ctakes-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chen, Pei" <Pei.C...@childrens.harvard.edu>
Subject RE: cTAKES resources and the classpath
Date Thu, 04 Oct 2012 22:54:01 GMT
Ohh, you meant Type System Descriptors files (my apologies- I misread that).  
Yes absolutely agree on that point!


> -----Original Message-----
> From: Steven Bethard [mailto:steven.bethard@Colorado.EDU]
> Sent: Thursday, October 04, 2012 6:50 PM
> To: ctakes-dev@incubator.apache.org
> Subject: Re: cTAKES resources and the classpath
> 
> On Oct 4, 2012, at 4:44 PM, "Chen, Pei" <Pei.Chen@childrens.harvard.edu>
> wrote:
> >> The other thing that should probably be moved into src/main/resources
> >> are the type system descriptor files. You can reference those on the
> >> class path, so they definitely belong in the jar as well.
> >
> > Yes, just need to mull over that just for a bit more (putting configuration
> files inside jars).  With my end-user hat on: some of those xml contain simple
> configuration parameters such as max permutation where end-users may
> just want to tweak without rebuilding code.
> 
> But I wasn't talking about the analysis engine descriptor files. Just the type
> system descriptor files. People definitely shouldn't be tweaking the type
> system - changes to the type system will almost certainly break cTAKES code.
> 
> Steve
> 
> 
> >  When those xml's are outside of the jars, it's very easy to do and
> > re-run a pipeline.  If we distribute and package these up, we may need
> > to provide some mechanism for end-users to easily modify configuration
> > parameters (either thru uiafit style or gui of some sort.)
> >
> > Yes, even if it's in src/main/resources, we can always configure maven to
> move it anywhere during assembly time, but just wanted clearly on this
> before opening up Eclipse again.
> >
> >> -----Original Message-----
> >> From: Steven Bethard [mailto:steven.bethard@Colorado.EDU]
> >> Sent: Thursday, October 04, 2012 6:30 PM
> >> To: ctakes-dev@incubator.apache.org
> >> Subject: Re: cTAKES resources and the classpath
> >>
> >> On Oct 4, 2012, at 4:26 PM, "Chen, Pei"
> >> <Pei.Chen@childrens.harvard.edu>
> >> wrote:
> >>> The individual resources have been moved to src/main/resources so
> >>> that
> >> maven will automatically add it to the classpath and individual jar(s).
> >>
> >> Thanks for doing this!
> >>
> >>
> >> Steve
> >>
> >>> It was also an opportunity to add a fully qualified namespace such
> >>> as /org/apach/ctakes/{project}/somemodelsresourcedirectory.  This
> >>> will be important to avoid clashes if we decide to distribute some
> >>> uber-jar/war or move resources around again. [Even though the
> >>> resources were updated in the descriptor xml files, still haven't
> >>> decided to move those yet since it's working now and we may move to
> >>> uimaFIT style in the future.]
> >>>
> >>> I just tested the default clinical documents pipeline in trunk with
> >>> the CVD
> >> and it should be working now.
> >>> Feel free to check out and work off trunk in ASF SVN now: svn co
> >>> https://svn.apache.org/repos/asf/incubator/ctakes/trunk
> >>> (The OrangeBook will still require Lucene 3.0.2 so you might have to
> >>> update that from 3.5.0 if you plan to use the dictionary lookup)
> >>>
> >>> [Disclaimer: over ~5000-6000 files were touched since we updated
> >>> java namespaces, license headers, resource paths, etc. and wouldn't
> >>> be surprised if we missed something.  Feel free to let us know if
> >>> you encounter any issues or even better- open up a Jira and fix it
> >>> :)]
> >>>
> >>> <-- Will get some coffee and start looking into some of the open
> >>> Jira issues
> >> now.  I think we should have a Release Candidate for 3.0.0-incubating
> >> or a baseline within ASF very soon.
> >>>
> >>> --Pei
> >>>
> >>>> -----Original Message-----
> >>>> From: Steven Bethard [mailto:steven.bethard@Colorado.EDU]
> >>>> Sent: Tuesday, October 02, 2012 5:25 PM
> >>>> To: ctakes-dev@incubator.apache.org
> >>>> Subject: Re: cTAKES resources and the classpath
> >>>>
> >>>> Yeah, I'm going to commit a change to the dependency parser that
> >>>> allows me to do this. (Currently the location of the lemmatized
> >>>> data file is hard coded.)
> >>>>
> >>>> Steve
> >>>>
> >>>> On Oct 2, 2012, at 3:12 PM, "Chen, Pei"
> >>>> <Pei.Chen@childrens.harvard.edu>
> >>>> wrote:
> >>>>
> >>>>> Hi Steve,
> >>>>> I didn't get a chance to do this last week, but will work on it
> >>>>> tonight and
> >>>> tomorrow.
> >>>>> Don't want this to stop you though, so feel free to read in the
> >>>>> resource
> >>>> from an another location temporarily...
> >>>>> --Pei
> >>>>>
> >>>>>> -----Original Message-----
> >>>>>> From: Steven Bethard [mailto:steven.bethard@Colorado.EDU]
> >>>>>> Sent: Tuesday, October 02, 2012 4:47 PM
> >>>>>> To: ctakes-dev@incubator.apache.org
> >>>>>> Subject: cTAKES resources and the classpath
> >>>>>>
> >>>>>> On Sep 25, 2012, at 3:22 PM, "Chen, Pei"
> >>>>>> <Pei.Chen@childrens.harvard.edu>
> >>>>>> wrote:
> >>>>>>> I think Brandyn and others had a reasonable suggestion to
use
> >>>>>>> the name
> >>>>>> instead of location paths in those pesky descriptor xml files-
> >>>>>> Assuming they're in the classpath (placed in src/main/resources).
> >>>>>>> I'll take a stab at this later this week if I get a chance.
> >>>>>>
> >>>>>> Pei, are you still working on this? I'm currently unable to
use
> >>>>>> the dependency parser from another project because it can't
find
> >>>>>> it's resources. Specifically, it fails on this line:
> >>>>>>
> >>>>>> URL lemmatizerDataFileURL =
> >>>>>> this.getClass().getResource(ENG_LEMMATIZER_DATA_FILE);
> >>>>>>
> >>>>>> I would take a stab at fixing these myself, but the JCasGen
issue
> >>>>>> makes these kind of broad changes really difficult for me.
> >>>>>>
> >>>>>> Steve
> >>>>>
> >>>
> >


Mime
View raw message