Searching classes in SNOMEDCT-ICD11

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

Searching classes in SNOMEDCT-ICD11

Campbell, James R

I have loaded SNOMED CT and a mapped  working version of ICD-11 into Protégé v5 for analysis.  SNOROCKET will classify them in 70-80 seconds on my laptop and I can browse the class hierarchy acceptably well.  Unfortunately the SEARCH window seems to be in advanced stages of rigor mortis and I cannot look up class members; it takes forever and nothing seems to happen.

 

I revised my run parameters to use all of my system 8Gbytes of memory.  Are there other setup features that are required to enable class searches in a very large ontology?

 

Jim Campbell


The information in this e-mail may be privileged and confidential, intended only for the use of the addressee(s) above. Any unauthorized use or disclosure of this information is prohibited. If you have received this e-mail by mistake, please delete it and immediately contact the sender.
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: Searching classes in SNOMEDCT-ICD11

Matthew Horridge-2
Administrator
Hi Jim,

Are you using the search window in the top right corner, or are you using CTRL+F in the class hierarchy?  The former should be more scalable than the latter.  

From your description below (i.e. no search result at all) it sounds like there’s not enough memory to build the search indices.  If the available memory is low, it’s possible that paging is taking place and this is the cause of the effects that you’re seeing.  I’ve file an issue [1] and I'll try and reproduce this with a version of SNOMED and ICD 11 loaded.

Cheers,

Matthew



On 19 Apr 2015, at 22:52, Campbell, James R <[hidden email]> wrote:

I have loaded SNOMED CT and a mapped  working version of ICD-11 into Protégé v5 for analysis.  SNOROCKET will classify them in 70-80 seconds on my laptop and I can browse the class hierarchy acceptably well.  Unfortunately the SEARCH window seems to be in advanced stages of rigor mortis and I cannot look up class members; it takes forever and nothing seems to happen.

 

I revised my run parameters to use all of my system 8Gbytes of memory.  Are there other setup features that are required to enable class searches in a very large ontology?

 

Jim Campbell


The information in this e-mail may be privileged and confidential, intended only for the use of the addressee(s) above. Any unauthorized use or disclosure of this information is prohibited. If you have received this e-mail by mistake, please delete it and immediately contact the sender.
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: Searching classes in SNOMEDCT-ICD11

Campbell, James R
Thank you for responding.  I was using the search window and have max 8Gb committed to Protege.  I noticed better response on my desktop with 128Gb but boy, is all that required?
Jim

James R. Campbell MD
[hidden email]
Office: 402-559-7505
Secretary: 402-559-7299
Pager: 402-888-1230

> On Apr 19, 2015, at 4:00 PM, "Matthew Horridge" <[hidden email]> wrote:
>
> wrote:

The information in this e-mail may be privileged and confidential, intended only for the use of the addressee(s) above. Any unauthorized use or disclosure of this information is prohibited. If you have received this e-mail by mistake, please delete it and immediately contact the sender.

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: Searching classes in SNOMEDCT-ICD11

Matthew Horridge-2
Administrator
Hi Jim,

A couple of things, (1) we will be upgrading to the OWL API version 4 soon and this has better memory consumption characteristics compared with the current version that is used by Protege (in some cases the savings are significant, but it depends on the ontology), (2) Have you tried ELK instead of SNOROCKET?  I’m not sure whether this will reduce memory consumption or make it worse, but it might be worth a try.  

Cheers,

Matthew



> On 20 Apr 2015, at 12:29, Campbell, James R <[hidden email]> wrote:
>
> Thank you for responding.  I was using the search window and have max 8Gb committed to Protege.  I noticed better response on my desktop with 128Gb but boy, is all that required?
> Jim
>
> James R. Campbell MD
> [hidden email]
> Office: 402-559-7505
> Secretary: 402-559-7299
> Pager: 402-888-1230
>
>> On Apr 19, 2015, at 4:00 PM, "Matthew Horridge" <[hidden email]> wrote:
>>
>> wrote:
>
> The information in this e-mail may be privileged and confidential, intended only for the use of the addressee(s) above. Any unauthorized use or disclosure of this information is prohibited. If you have received this e-mail by mistake, please delete it and immediately contact the sender.
>
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: Searching classes in SNOMEDCT-ICD11

Tania Tudorache
In reply to this post by Campbell, James R
Hi Jim,

We'll also be working on a Lucene search for Protege 5, and that would
also offer an alternative. With Lucene, there will be a one time cost to
create the index (which would be stored locally), but searching would be
very fast after that, and there would be no delay when Protege starts.
We hope to have the Lucene search implemented this year.

Cheers,
Tania



On 04/19/2015 08:29 PM, Campbell, James R wrote:

> Thank you for responding.  I was using the search window and have max 8Gb committed to Protege.  I noticed better response on my desktop with 128Gb but boy, is all that required?
> Jim
>
> James R. Campbell MD
> [hidden email]
> Office: 402-559-7505
> Secretary: 402-559-7299
> Pager: 402-888-1230
>
>> On Apr 19, 2015, at 4:00 PM, "Matthew Horridge" <[hidden email]> wrote:
>>
>> wrote:
> The information in this e-mail may be privileged and confidential, intended only for the use of the addressee(s) above. Any unauthorized use or disclosure of this information is prohibited. If you have received this e-mail by mistake, please delete it and immediately contact the sender.
>
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user