ctakes-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Murphy, Sean P." <Murphy.S...@mayo.edu>
Subject RE: How to use Drug NER
Date Tue, 11 Mar 2014 15:48:31 GMT
Ravi,
                Can you look at the list of IdentifiedAnnotation to see if the entities include
the drugs in question.   Specifically, the 'ctakes.typesystem.type.textsem.MedicationMention'
should be populated. Additionally, there should be the missing drug related annotations, such
as strength in that list.
                The problem is that the drug ner was migrated to use the new type system a
release or two back.  However, the old type system format was allowed to remain in the event
that someone required that legacy handling.  These can actually be populated by uncommenting
some code that exists in the DrugMention source code and rerunning.
                I had forgotten about these changes.  Hopefully, that explains what you're
seeing.
                Thanks,
                                ~Sean

From: ravi garg [mailto:ravigarg27@gmail.com]
Sent: Tuesday, March 11, 2014 10:15 AM
To: user@ctakes.apache.org
Subject: Re: How to use Drug NER

Hey,
I am using ctakes-drug-ner/desc/analysis_engine/AggregatePlainTextUMLSProcessor.xml as the
Analysis Engine in the CVD where in I am getting the other annotations such as Medication
Annotations Fraction Annotation which are of ctakes.typesystem.type.textsem.IdentifiedAnnotation
but do not get any of the ctakes.drugner.type annotations for eg Frequency Annotation Drug
Strength Annotations etc. I dont get any errors though but I feel DrugMentionAnnotator descriptor
which extracts the strength, frequency, route etc is not working for me.
Thanks,
Ravi

On Tue, Mar 11, 2014 at 8:18 PM, Murphy, Sean P. <Murphy.Sean@mayo.edu<mailto:Murphy.Sean@mayo.edu>>
wrote:
Ravi,
                I assume you're just using plain text as the input for the pipeline.  Did
you use the provided sample DrugNER_PlainText_CPE.xml available in the ctakes-drug-ner/desc/collection_processing_engine/
folder to see if it can run the provided sample?
                If my assumption is correct than make sure you're using the Analysis Engines
provided for the plain text path (i.e. no aggregates with 'CDA' in the name should be utilized).
 This also pertains to the consumer which output ( make sure to use this one 'AggregatePlaintextProcessor.xml').
                If you are using the CVD (cas visual debugger) then everything but the consumer
would still apply.   Just be sure that you are loading the DrugAggregatePlaintextUMLSProcessor
AE specifier.
                BTW.  Are you seeing other annotations other than the drugner type?

                Thanks,
                                ~Sean

From: ravi garg [mailto:ravigarg27@gmail.com<mailto:ravigarg27@gmail.com>]
Sent: Tuesday, March 11, 2014 2:38 AM

To: user@ctakes.apache.org<mailto:user@ctakes.apache.org>
Subject: Re: How to use Drug NER

Ohh Okay, I am sorry Sean, I didn't actually knew Rxnorm is an umls dictionary and you need
to have a license to use it. I have got that working but still I dont get any annotations
in drugner.type for example RouteAnnotation, DurationAnnotation etc. Any ideas on how to get
that working
Thanks,
Ravi

On Mon, Mar 10, 2014 at 11:56 PM, ravi garg <ravigarg27@gmail.com<mailto:ravigarg27@gmail.com>>
wrote:
Hey Xiayuan,
I ensured that the directory of the resource that is the rxnorm_index is correct and cTAKES
is also actually using it. I can see the logs wherein cTAKES actually parses 16k docs. I get
no errors though on using the pipeline but nothing is extracted or annotated. Even the dosage,
frequency and route which I believe are not dependent on drug dictionary are also not getting
identified. I have worked on the main clinical pipeline before but am working on the Drug
pipeline for the first time so no idea whats wrong.
Regards,
Ravi

On Mon, Mar 10, 2014 at 11:38 PM, Xiayuan Huang <huangxy@cs.wisc.edu<mailto:huangxy@cs.wisc.edu>>
wrote:
Hi Ravi,

Currently I am doing one project using cTAKES. I am trying to use a pipeline for identifying
smoking status, SimulatedProdSmokingTAE.xml. And I also encountered some problem on executing
those pipelines. Some possible errors will be the directory of UIMA resource, you have to
set that correctly. And I got some exceptions when I try to run pipelines. We can discuss
how to implement cTAKES later on, I guess I really need someone to discuss and help out on
this cTAKES system.

Best,
Xiayuan Huang



On 2014-03-10 11:05, ravi garg wrote:
Hey,

I am using DrugAggregatePlainTextProcessor.xml as my analysis engine
annotator. I have downloaded all the required resources by cTAKES 3.1
. I guess it uses the rxnorm_index as the drug database. But I am not
getting any annotations as the output. Please let me know the correct
way using Drug NER

Thanks,
Ravi

--

Ravi Garg
4th Year
MSc (hons) Biological Sciences
B.E (hons) Computer Science and Engineering
BITS Pilani KK Birla Goa Campus



--
Ravi Garg
4th Year
MSc (hons) Biological Sciences
B.E (hons) Computer Science and Engineering
BITS Pilani KK Birla Goa Campus



--
Ravi Garg
4th Year
MSc (hons) Biological Sciences
B.E (hons) Computer Science and Engineering
BITS Pilani KK Birla Goa Campus



--
Ravi Garg
4th Year
MSc (hons) Biological Sciences
B.E (hons) Computer Science and Engineering
BITS Pilani KK Birla Goa Campus

Mime
View raw message