Manually loading rulebases into Classification Server

To load rulebases in a classification server instance the simplest way is of course to use publisher! There are however occasions where manually loading rulebases may be required for troubleshooting or experimenting.

Publishing manually involve three steps:

• Creating a new publish set. A publish set is the container for all the rulebases associated with the model or the use case
• Uploading the PAK files one by one to the publish set
• Committing the publish set and reloading the rulenet so the rulebases become active.

Those three steps can be achieved via the Classification Server Test Interface, using the XML request section at the bottom of the interface.

First create a new publish set using the following XML request:

 <request op="publish_set_init">
     <publish_set>the_name_of_my_set</publish_set>
 </request>

and submit without any file payload

Then for each PAK or rulebase you want to publish issue the following:

 <request op="publish_set_add">
    <publish_set>the_name_of_my_set</publish_set>
 </request>

with the pak file or rulebase file added to the payload using the “Choose file” button at the bottom of the screen.

When all the assets are uploaded issue the following command to recompile the rulenet and make it live:

 <request op="publish_set_activate">
     <publish_set>the_name_of_my_set</publish_set>
 </request>

When successful the message “RULEBASE RELOADING” will be displayed in the web browser.