Configure media filters

From DSpace Wiki

[edit] Files:

• [dspace]/bin/filter-media
• [dspace]/config/dspace.cfg
• A scheduling program (e.g. Unix cron or Windows Scheduled Tasks)

[edit] Instructions:

1. In DSpace, “media filters” are what control both full-text indexing and automated creation of thumbnail images. Both can be scheduled by calling the filter-media script (which in turn calls the org.dspace.app.mediafilter.MediaFilterManager class).
2. The list of all currently enabled “media filters” is available in your dspace.cfg configuration file under the section labeled:

   #### Media Filter plugins (through PluginManager) ####

3. If you wish to disable or enable a specific “media filter”, you can remove or add them from the sequence list in that section: plugin.sequence.org.dspace.app.mediafilter.MediaFilter = org.dspace.app.mediafilter.PDFFilter, org.dspace.app.mediafilter.HTMLFilter, org.dspace.app.mediafilter.WordFilter, org.dspace.app.mediafilter.JPEGFilter
4. (Make sure that each separate line ends with a backslash \ character! A list of what each “media filter” does is available in the Notes section)
5. For Linux or Mac OSX, you can schedule the filter-media shell script to run by adding a cron entry similar to the following to the crontab for the user who installed DSpace:

   0 2 * * * ''[dspace]''/bin/filter-media

6. (The above entry would schedule filter-media to run nightly at 2am. You would need to change [dspace] to the full path of your DSpace installation directory.)
7. For Windows, you will be unable to use the filter-media shell script. Instead, you should use Windows Scheduled Tasks to schedule the following command to run at the appropriate time of day:

   ''[dspace]''/bin/dsrun.bat org.dspace.app.mediafilter.MediaFilterManager

8. (The above command should appear on a single line.)

[edit] Notes:

• Below is a listing of all currently available Media Filters, and what they actually do:
  • HTMLFilter – extracts the full text of HTML documents for full text indexing.
  • JPEGFilter – creates thumbnail images of GIF, JPEG and PNG files
  • BrandedPreviewJPEGFilter – creates a branded preview image for GIF, JPEG and PNG files (disabled by default)
  • PDFFilter – extracts the full text of Adobe PDF documents (only if text-based or OCRed) for full text indexing
  • WordFilter – extracts the full text of Microsoft Word or Plain Text documents for full text indexing
• Please note that the filter-media or MediaFilterManager will automatically update the DSpace search index by default (see Re-index DSpace). This is the recommended way to run these scripts. But, should you wish to disable it, you can pass the -n flag to either script to do so.
• The following additional options are also available for either the filter-media or MediaFilterManager scripts:</p>
  • -v = verbose mode (prints out all extracted text and additional messages)
  • -f = forces reprocessing of all bitstreams (by default only unprocessed bitstreams are processed by this script)
  • -n = do not update the search index after completion
  • -i <handle> = only process bitstreams within a particular community/collection/item represented by the <handle>
  • -m <#> = only process a maximum number of bitstreams (specified by <#>) . This is useful if you want to process bitstreams little-by-little in order to avoid taxing your server too much.