missing medication mentions (tika cTAKESParser) Inbox x

classic Classic list List threaded Threaded
10 messages Options
Reply | Threaded
Open this post in threaded view
|

missing medication mentions (tika cTAKESParser) Inbox x

Patrick Young
I am using tika-app-1.19.jar & ctakes4.0.0 to populate neo4j with ctakes event mentions extracted from biomedical articles. However, I've noticed some  medication mentions e.g., indinavir, zidovudine are missed while other antiretrovirals such as lamivudine are detected.

The default CVD spots these meds properly though... any ideas why this might be happening?

Many thanks,
Paddy Young


--
Dr Patrick M Young
Reply | Threaded
Open this post in threaded view
|

Re: missing medication mentions (tika cTAKESParser) Inbox x

Tim Allison
Chris,
  I know nothing about ctakes...any ideas?
On Wed, Oct 10, 2018 at 4:33 AM Patrick Young <[hidden email]> wrote:

>
> I am using tika-app-1.19.jar & ctakes4.0.0 to populate neo4j with ctakes event mentions extracted from biomedical articles. However, I've noticed some  medication mentions e.g., indinavir, zidovudine are missed while other antiretrovirals such as lamivudine are detected.
>
> The default CVD spots these meds properly though... any ideas why this might be happening?
>
> Many thanks,
> Paddy Young
>
>
> --
> Dr Patrick M Young
Reply | Threaded
Open this post in threaded view
|

Re: missing medication mentions (tika cTAKESParser) Inbox x

Steph van Schalkwyk
This is not Tika functionality - unless the items above are not output by Tika, which is unlikely.
Use the cTakes GUI to check if those items exist in the RRF files.
Steph

+1.314.452.2896 (Tel/SMS)


On Wed, Oct 10, 2018 at 10:04 AM Tim Allison <[hidden email]> wrote:
Chris,
  I know nothing about ctakes...any ideas?
On Wed, Oct 10, 2018 at 4:33 AM Patrick Young <[hidden email]> wrote:
>
> I am using tika-app-1.19.jar & ctakes4.0.0 to populate neo4j with ctakes event mentions extracted from biomedical articles. However, I've noticed some  medication mentions e.g., indinavir, zidovudine are missed while other antiretrovirals such as lamivudine are detected.
>
> The default CVD spots these meds properly though... any ideas why this might be happening?
>
> Many thanks,
> Paddy Young
>
>
> --
> Dr Patrick M Young
Reply | Threaded
Open this post in threaded view
|

Re: missing medication mentions (tika cTAKESParser) Inbox x

Chris Mattmann
In reply to this post by Tim Allison

Hmm, perhaps Patrick take a look at the CTAKESContentHandler code here [1] and
the wiki here:

 

https://wiki.apache.org/tika/cTAKESParser

 

We may be pinned to an older version of cTAKES and/or we may not be flowing it
out properly (we take CTAKES output and then format it for TIka).

 

Cheers,

Chris

 

 

[1] https://github.com/apache/tika/blob/master/tika-parsers/src/main/java/org/apache/tika/parser/ctakes/CTAKESContentHandler.java

 

From: Tim Allison <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 8:05 AM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Chris,

  I know nothing about ctakes...any ideas?

On Wed, Oct 10, 2018 at 4:33 AM Patrick Young <[hidden email]> wrote:

 

I am using tika-app-1.19.jar & ctakes4.0.0 to populate neo4j with ctakes event mentions extracted from biomedical articles. However, I've noticed some  medication mentions e.g., indinavir, zidovudine are missed while other antiretrovirals such as lamivudine are detected.

 

The default CVD spots these meds properly though... any ideas why this might be happening?

 

Many thanks,

Paddy Young

 

 

--

Dr Patrick M Young

 

Reply | Threaded
Open this post in threaded view
|

Re: missing medication mentions (tika cTAKESParser) Inbox x

Patrick Young
Steph, the cTakes GUI is the default CVD I mentioned above & it does find the relevant medications. What are RRF files?

Chris, ***good shout***. Made me realise what the issue was: the analysis engine...

I altered CTakesConfig.properties file to point at AggregatePlaintextFastUMLSProcessor.xml rather than AggregatePlaintextUMLSProcessor.xml  by editing this line

aeDescriptorPath=/ctakes-clinical-pipeline/desc/analysis_engine/AggregatePlaintextFastUMLSProcessor.xml

For the benefit of future travellers, I also had to reboot my computer afterwards... the database gets locked otherwise & it won't work.

Thank you all for responding to me so promptly. I'm most happy.


On Wed, 10 Oct 2018 at 16:35, Chris Mattmann <[hidden email]> wrote:

Hmm, perhaps Patrick take a look at the CTAKESContentHandler code here [1] and
the wiki here:

 

https://wiki.apache.org/tika/cTAKESParser

 

We may be pinned to an older version of cTAKES and/or we may not be flowing it
out properly (we take CTAKES output and then format it for TIka).

 

Cheers,

Chris

 

 

[1] https://github.com/apache/tika/blob/master/tika-parsers/src/main/java/org/apache/tika/parser/ctakes/CTAKESContentHandler.java

 

From: Tim Allison <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 8:05 AM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Chris,

  I know nothing about ctakes...any ideas?

On Wed, Oct 10, 2018 at 4:33 AM Patrick Young <[hidden email]> wrote:

 

I am using tika-app-1.19.jar & ctakes4.0.0 to populate neo4j with ctakes event mentions extracted from biomedical articles. However, I've noticed some  medication mentions e.g., indinavir, zidovudine are missed while other antiretrovirals such as lamivudine are detected.

 

The default CVD spots these meds properly though... any ideas why this might be happening?

 

Many thanks,

Paddy Young

 

 

--

Dr Patrick M Young

 



--
Dr Patrick M Young
Reply | Threaded
Open this post in threaded view
|

Re: missing medication mentions (tika cTAKESParser) Inbox x

Chris Mattmann

Thanks Patrick, would you consider updating this in an FAQ on the wiki page I referenced below?

If you register for an account there and tell me the name I can grant you permissions to edit the
page.


thanks!

 

Cheers,

Chris

 

 

 

 

From: Patrick Young <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 12:59 PM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Steph, the cTakes GUI is the default CVD I mentioned above & it does find the relevant medications. What are RRF files?

 

Chris, ***good shout***. Made me realise what the issue was: the analysis engine...

 

I altered CTakesConfig.properties file to point at AggregatePlaintextFastUMLSProcessor.xml rather than AggregatePlaintextUMLSProcessor.xml  by editing this line

 

aeDescriptorPath=/ctakes-clinical-pipeline/desc/analysis_engine/AggregatePlaintextFastUMLSProcessor.xml

 

For the benefit of future travellers, I also had to reboot my computer afterwards... the database gets locked otherwise & it won't work.

 

Thank you all for responding to me so promptly. I'm most happy.

 

 

On Wed, 10 Oct 2018 at 16:35, Chris Mattmann <[hidden email]> wrote:

Hmm, perhaps Patrick take a look at the CTAKESContentHandler code here [1] and
the wiki here:

 

https://wiki.apache.org/tika/cTAKESParser

 

We may be pinned to an older version of cTAKES and/or we may not be flowing it
out properly (we take CTAKES output and then format it for TIka).

 

Cheers,

Chris

 

 

[1] https://github.com/apache/tika/blob/master/tika-parsers/src/main/java/org/apache/tika/parser/ctakes/CTAKESContentHandler.java

 

From: Tim Allison <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 8:05 AM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Chris,

  I know nothing about ctakes...any ideas?

On Wed, Oct 10, 2018 at 4:33 AM Patrick Young <[hidden email]> wrote:

 

I am using tika-app-1.19.jar & ctakes4.0.0 to populate neo4j with ctakes event mentions extracted from biomedical articles. However, I've noticed some  medication mentions e.g., indinavir, zidovudine are missed while other antiretrovirals such as lamivudine are detected.

 

The default CVD spots these meds properly though... any ideas why this might be happening?

 

Many thanks,

Paddy Young

 

 

--

Dr Patrick M Young

 



--

Dr Patrick M Young

Reply | Threaded
Open this post in threaded view
|

Re: missing medication mentions (tika cTAKESParser) Inbox x

Patrick Young
Chris, I've registered as Patrick Young and gone to


Is this what you meant?




On Wed, 10 Oct 2018 at 21:03, Chris Mattmann <[hidden email]> wrote:

Thanks Patrick, would you consider updating this in an FAQ on the wiki page I referenced below?

If you register for an account there and tell me the name I can grant you permissions to edit the
page.


thanks!

 

Cheers,

Chris

 

 

 

 

From: Patrick Young <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 12:59 PM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Steph, the cTakes GUI is the default CVD I mentioned above & it does find the relevant medications. What are RRF files?

 

Chris, ***good shout***. Made me realise what the issue was: the analysis engine...

 

I altered CTakesConfig.properties file to point at AggregatePlaintextFastUMLSProcessor.xml rather than AggregatePlaintextUMLSProcessor.xml  by editing this line

 

aeDescriptorPath=/ctakes-clinical-pipeline/desc/analysis_engine/AggregatePlaintextFastUMLSProcessor.xml

 

For the benefit of future travellers, I also had to reboot my computer afterwards... the database gets locked otherwise & it won't work.

 

Thank you all for responding to me so promptly. I'm most happy.

 

 

On Wed, 10 Oct 2018 at 16:35, Chris Mattmann <[hidden email]> wrote:

Hmm, perhaps Patrick take a look at the CTAKESContentHandler code here [1] and
the wiki here:

 

https://wiki.apache.org/tika/cTAKESParser

 

We may be pinned to an older version of cTAKES and/or we may not be flowing it
out properly (we take CTAKES output and then format it for TIka).

 

Cheers,

Chris

 

 

[1] https://github.com/apache/tika/blob/master/tika-parsers/src/main/java/org/apache/tika/parser/ctakes/CTAKESContentHandler.java

 

From: Tim Allison <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 8:05 AM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Chris,

  I know nothing about ctakes...any ideas?

On Wed, Oct 10, 2018 at 4:33 AM Patrick Young <[hidden email]> wrote:

 

I am using tika-app-1.19.jar & ctakes4.0.0 to populate neo4j with ctakes event mentions extracted from biomedical articles. However, I've noticed some  medication mentions e.g., indinavir, zidovudine are missed while other antiretrovirals such as lamivudine are detected.

 

The default CVD spots these meds properly though... any ideas why this might be happening?

 

Many thanks,

Paddy Young

 

 

--

Dr Patrick M Young

 



--

Dr Patrick M Young



--
Dr Patrick M Young
Reply | Threaded
Open this post in threaded view
|

Re: missing medication mentions (tika cTAKESParser) Inbox x

Chris Mattmann

I’ve granted your permissions to edit the wiki.

Can you edit this page:

 

https://wiki.apache.org/tika/cTAKESParser

 

With a note on if you want to use a different engine and how?

 

 

 

From: Patrick Young <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 1:16 PM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Chris, I've registered as Patrick Young and gone to

 

Is this what you meant?

 

 

 

 

On Wed, 10 Oct 2018 at 21:03, Chris Mattmann <[hidden email]> wrote:

Thanks Patrick, would you consider updating this in an FAQ on the wiki page I referenced below?

If you register for an account there and tell me the name I can grant you permissions to edit the
page.


thanks!

 

Cheers,

Chris

 

 

 

 

From: Patrick Young <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 12:59 PM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Steph, the cTakes GUI is the default CVD I mentioned above & it does find the relevant medications. What are RRF files?

 

Chris, ***good shout***. Made me realise what the issue was: the analysis engine...

 

I altered CTakesConfig.properties file to point at AggregatePlaintextFastUMLSProcessor.xml rather than AggregatePlaintextUMLSProcessor.xml  by editing this line

 

aeDescriptorPath=/ctakes-clinical-pipeline/desc/analysis_engine/AggregatePlaintextFastUMLSProcessor.xml

 

For the benefit of future travellers, I also had to reboot my computer afterwards... the database gets locked otherwise & it won't work.

 

Thank you all for responding to me so promptly. I'm most happy.

 

 

On Wed, 10 Oct 2018 at 16:35, Chris Mattmann <[hidden email]> wrote:

Hmm, perhaps Patrick take a look at the CTAKESContentHandler code here [1] and
the wiki here:

 

https://wiki.apache.org/tika/cTAKESParser

 

We may be pinned to an older version of cTAKES and/or we may not be flowing it
out properly (we take CTAKES output and then format it for TIka).

 

Cheers,

Chris

 

 

[1] https://github.com/apache/tika/blob/master/tika-parsers/src/main/java/org/apache/tika/parser/ctakes/CTAKESContentHandler.java

 

From: Tim Allison <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 8:05 AM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Chris,

  I know nothing about ctakes...any ideas?

On Wed, Oct 10, 2018 at 4:33 AM Patrick Young <[hidden email]> wrote:

 

I am using tika-app-1.19.jar & ctakes4.0.0 to populate neo4j with ctakes event mentions extracted from biomedical articles. However, I've noticed some  medication mentions e.g., indinavir, zidovudine are missed while other antiretrovirals such as lamivudine are detected.

 

The default CVD spots these meds properly though... any ideas why this might be happening?

 

Many thanks,

Paddy Young

 

 

--

Dr Patrick M Young

 



--

Dr Patrick M Young



--

Dr Patrick M Young

Reply | Threaded
Open this post in threaded view
|

Re: missing medication mentions (tika cTAKESParser) Inbox x

Patrick Young
I edited the wiki paragraph that describes the ctakes properties configuration file as requested...

On Thu, 11 Oct 2018 at 18:57, Chris Mattmann <[hidden email]> wrote:

I’ve granted your permissions to edit the wiki.

Can you edit this page:

 

https://wiki.apache.org/tika/cTAKESParser

 

With a note on if you want to use a different engine and how?

 

 

 

From: Patrick Young <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 1:16 PM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Chris, I've registered as Patrick Young and gone to

 

Is this what you meant?

 

 

 

 

On Wed, 10 Oct 2018 at 21:03, Chris Mattmann <[hidden email]> wrote:

Thanks Patrick, would you consider updating this in an FAQ on the wiki page I referenced below?

If you register for an account there and tell me the name I can grant you permissions to edit the
page.


thanks!

 

Cheers,

Chris

 

 

 

 

From: Patrick Young <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 12:59 PM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Steph, the cTakes GUI is the default CVD I mentioned above & it does find the relevant medications. What are RRF files?

 

Chris, ***good shout***. Made me realise what the issue was: the analysis engine...

 

I altered CTakesConfig.properties file to point at AggregatePlaintextFastUMLSProcessor.xml rather than AggregatePlaintextUMLSProcessor.xml  by editing this line

 

aeDescriptorPath=/ctakes-clinical-pipeline/desc/analysis_engine/AggregatePlaintextFastUMLSProcessor.xml

 

For the benefit of future travellers, I also had to reboot my computer afterwards... the database gets locked otherwise & it won't work.

 

Thank you all for responding to me so promptly. I'm most happy.

 

 

On Wed, 10 Oct 2018 at 16:35, Chris Mattmann <[hidden email]> wrote:

Hmm, perhaps Patrick take a look at the CTAKESContentHandler code here [1] and
the wiki here:

 

https://wiki.apache.org/tika/cTAKESParser

 

We may be pinned to an older version of cTAKES and/or we may not be flowing it
out properly (we take CTAKES output and then format it for TIka).

 

Cheers,

Chris

 

 

[1] https://github.com/apache/tika/blob/master/tika-parsers/src/main/java/org/apache/tika/parser/ctakes/CTAKESContentHandler.java

 

From: Tim Allison <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 8:05 AM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Chris,

  I know nothing about ctakes...any ideas?

On Wed, Oct 10, 2018 at 4:33 AM Patrick Young <[hidden email]> wrote:

 

I am using tika-app-1.19.jar & ctakes4.0.0 to populate neo4j with ctakes event mentions extracted from biomedical articles. However, I've noticed some  medication mentions e.g., indinavir, zidovudine are missed while other antiretrovirals such as lamivudine are detected.

 

The default CVD spots these meds properly though... any ideas why this might be happening?

 

Many thanks,

Paddy Young

 

 

--

Dr Patrick M Young

 



--

Dr Patrick M Young



--

Dr Patrick M Young



--
Dr Patrick M Young
Reply | Threaded
Open this post in threaded view
|

Re: missing medication mentions (tika cTAKESParser) Inbox x

Chris Mattmann

Perfect great job!

 

 

 

From: Patrick Young <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Thursday, October 11, 2018 at 1:59 PM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

I edited the wiki paragraph that describes the ctakes properties configuration file as requested...

 

On Thu, 11 Oct 2018 at 18:57, Chris Mattmann <[hidden email]> wrote:

I’ve granted your permissions to edit the wiki.

Can you edit this page:

 

https://wiki.apache.org/tika/cTAKESParser

 

With a note on if you want to use a different engine and how?

 

 

 

From: Patrick Young <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 1:16 PM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Chris, I've registered as Patrick Young and gone to

 

Is this what you meant?

 

 

 

 

On Wed, 10 Oct 2018 at 21:03, Chris Mattmann <[hidden email]> wrote:

Thanks Patrick, would you consider updating this in an FAQ on the wiki page I referenced below?

If you register for an account there and tell me the name I can grant you permissions to edit the
page.


thanks!

 

Cheers,

Chris

 

 

 

 

From: Patrick Young <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 12:59 PM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Steph, the cTakes GUI is the default CVD I mentioned above & it does find the relevant medications. What are RRF files?

 

Chris, ***good shout***. Made me realise what the issue was: the analysis engine...

 

I altered CTakesConfig.properties file to point at AggregatePlaintextFastUMLSProcessor.xml rather than AggregatePlaintextUMLSProcessor.xml  by editing this line

 

aeDescriptorPath=/ctakes-clinical-pipeline/desc/analysis_engine/AggregatePlaintextFastUMLSProcessor.xml

 

For the benefit of future travellers, I also had to reboot my computer afterwards... the database gets locked otherwise & it won't work.

 

Thank you all for responding to me so promptly. I'm most happy.

 

 

On Wed, 10 Oct 2018 at 16:35, Chris Mattmann <[hidden email]> wrote:

Hmm, perhaps Patrick take a look at the CTAKESContentHandler code here [1] and
the wiki here:

 

https://wiki.apache.org/tika/cTAKESParser

 

We may be pinned to an older version of cTAKES and/or we may not be flowing it
out properly (we take CTAKES output and then format it for TIka).

 

Cheers,

Chris

 

 

[1] https://github.com/apache/tika/blob/master/tika-parsers/src/main/java/org/apache/tika/parser/ctakes/CTAKESContentHandler.java

 

From: Tim Allison <[hidden email]>
Reply-To: "[hidden email]" <[hidden email]>
Date: Wednesday, October 10, 2018 at 8:05 AM
To: "[hidden email]" <[hidden email]>
Subject: Re: missing medication mentions (tika cTAKESParser) Inbox x

 

Chris,

  I know nothing about ctakes...any ideas?

On Wed, Oct 10, 2018 at 4:33 AM Patrick Young <[hidden email]> wrote:

 

I am using tika-app-1.19.jar & ctakes4.0.0 to populate neo4j with ctakes event mentions extracted from biomedical articles. However, I've noticed some  medication mentions e.g., indinavir, zidovudine are missed while other antiretrovirals such as lamivudine are detected.

 

The default CVD spots these meds properly though... any ideas why this might be happening?

 

Many thanks,

Paddy Young

 

 

--

Dr Patrick M Young

 



--

Dr Patrick M Young



--

Dr Patrick M Young



--

Dr Patrick M Young