This is more of a newbie question but I have not found a good document
about it. Please help.
As of now, my requirements appear to be :
1> use existing ctakes pipelines;
2> maybe develop one or two on my own
3> provide services (like a web server)
To do that, it seems I could do the followi
Personally, I would prefer Sean's idea and I think it'd be easy to
implement using org.junit.Assume [1]. The drawback is that one *must*
remember to setup the umls credentials to run the tests during a
release, whereas the maven profiles enforce that. But either is fine.
My $0.02,
Hadrian
[1]
Alex,
I forgot to add: thanks for all your work on these changes and improvements.
regarding your email Sean,
I think a small non-umls custom dictionary for a pipeline test would be
great.
And test(s) with combined hsql and bsv would also be great. I don't know
they necessarily need to be outside
Alex, I like the idea of "*2 profiles (in pom.xml), one *with* UMLS **account
and one *without**". However, I would have just the one without the
credentials be part of the Jenkins job and someone would manually run the
other one as part of the release process (or whenever someone felt it was
warr
Hi Alex,
I think that it goes against the umls license to have credentials available to
the public. That might be what you were saying in a previous email:
> Exporting the ctakes_umlsuser, ctakes_umlspw makes the whole build
> succeed. But having some credentials in the Jenkins job (official)
I started to look for ways to make the build stable. After applying patch
for CTAKES-334, the only issue remains with:
testCPE(org.apache.ctakes.regression.test.RegressionPipelineTest):
Initialization of CAS Processor with name "RegressionPipelineAggregateTest"
failed.
which is caused by:
ERROR
