\d\d):(\d\d)#', $s, $match);
$d = $match[1];
$m = $month_arr[ $match[2] ];
$Y = $match[3];
$H = $match[4];
$i = $match[5];
$s = $match[6];
$str = $Y.'-'.$m.'-'.$d.' '.$H.':'.$i.':'.$s;
return strtotime($str);
}
/* =========================================================
header_gzip
========================================================= */
function header_gzip() {
// use gzencode !
if ( headers_sent() ) {
return FALSE;
}
if ( ! empty($_SERVER['HTTP_ACCEPT_ENCODING']) ) {
if ( preg_match('{((?:x-)?gzip)}i', $_SERVER['HTTP_ACCEPT_ENCODING'], $m) ) {
$gzip = $m[1];
header('Content-Encoding: ' . $gzip);
return TRUE;
}
}
return FALSE;
}
?>