Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[bioontology-support] Can't generate ChEBI mappings in my OntoPortal 3 VM

Jennifer Leigh Vendetti vendetti at stanford.edu
Thu Sep 10 17:42:10 PDT 2020


Hi Mark,


On Sep 9, 2020, at 6:45 AM, Miller, Mark <markampa at pennmedicine.upenn.edu<mailto:markampa at pennmedicine.upenn.edu>> wrote:

Currently, one of my main uses for BioPortal/OntoPortal is retrieving all mappings between ChEBI, DrOn, and RxNorm. I prefer to do that from the VM because it’s faster and I can load new versions as soon as they become available. And it’s better citizenship, right?

I recently upgraded to OntoPortal 3, and I can’t seem to get mappings to or from ChEBI.

  *
     *
        *
           *
              *


OK, I’ve tracked down the underlying reason that you’re seeing this behavior.

In order to retrieve all of the mappings between two ontologies, the system depends on the existence of a persistent count of the number of mappings for each ontology. You can see these persisted counts using a REST call like the following:

http://your_appliance_id:8080/mappings/statistics/ontologies

These counts are generated and stored as part of a cron job that is configured to run once every Saturday, i.e.:

https://github.com/ncbo/ncbo_cron/blob/8c5d9f75ae62e5d311814d4ee5c22b12dccdc966/lib/ncbo_cron/config.rb#L58

The historical reasons for this have to do with the 4store backend triplestore that is used in OntoPortal’s stack. Unfortunately, COUNT queries in 4store aren’t performant and we’ve had to add some workarounds in the system that prevent these types of queries from happening frequently in the production BioPortal that we host at Stanford. These cron job settings were carried over into our OntoPortal virtual appliance distribution.

All of that is to say that you have 2 options for seeing the mappings between ontologies in your appliance:

1). The easiest option is to wait until after the cron job has executed on Sat. at 12:30am, after which time you’ll be able to see all of the mappings between ontologies. (You may have to clear the caches via the Admin page).

2). If you have an urgent need to see the mappings, you can go the more difficult route and manually modify the schedule for this cron job. To follow are the instructions for how you would do this from the administration console:

a). Change to the ontoportal user:

sudo su - ontoportal

b). Shut down the current running cron process:

cd /srv/ontoportal/ncbo_cron/
bin/ncbo_cron —kill

c). Edit the crontab schedule for the “create_mapping_counts” job on line 56 of the configuration file (/srv/ontoportal/ncbo_cron/lib/ncbo_cron/config.rb). This page is helpful for showing examples of crontab schedule syntax: https://crontab.guru/examples.html.

d). After editing / saving / closing the configuration file, reboot your appliance.

After rebooting your appliance, the cron job scheduler will have picked up your new setting on how often you’d like to run the cron_mapping_counts job. If you want to check the status of whether the job is running successfully, the log output is stored here: /srv/ontoportal/ncbo_cron/logs/scheduler-mapping-counts.log.

Kind regards,
Jennifer






-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/bioontology-support/attachments/20200911/856cb8c9/attachment.html>


More information about the bioontology-support mailing list