2019-10-07 Meeting notes

• Andrew Atterton has finished off the tickets on validation he had outstanding leaving just the tickets George Swinnerton needs to complete. He will aim to finish these off today ready for the validation meeting tomorrow. This should allow the definition of the realease to be completed in the meeting. The original bamboo issue has been fixed however it has caused another build issue where an imported library to to an imported library appears not to have used a pinned version and is now breaking. The longterm solution to these issues would be to move to python 3 however ther are issues with the SVN libraries and the version of svn repo used which will need resolving or the dependency needs to be removed by finding an alternative solution.
• james.griffin (Unlicensed)  has made a start on the RaDaR tickets reloading the Meds now the importer is fixed. and RDR-663 removing the To Field on the Comorbidity page which is deployed to staging for review. He is now looking coloring of red dead patients by checking they have a data of death on the radar record by preference and not just the last updated record.
• George Swinnerton has some code for the duplicate radar memberships ready for testing and for RaDaRs case the plan is to delete them all and let RaDaR recreate the memberships however it needs to be verified that this will work/can be triggered. He has also realised the test process is inaccurate because the data that is used for testing is not necessarily matching what would be created when a file is imported meaning that tests may not be testing what we think. ideally the test database should be populated using the code that is used in the live system in some way. This needs more work.
• Kings feed Fran has sent off/is sending off the first quarters error file. An issue was found with the extract where it was using a laborder date and not the specimen date for resultitems. This has now been corrected. Fran has also noted there are no BPs in the file George is unclear what this means as there are in dialysis sessions but may not be in general results? Also there is a need to add the registry numbers to patients which will require a process similar to that used in PV to sync memberships that is triggered when the demographics are updated. This will need osme development work.

 George Swinnerton

 George Swinnertonhighlighted an issue where a DM incorrectly assigned a date of death to the wrong patient. This has resulted in their PV account being deleted and a number of errors with the audit files. The DM appears to have edited the incoming file based on mismatching a patient number. Fran has asked about whether there is any in validation highlighting that might help highlight this besides the errors that were reported due to this error 

Tim Whitlock (Deactivated)

• Simplified is still not corrected and appears to need the models library updated in the code. Currently it is still querying for the abormalflags which was renamed with the latest update. George Swinnertonwill sort this this week.

Tim Whitlock (Deactivated)

MyPY is a static type checker if the code is suitably annotated George Swinnerton has looked as this MyPy and highlighted some potential issues due to the way we use imports and also was not sure on what command line should be used in tox for it. Dropbox have used it to great effect https://blogs.dropbox.com/tech/2019/09/our-journey-to-type-checking-4-million-lines-of-python/ not accessible on hospital network and describe how they have used it.

Barney has found some more EQUAL patients to be matched which George Swinnertonwill do this week.

George Swinnerton