ctakes-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Miller, Timothy" <Timothy.Mil...@childrens.harvard.edu>
Subject Re: Looking for cTakes deployment strategies [EXTERNAL]
Date Tue, 29 Jan 2019 11:45:11 GMT
I have seen this with SentenceDetectorAnnotatorBIO.xml annotator, but with the one you describe,
I thought it had a hard-coded rule to break on newlines and split them into sentences. Do
you have any log files that you can copy/paste the initialization lines so we can verify which
sentence segmenter you're running?

From: Joseph Erfani <yerfani@foreseemed.com>
Sent: Monday, January 28, 2019 6:19 PM
To: dev@ctakes.apache.org
Subject: Re: Looking for cTakes deployment strategies [EXTERNAL]

Hello everyone,
I have a question regarding the cTakes sentence detector. I am using
the "SentenceDetectorAnnotator.xml"
analysis engine located in the ctakes-core for sentence boundary detection.
It seems that the sentence boundary engine is not able to find the sentence
boundary, when a sentence is finished with a carriage return instead of a
period or several spaces.
e.g. the note
"He is a smoker
He has hypertension"

all the text is considered as one sentence, while there is a carriage
return after the word 'smoker' (at the end of the first sentence).
Have you encountered similar problem or do  you  have any suggestion for

Thank you

On Wed, Jan 16, 2019 at 10:47 AM Anusha Balasubramaniam <
anushab@foreseemed.com> wrote:

> Hello everyone,
> I am looking for a strategy to use cTakes to asynchronously process
> thousands of clinical notes by listening to a queue on AWS and maintaining
> a hot process with all the dictionaries loaded in memory. So far I've had
> some success using the REST server wrapper I found here:
> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_dirkweissenborn_ctakes-2Dserver&d=DwIBaQ&c=qS4goWBT7poplM69zy_3xhKwEW14JZMSdioCoppxeFU&r=Heup-IbsIg9Q1TPOylpP9FE4GTK-OqdTDRRNQXipowRLRjx0ibQrHEo8uYx6674h&m=NkatAgDcxp3wmhmcluDwWbJYocosvqWSD3kcmDDGtHU&s=e3GmZyP0_WM8lPbTwxDcmbT1Qspwfgj-tSbYM3Wk-Q0&e=,
but it's still a
> synchronous call, which I found hard to scale.
> Are there any other wrappers out there that could be used to enable cTakes
> to listen to a port for input? Can anyone share some strategies they used
> to implement cTakes on AWS to achieve similar requirements?
> Thanks and Regards,
> Anusha

View raw message