Peter M. Groen / knowledgetree

Commit d83344dcceb4f06e9910d591c966c0f57885369c

Authored by nbm
1 parent 555d3398

Add OpenDocument indexing support (requires unzip) 


git-svn-id: https://kt-dms.svn.sourceforge.net/svnroot/kt-dms/trunk@4989 c91229c3-7414-0410-bfa2-8a42b809f60b
Inline Side-by-side
Showing 2 changed files with 81 additions and 0 deletions
plugins/ktstandard/KTIndexer.php
  View file @d83344d
... ... @@ -16,6 +16,8 @@ class KTIndexerPlugin extends KTPlugin {
16 16 'ktstandard.indexer.triggers.pdf', 'contents/PdfIndexer.php');
17 17 $this->registerTrigger('content', 'transform', 'KTPostscriptIndexerTrigger',
18 18 'ktstandard.indexer.triggers.ps', 'contents/PsIndexer.php');
  19 + $this->registerTrigger('content', 'transform', 'KTOpenDocumentIndexerTrigger',
  20 + 'ktstandard.indexer.triggers.opendocument', 'contents/OpenDocumentIndexer.php');
19 21 }
20 22 }
21 23  
... ...
plugins/ktstandard/contents/OpenDocumentIndexer.php 0 → 100644
  View file @d83344d
  1 +<?php
  2 +
  3 +require_once(KT_DIR . '/plugins/ktstandard/contents/BaseIndexer.php');
  4 +
  5 +class KTOpenDocumentIndexerTrigger extends KTBaseIndexerTrigger {
  6 + var $mimetypes = array(
  7 + 'application/octet-stream' => true,
  8 + 'application/zip' => true,
  9 + 'application/x-zip' => true,
  10 + 'application/vnd.oasis.opendocument.text' => true,
  11 + 'application/vnd.oasis.opendocument.text-template' => true,
  12 + 'application/vnd.oasis.opendocument.presentation' => true,
  13 + 'application/vnd.oasis.opendocument.presentation-template' => true,
  14 + 'application/vnd.oasis.opendocument.spreadsheet' => true,
  15 + 'application/vnd.oasis.opendocument.spreadsheet-template' => true,
  16 + );
  17 +
  18 + function transform() {
  19 + $iMimeTypeId = $this->oDocument->getMimeTypeId();
  20 + $sMimeType = KTMime::getMimeTypeName($iMimeTypeId);
  21 + $sFileName = $this->oDocument->getFileName();
  22 + if (in_array($sMimeType, array('application/octet-stream', 'application/zip', 'application/x-zip'))) {
  23 + $sExtension = KTMime::stripAllButExtension($sFileName);
  24 + $sTable = KTUtil::getTableName('mimetypes');
  25 + $sQuery = "SELECT id FROM $sTable WHERE LOWER(filetypes) = ?";
  26 + $aParams = array($sExtension);
  27 + $id = DBUtil::getOneResultKey(array($sQuery, $aParams), 'id');
  28 + if ($id) {
  29 + $this->oDocument->setMimeTypeId($id);
  30 + $this->oDocument->update();
  31 + }
  32 + }
  33 + parent::transform();
  34 + }
  35 +
  36 + function extract_contents($sFilename, $sTmpFilename) {
  37 + $sUnzipCommand = KTUtil::findCommand("import/unzip", "unzip");
  38 + if (empty($sUnzipCommand)) {
  39 + return;
  40 + }
  41 + $this->sTmpPath = tempnam('/tmp', 'opendocumentextract');
  42 + if ($this->sTmpPath === false) {
  43 + return;
  44 + }
  45 + unlink($this->sTmpPath);
  46 + mkdir($this->sTmpPath, 0700);
  47 +
  48 + $sCmd = array(
  49 + $sUnzipCommand,
  50 + "-q", "-n",
  51 + "-d", $this->sTmpPath,
  52 + $sFilename,
  53 + );
  54 + KTUtil::pexec($sCmd);
  55 +
  56 + $sManifest = sprintf("%s/%s", $this->sTmpPath, "META-INF/manifest.xml");
  57 + if (!file_exists($sManifest)) {
  58 + $this->cleanup();
  59 + return;
  60 + }
  61 + $sContentFile = sprintf("%s/%s", $this->sTmpPath, "content.xml");
  62 + if (!file_exists($sContentFile)) {
  63 + $this->cleanup();
  64 + return;
  65 + }
  66 +
  67 + $sContent = file_get_contents($sContentFile);
  68 + $sContent = preg_replace ("@(</?[^>]*>)+@", " ", $sContent);
  69 +
  70 + $this->cleanup();
  71 + return $sContent;
  72 + }
  73 +
  74 + function cleanup() {
  75 + KTUtil::deleteDirectory($this->sTmpPath);
  76 + }
  77 +}
  78 +
  79 +?>
... ...