typo3/sysext/indexed_search/mod/index.php

   1 <?php
   2 /***************************************************************
   3 *  Copyright notice
   4 *
   5 *  (c) 1999-2005 Kasper Skaarhoj (kasperYYYY@typo3.com)
   6 *  All rights reserved
   7 *
   8 *  This script is part of the TYPO3 project. The TYPO3 project is
   9 *  free software; you can redistribute it and/or modify
  10 *  it under the terms of the GNU General Public License as published by
  11 *  the Free Software Foundation; either version 2 of the License, or
  12 *  (at your option) any later version.
  13 *
  14 *  The GNU General Public License can be found at
  15 *  http://www.gnu.org/copyleft/gpl.html.
  16 *  A copy is found in the textfile GPL.txt and important notices to the license
  17 *  from the author is found in LICENSE.txt distributed with these scripts.
  18 *
  19 *
  20 *  This script is distributed in the hope that it will be useful,
  21 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  22 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  23 *  GNU General Public License for more details.
  24 *
  25 *  This copyright notice MUST APPEAR in all copies of the script!
  26 ***************************************************************/
  27 /**
  28  * Module: Indexing Engine Overview
  29  *
  30  * @author      Kasper Skårhøj <kasperYYYY@typo3.com>
  31  */
  32 /**
  33  * [CLASS/FUNCTION INDEX of SCRIPT]
  34  *
  35  *
  36  *
  37  *   83: class SC_mod_tools_isearch_index
  38  *   97:     function init()
  39  *  110:     function jumpToUrl(URL)
  40  *  133:     function menuConfig()
  41  *  156:     function main()
  42  *  193:     function printContent()
  43  *
  44  *              SECTION: OTHER FUNCTIONS:
  45  *  216:     function getRecordsNumbers()
  46  *  234:     function tableHead($str)
  47  *  243:     function getPhashStat()
  48  *  278:     function getPhashT3pages()
  49  *  347:     function getPhashExternalDocs()
  50  *  416:     function formatFeGroup($fegroup_recs)
  51  *  432:     function formatCHash($arr)
  52  *  447:     function getNumberOfSections($phash)
  53  *  459:     function getNumberOfWords($phash)
  54  *  471:     function getGrlistRecord($phash)
  55  *  487:     function getNumberOfFulltext($phash)
  56  *  498:     function getPhashTypes()
  57  *  528:     function countUniqueTypes($item_type)
  58  *
  59  * TOTAL FUNCTIONS: 18
  60  * (This index is automatically created/updated by the extension "extdeveval")
  61  *
  62  */
  63
  64
  65 unset($MCONF);
  66 require ("conf.php");
  67 require ($BACK_PATH."init.php");
  68 require ($BACK_PATH."template.php");
  69 $BE_USER->modAccess($MCONF,1);
  70
  71 t3lib_extMgm::isLoaded("indexed_search",1);
  72 require_once(t3lib_extMgm::extPath('indexed_search').'class.indexer.php');
  73
  74
  75
  76 /**
  77  * Backend module providing boring statistics of the index-tables.
  78  *
  79  * @author      Kasper Skaarhoj <kasperYYYY@typo3.com>
  80  * @package TYPO3
  81  * @subpackage tx_indexedsearch
  82  */
  83 class SC_mod_tools_isearch_index {
  84         var $MCONF=array();
  85         var $MOD_MENU=array();
  86         var $MOD_SETTINGS=array();
  87
  88         /**
  89          * document template object
  90          *
  91          * @var noDoc
  92          */
  93         var $doc;
  94
  95         var $include_once=array();
  96         var $content;
  97
  98         /**
  99          * Initialization
 100          *
 101          * @return      [type]          ...
 102          */
 103         function init() {
 104                 global $BE_USER,$LANG,$BACK_PATH,$TCA_DESCR,$TCA,$CLIENT,$TYPO3_CONF_VARS;
 105                 $this->MCONF = $GLOBALS["MCONF"];
 106
 107                 $this->menuConfig();
 108
 109                 $this->doc = t3lib_div::makeInstance("noDoc");
 110                 $this->doc->form='<form action="" method="POST">';
 111                 $this->doc->backPath = $BACK_PATH;
 112                                 // JavaScript
 113                 $this->doc->JScode = '
 114                 <script language="javascript" type="text/javascript">
 115                         script_ended = 0;
 116                         function jumpToUrl(URL) {
 117                                 window.location.href = URL;
 118                         }
 119                 </script>
 120                 ';
 121                 $this->doc->tableLayout = Array (
 122                         "defRow" => Array (
 123                                 "0" => Array('<td valign="top" nowrap>','</td>'),
 124                                 "defCol" => Array('<TD><img src="'.$this->doc->backPath.'clear.gif" width=10 height=1></td><td valign="top" nowrap>','</td>')
 125                         )
 126                 );
 127
 128                 $indexer = t3lib_div::makeInstance('tx_indexedsearch_indexer');
 129                 $indexer->initializeExternalParsers();
 130                 #debug(array_keys($indexer->external_parsers));
 131                 #debug($indexer->internal_log);
 132         }
 133
 134         /**
 135          * [Describe function...]
 136          *
 137          * @return      [type]          ...
 138          */
 139         function menuConfig()   {
 140                 global $BE_USER,$LANG,$BACK_PATH,$TCA_DESCR,$TCA,$CLIENT,$TYPO3_CONF_VARS;
 141
 142                 // MENU-ITEMS:
 143                         // If array, then it's a selector box menu
 144                         // If empty string it's just a variable, that'll be saved.
 145                         // Values NOT in this array will not be saved in the settings-array for the module.
 146                 $this->MOD_MENU = array(
 147                         "function" => array(
 148                                 "stat" => "General statistics",
 149                                 "typo3pages" => "List: TYPO3 Pages",
 150                                 "externalDocs" => "List: External documents",
 151                         )
 152                 );
 153                         // CLEANSE SETTINGS
 154                 $this->MOD_SETTINGS = t3lib_BEfunc::getModuleData($this->MOD_MENU, t3lib_div::_GP("SET"), $this->MCONF["name"], "ses");
 155         }
 156
 157         /**
 158          * [Describe function...]
 159          *
 160          * @return      [type]          ...
 161          */
 162         function main() {
 163                 global $BE_USER,$LANG,$BACK_PATH,$TCA_DESCR,$TCA,$CLIENT,$TYPO3_CONF_VARS;
 164
 165                 $this->content="";
 166                 $this->content.=$this->doc->startPage("Indexing Engine Statistics");
 167
 168                 $menu=t3lib_BEfunc::getFuncMenu(0,"SET[function]",$this->MOD_SETTINGS["function"],$this->MOD_MENU["function"]);
 169
 170                 $this->content.=$this->doc->header("Indexing Engine Statistics");
 171                 $this->content.=$this->doc->spacer(5);
 172                 $this->content.=$this->doc->section('',$menu);
 173
 174                 switch($this->MOD_SETTINGS["function"]) {
 175                         case "stat":
 176                                 $this->content.=$this->doc->section('Records',$this->doc->table($this->getRecordsNumbers()),0,1);
 177                                 $this->content.=$this->doc->spacer(15);
 178                 //              $this->content.=$this->doc->section('index_phash STATISTICS',$this->doc->table($this->getPhashStat()),1);
 179                 //              $this->content.=$this->doc->spacer(15);
 180                                 $this->content.=$this->doc->section('index_phash TYPES',$this->doc->table($this->getPhashTypes()),1);
 181                                 $this->content.=$this->doc->spacer(15);
 182                         break;
 183                         case "externalDocs":
 184                                 $this->content.=$this->doc->section('External documents',$this->doc->table($this->getPhashExternalDocs()),0,1);
 185                                 $this->content.=$this->doc->spacer(15);
 186                         break;
 187                         case "typo3pages":
 188                                 $this->content.=$this->doc->section('TYPO3 Pages',$this->doc->table($this->getPhashT3pages()),0,1);
 189                                 $this->content.=$this->doc->spacer(15);
 190                         break;
 191                 }
 192         }
 193
 194         /**
 195          * [Describe function...]
 196          *
 197          * @return      [type]          ...
 198          */
 199         function printContent() {
 200                 $this->content.=$this->doc->endPage();
 201                 echo $this->content;
 202         }
 203
 204
 205
 206
 207
 208
 209
 210
 211
 212
 213         /***************************
 214          *
 215          * OTHER FUNCTIONS:
 216          *
 217          ***************************/
 218
 219         /**
 220          * @return      [type]          ...
 221          */
 222         function getRecordsNumbers()    {
 223                 $tables=explode(",","index_phash,index_words,index_rel,index_grlist,index_section,index_fulltext");
 224                 $recList=array();
 225                 reset($tables);
 226                 while(list(,$t)=each($tables))  {
 227                         $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*)', $t, '');
 228                         $row = $GLOBALS['TYPO3_DB']->sql_fetch_row($res);
 229                         $recList[] = array($this->tableHead($t), $row[0]);
 230                 }
 231                 return $recList;
 232         }
 233
 234         /**
 235          * [Describe function...]
 236          *
 237          * @param       [type]          $str: ...
 238          * @return      [type]          ...
 239          */
 240         function tableHead($str)        {
 241                 return "<strong>".$str.":&nbsp;&nbsp;&nbsp;</strong>";
 242         }
 243
 244         /**
 245          * [Describe function...]
 246          *
 247          * @return      [type]          ...
 248          */
 249         function getPhashStat() {
 250                 $recList = array();
 251
 252                         // TYPO3 pages, unique
 253                 $items = array();
 254                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*),phash', 'index_phash', 'data_page_id!=0', 'phash_grouping,pcount,phash');
 255                 while($row = $GLOBALS['TYPO3_DB']->sql_fetch_row($res)) {
 256                         $items[] = $row;
 257                 }
 258                 $recList[] = array($this->tableHead("TYPO3 pages"), count($items));
 259
 260                         // TYPO3 pages:
 261                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*)', 'index_phash', 'data_page_id!=0');
 262                 $row = $GLOBALS['TYPO3_DB']->sql_fetch_row($res);
 263                 $recList[] = array($this->tableHead("TYPO3 pages, raw"), $row[0]);
 264
 265                         // External files, unique
 266                 $items = array();
 267                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*),phash', 'index_phash', 'data_filename!=\'\'', 'phash_grouping');
 268                 $row = $GLOBALS['TYPO3_DB']->sql_fetch_row($res);
 269                 $recList[] = array($this->tableHead("External files"), $row[0]);
 270
 271                         // External files
 272                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*)', 'index_phash', 'data_filename!=\'\'');
 273                 $row = $GLOBALS['TYPO3_DB']->sql_fetch_row($res);
 274                 $recList[] = array($this->tableHead("External files, raw"), $row[0]);
 275
 276                 return $recList;
 277         }
 278
 279         /**
 280          * [Describe function...]
 281          *
 282          * @return      [type]          ...
 283          */
 284         function getPhashT3pages()      {
 285                 $recList[]=array(
 286                         $this->tableHead("id/type"),
 287                         $this->tableHead("Title"),
 288                         $this->tableHead("Size"),
 289                         $this->tableHead("Words"),
 290                         $this->tableHead("mtime"),
 291                         $this->tableHead("Indexed"),
 292                         $this->tableHead("Updated"),
 293                         $this->tableHead("Parsetime"),
 294                         $this->tableHead("#sec/gr/full"),
 295                         $this->tableHead("#sub"),
 296                         $this->tableHead("Lang"),
 297                         $this->tableHead("cHash"),
 298                         $this->tableHead("phash")
 299                 );
 300
 301                         // TYPO3 pages, unique
 302                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*) AS pcount,index_phash.*', 'index_phash', 'data_page_id!=0', 'phash_grouping,phash,cHashParams,data_filename,data_page_id,data_page_reg1,data_page_type,data_page_mp,gr_list,item_type,item_title,item_description,item_mtime,tstamp,item_size,contentHash,crdate,parsetime,sys_language_uid,item_crdate,externalUrl,recordUid,freeIndexUid', 'data_page_id');
 303                 while($row = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res))       {
 304
 305                         $cHash = count(unserialize($row["cHashParams"])) ? $this->formatCHash(unserialize($row["cHashParams"])) : "";
 306                         $grListRec = $this->getGrlistRecord($row["phash"]);
 307                         $recList[] = array(
 308                                 $row["data_page_id"].($row["data_page_type"]?"/".$row["data_page_type"]:""),
 309                                 t3lib_div::fixed_lgd($row["item_title"],30),
 310                                 t3lib_div::formatSize($row["item_size"]),
 311                                 $this->getNumberOfWords($row["phash"]),
 312                                 t3lib_BEfunc::datetime($row["item_mtime"]),
 313                                 t3lib_BEfunc::datetime($row["crdate"]),
 314                                 ($row["tstamp"]!=$row["crdate"] ? t3lib_BEfunc::datetime($row["tstamp"]) : ""),
 315                                 $row["parsetime"],
 316                                 $this->getNumberOfSections($row["phash"])."/".$grListRec[0]["pcount"]."/".$this->getNumberOfFulltext($row["phash"]),
 317                                 $row["pcount"]."/".$this->formatFeGroup($grListRec),
 318                                 $row["sys_language_uid"],
 319                                 $cHash,
 320                                 $row["phash"]
 321                         );
 322
 323                         if ($row["pcount"]>1)   {
 324                                 $res2 = $GLOBALS['TYPO3_DB']->exec_SELECTquery('index_phash.*', 'index_phash', 'phash_grouping='.intval($row['phash_grouping']).' AND phash!='.intval($row['phash']));
 325                                 while($row2 = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res2))     {
 326                                         $grListRec = $this->getGrlistRecord($row2["phash"]);
 327                                         $recList[] = array(
 328                                                 "",
 329                                                 "",
 330                                                 t3lib_div::formatSize($row2["item_size"]),
 331                                                 $this->getNumberOfWords($row2["phash"]),
 332                                                 t3lib_BEfunc::datetime($row2["item_mtime"]),
 333                                                 t3lib_BEfunc::datetime($row2["crdate"]),
 334                                                 ($row2["tstamp"]!=$row2["crdate"] ? t3lib_BEfunc::datetime($row2["tstamp"]) : ""),
 335                                                 $row2["parsetime"],
 336                                                 $this->getNumberOfSections($row2["phash"])."/".$grListRec[0]["pcount"]."/".$this->getNumberOfFulltext($row2["phash"]),
 337                                                 "-/".$this->formatFeGroup($grListRec),
 338                                                 "",
 339                                                 "",
 340                                                 $row2["phash"]
 341                                         );
 342                                 }
 343                         }
 344                 }
 345                 return $recList;
 346         }
 347
 348         /**
 349          * [Describe function...]
 350          *
 351          * @return      [type]          ...
 352          */
 353         function getPhashExternalDocs() {
 354                 $recList[]=array(
 355                         $this->tableHead("Filename"),
 356                         $this->tableHead("Size"),
 357                         $this->tableHead("Words"),
 358                         $this->tableHead("mtime"),
 359                         $this->tableHead("Indexed"),
 360                         $this->tableHead("Updated"),
 361                         $this->tableHead("Parsetime"),
 362                         $this->tableHead("#sec/gr/full"),
 363                         $this->tableHead("#sub"),
 364                         $this->tableHead("cHash"),
 365                         $this->tableHead("phash"),
 366                         $this->tableHead("Path")
 367                 );
 368
 369                         // TYPO3 pages, unique
 370                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*) AS pcount,index_phash.*', 'index_phash', 'item_type>0', 'phash_grouping,phash,cHashParams,data_filename,data_page_id,data_page_reg1,data_page_type,data_page_mp,gr_list,item_type,item_title,item_description,item_mtime,tstamp,item_size,contentHash,crdate,parsetime,sys_language_uid,item_crdate,externalUrl,recordUid,freeIndexUid', 'item_type');
 371                 while($row = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res))       {
 372
 373                         $cHash = count(unserialize($row["cHashParams"])) ? $this->formatCHash(unserialize($row["cHashParams"])) : "";
 374                         $grListRec = $this->getGrlistRecord($row["phash"]);
 375                         $recList[]=array(
 376                                 t3lib_div::fixed_lgd($row["item_title"],30),
 377                                 t3lib_div::formatSize($row["item_size"]),
 378                                 $this->getNumberOfWords($row["phash"]),
 379                                 t3lib_BEfunc::datetime($row["item_mtime"]),
 380                                 t3lib_BEfunc::datetime($row["crdate"]),
 381                                 ($row["tstamp"]!=$row["crdate"] ? t3lib_BEfunc::datetime($row["tstamp"]) : ""),
 382                                 $row["parsetime"],
 383                                 $this->getNumberOfSections($row["phash"])."/".$grListRec[0]["pcount"]."/".$this->getNumberOfFulltext($row["phash"]),
 384                                 $row["pcount"],
 385                                 $cHash,
 386                                 $row["phash"],
 387                                 t3lib_div::fixed_lgd($row["data_filename"],100)
 388                         );
 389
 390                         if ($row["pcount"]>1)   {
 391                                 $res2 = $GLOBALS['TYPO3_DB']->exec_SELECTquery('index_phash.*', 'index_phash', 'phash_grouping='.intval($row['phash_grouping']).' AND phash!='.intval($row['phash']));
 392                                 while($row2 = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res2))     {
 393                                         $cHash = count(unserialize($row2["cHashParams"])) ? $this->formatCHash(unserialize($row2["cHashParams"])) : "";
 394                                         $grListRec = $this->getGrlistRecord($row2["phash"]);
 395                                         $recList[]=array(
 396                                                 "",
 397                                                 "",
 398                                                 $this->getNumberOfWords($row2["phash"]),
 399                                                 "",
 400                                                 t3lib_BEfunc::datetime($row2["crdate"]),
 401                                                 ($row2["tstamp"]!=$row2["crdate"] ? t3lib_BEfunc::datetime($row2["tstamp"]) : ""),
 402                                                 $row2["parsetime"],
 403                                                 $this->getNumberOfSections($row2["phash"])."/".$grListRec[0]["pcount"]."/".$this->getNumberOfFulltext($row2["phash"]),
 404                                                 "",
 405                                                 $cHash,
 406                                                 $row2["phash"],
 407                                                 ""
 408                                         );
 409                                 }
 410                         }
 411         //              debug($row);
 412                 }
 413                 return $recList;
 414         }
 415
 416         /**
 417          * [Describe function...]
 418          *
 419          * @param       [type]          $fegroup_recs: ...
 420          * @return      [type]          ...
 421          */
 422         function formatFeGroup($fegroup_recs)   {
 423                 reset($fegroup_recs);
 424                 $str = array();
 425                 while(list(,$row)=each($fegroup_recs))  {
 426                         $str[] = $row["gr_list"]=="0,-1" ? "NL" : $row["gr_list"];
 427                 }
 428                 arsort($str);
 429                 return implode("|",$str);
 430         }
 431
 432         /**
 433          * [Describe function...]
 434          *
 435          * @param       [type]          $arr: ...
 436          * @return      [type]          ...
 437          */
 438         function formatCHash($arr)      {
 439                 reset($arr);
 440                 $list=array();
 441                 while(list($k,$v)=each($arr))   {
 442                         $list[]=$k."=".$v;
 443                 }
 444                 return implode("<BR>",$list);
 445         }
 446
 447         /**
 448          * [Describe function...]
 449          *
 450          * @param       [type]          $phash: ...
 451          * @return      [type]          ...
 452          */
 453         function getNumberOfSections($phash)    {
 454                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*)', 'index_section', 'phash='.intval($phash));
 455                 $row = $GLOBALS['TYPO3_DB']->sql_fetch_row($res);
 456                 return $row[0];
 457         }
 458
 459         /**
 460          * [Describe function...]
 461          *
 462          * @param       [type]          $phash: ...
 463          * @return      [type]          ...
 464          */
 465         function getNumberOfWords($phash)       {
 466                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*)', 'index_rel', 'phash='.intval($phash));
 467                 $row = $GLOBALS['TYPO3_DB']->sql_fetch_row($res);
 468                 return $row[0];
 469         }
 470
 471         /**
 472          * [Describe function...]
 473          *
 474          * @param       [type]          $phash: ...
 475          * @return      [type]          ...
 476          */
 477         function getGrlistRecord($phash)        {
 478                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('index_grlist.*', 'index_grlist', 'phash='.intval($phash));
 479                 $allRows = array();
 480                 while($row = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res))       {
 481                         $row["pcount"] = $GLOBALS['TYPO3_DB']->sql_num_rows($res);
 482                         $allRows[] = $row;
 483                 }
 484                 return $allRows;
 485         }
 486
 487         /**
 488          * [Describe function...]
 489          *
 490          * @param       [type]          $phash: ...
 491          * @return      [type]          ...
 492          */
 493         function getNumberOfFulltext($phash)    {
 494                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*)', 'index_fulltext', 'phash='.intval($phash));
 495                 $row = $GLOBALS['TYPO3_DB']->sql_fetch_row($res);
 496                 return $row[0];
 497         }
 498
 499         /**
 500          * [Describe function...]
 501          *
 502          * @return      [type]          ...
 503          */
 504         function getPhashTypes()        {
 505                 $recList=array();
 506
 507                 // Types:
 508                 $Itypes = array(
 509                         "html" => 1,
 510                         "htm" => 1,
 511                         "pdf" => 2,
 512                         "doc" => 3,
 513                         "txt" => 4
 514                 );
 515
 516                 $revTypes=array_flip($Itypes);
 517                 $revTypes[0]="TYPO3 page";
 518
 519                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*),item_type', 'index_phash', '', 'item_type', 'item_type');
 520                 while($row = $GLOBALS['TYPO3_DB']->sql_fetch_row($res)) {
 521                         $iT = $row[1];
 522                         $recList[] = array($this->tableHead($revTypes[$iT]." ($iT)"), $this->countUniqueTypes($iT)."/".$row[0]);
 523                 }
 524
 525                 return $recList;
 526         }
 527
 528         /**
 529          * [Describe function...]
 530          *
 531          * @param       [type]          $item_type: ...
 532          * @return      [type]          ...
 533          */
 534         function countUniqueTypes($item_type)   {
 535                         // TYPO3 pages, unique
 536                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery('count(*)', 'index_phash', 'item_type='.intval($item_type), 'phash_grouping');
 537                 $items = array();
 538                 while($row = $GLOBALS['TYPO3_DB']->sql_fetch_row($res)) {
 539                         $items[] = $row;
 540                 }
 541                 return count($items);
 542         }
 543 }
 544
 545 // Include extension?
 546 if (defined("TYPO3_MODE") && $TYPO3_CONF_VARS[TYPO3_MODE]["XCLASS"]["ext/indexed_search/mod/index.php"])        {
 547         include_once($TYPO3_CONF_VARS[TYPO3_MODE]["XCLASS"]["ext/indexed_search/mod/index.php"]);
 548 }
 549
 550
 551
 552
 553
 554
 555
 556
 557
 558
 559
 560
 561 // Make instance:
 562 $SOBE = t3lib_div::makeInstance("SC_mod_tools_isearch_index");
 563 $SOBE->init();
 564 $SOBE->main();
 565 $SOBE->printContent();
 566 ?>