Wikis - Page

Configure ControlPoint with MediaServer OCR

2 Likes

ISSUE:

Some may find the instructions in the installation guide difficult to follow and implement concerning integration with IDOL MediaServer. This is a simple way to configure MediaServer to OCR images that will be then treated by CP.

HOW TO STEPS:

Download Media Server.

Install it (this could be just a zip unpack) in Program Files\Micro Focus\MediaServer for example.

You can make it a service with a command like:

New-Service -Name "ControlPoint MediaServer" -BinaryPathName "C:\Program Files\Micro Focus\MediaServer\mediaserver.exe" -StartupType Automatic -Description "MediaServer for ControlPoint"

Be sure that you have a license for Media Server, it is not in the CP base license, it is an option.

Open ...MediaServer\mediaserver.cfg file.

  • Note the port mentionned in the [Server] section: default14000
  • Check VisualChannels=1 in [Channel] section

Edit ...ControlPoint\Indexer\FileSystem Connector Framework\ControlPoint FileSystem Connector Framework.cfg

  • add in the [ImportTasks] section, at the end of the PreXX,

Pre2=lua:scripts/mediaserver/OCR.lua

Rename existing Pre2 if needed to Pre3.

  • Add a section (after ImportTasks)

[MediaServerSettings]
MediaServerHost=localhost
MediaServerPort=14000
ReadFromOriginalLocation=true

Notes:

  • Change localhost with the name of server when MediaServer is installed on a different server than CFS
  • 14000 is the default port: you can change it in the mediaserver.cfg file

Edit ...ControlPoint\Indexer\FileSystem Connector Framework\scripts\mediaserver\OCR.lua

  • Move line 63 (with the "End") to line 67, after the document:addField... line so that it looks like

});

end
-- If analysis was performed successfully, don't extract

-- text using KeyView, just get the metadata.

document:addField("AUTN_FILTER_META_ONLY", "");
end

return true;

end

Save the files 

Restart FileSystem Connector Framework service.

Start ...MediaServer\mediaserver.exe or the service and wait few minutes.

You're ready!

Labels:

How To-Best Practice
Comment List
  • Note that within the OCR.lua, by default only the following Document Classes are sent to Media Server for analysis:

    local supportedDocumentClasses = {
    ["4"] = "Image", -- adRASTERIMAGE
    ["20"] = "AudioVideo" -- adMOVIE
    }

    Refer to the CFS Admin Guide - Appendix C:KeyView Classes for the full list of classes and corresponding Attribute Numbers.
    If you wish to OCR images and PDF files then you will need the following:


    local supportedDocumentClasses = {
    ["1"] = "Image", -- adWordProcessor
    ["4"] = "Image", -- adRASTERIMAGE
    ["5"] = "Image" -- adVECTORGRAPHIC
    }

Related
Recommended