2021-02-22 Meeting notes

• Andy hasn't managed to progress James xml tracing work.
• George hasn't heard back from renalware but first he needs to complete the schema changes and push them to github before chasing too hard.
• NHSWales changes still need to be made to the codelists on the ukrdc. These should be pushed as part of the next pkb deploy. This should mean we are able to test files supplied from wales. It is likely we will receive this some time in March.
• PYXBx to start using this the schema libraries need to be rebuilt using the new version and rereleased. Then the dependencies updated to the new library. Joel has not found where/how the library gets built and  was awaiting George's return to explain the process. In theory that is the only thing preventing its use. In addition it may be that the strick rule that prevents a schema built with a different version of pyxb should apply as our version is just ensuring python version compatibility.
• Andrew has contacted Afonso and checked he has all the table information he needs.
• Joel has a pull request in to move from using devpi in the deployment to using specific git commits. This would then ensure the accidental deployment of the wrong code will not happen as it has with the current devpi approach.
• Pentesting we should be receiving the Internet report this week and the N3 report Early next week for our review. Early indications are nothing significant has been found except for the API issue found on PV live which SSG are currently trying to resolve.
• The final validation release is waiting on the some final testing once this has been done and signed off it should be possible to make a release. Shaun asked what needed testing and George highlighted the checking of the CKD1/2/3 classifier to ensure it works as expected. SHaun would need to create some test files for this and suggested Fran may be in a better position to test this sooner.
• Andre has almost completed the changes requested by the Magic team to add multiple unit associations to users instead of the current one to one relation. This she hopes to demonstrate to the magic team this Friday.
• Andre has reloaded all the (loaded) AKI files into a new database. This amounts to around 8.5K files of which 245 failed for some reason. It seems from the files so far investigated this was because the date format failed because it had a two digit year. Also a number of creatine files didn't load because the associated Alert file was missing. Andre is currently working through the reasons files failed to load and identifying anything that needs to be queried with Sarah. Once this is completed the data can be compared with the current live  database and checks made that the missing data is now present.
• No news on the Star registry.
• Andrew has completed the NHSBT file for this quarter.
• Nothing significant to report on RaDaR tickets this week.

 Joel highlighted a couple of things to do with checking code for issues.

github now has a free security scanning service which can be applied to our repos which will help flag if a security issue has been flagged for any libraries used etc. This was a paid for service but is now free. Joel will turn it on at the organisation level and document how it can be enabled for a repo.

bandit (Welcome to Bandit’s developer documentation! — Bandit documentation) is a tool designed to find common security issues in Python code. As the current Bitbucket module has not caught all issues it would seem sensible to add this library to the dev requirements and ensure it gets run. Whilst this may not catch all issues it should help and can be augmented using test plugins.