Text Processing Nodes Don't Work When Installed Via Command Line

[kwhipke1]

OS: Oracle Linux Server release 7.2
Knime Version: 3.7.0

I’m trying to figure out the steps to set up KNIME in a headless environment. I’m testing this out on an Oracle Linux Server VM. Everything seems to work fine except for the text processing nodes (which I need to use). I downloaded and unzipped the Linux version of KNIME. I also downloaded the update site zip and installed the extensions via the command line, like so:

./knime -application org.eclipse.equinox.p2.director -nosplash \
	-repository jar:file:///home/atg/apps/endeca/apps/CompanyName/config/knime/org.knime.update.analytics-platform_3.7.0.zip\!/ \
	-installIU org.knime.ext.textprocessing,org.knime.datageneration
	-consoleLog

I then tried running my workflow from the command line. It failed with messages like this for all the text processing nodes:

ERROR	KNIME-Worker-8 GenericStreamablePreprocessingNodeModel	Preprocessing instance could not be created!
WARN	KNIME-Worker-8 Node	java.lang.NullPointerException

I get the same error trying to run the workflow in the GUI on this VM.

Just as a quick sanity check, I loaded up the KNIME GUI on this VM and opened up the example 50_Applications/07_Forum_Analysis_of_the_KNIME_Forum/02_Text_Classification_from_Forum_Posts. The workflow failed to load with the following errors:

ERROR LoadWorkflowRunnable            Errors during load: Status: Error: 02_Text_Classification_from_Forum_Posts 0 loaded with errors
ERROR LoadWorkflowRunnable              Status: Error: 02_Text_Classification_from_Forum_Posts 0
ERROR LoadWorkflowRunnable                Status: Error: data preparation 0:171
ERROR LoadWorkflowRunnable                  Status: Error: Unable to load node with ID suffix 206 into workflow, skipping it: Could not initialize class org.knime.ext.textprocessing.nodes.preprocessing.stopwordfilter.BuildInStopwordListFactory
ERROR LoadWorkflowRunnable                  Status: Warning: Porter Stemmer 0:171:207
ERROR LoadWorkflowRunnable                    Status: Warning: State has changed from CONFIGURED to IDLE
ERROR LoadWorkflowRunnable                  Status: Warning: State has changed from CONFIGURED to IDLE
ERROR LoadWorkflowRunnable                Status: Error: data preparation 0:99
ERROR LoadWorkflowRunnable                  Status: Error: Unable to load node with ID suffix 206 into workflow, skipping it: Could not initialize class org.knime.ext.textprocessing.nodes.preprocessing.stopwordfilter.BuildInStopwordListFactory

These errors don’t occur when I install the extension via the GUI. They only happen when I install via the command line.

What do I need to do on the command line to ensure the extension is properly installed and initialized?

[kwhipke1]

Haven’t tested it yet, but to avoid wasting anyone’s time:

I think the problem is that I didn’t use the correct feature ID. I need to go to the GUI and look at the ID listed in the “Install new software” for the features I want to install. It should be:

./knime -application org.eclipse.equinox.p2.director -nosplash \
	-repository jar:file:///home/atg/apps/endeca/apps/CompanyName/config/knime/org.knime.update.analytics-platform_3.7.0.zip\!/ \
	-installIU org.knime.features.ext.textprocessing.feature.group,org.knime.features.datageneration.feature.group
	-consoleLog