My local build was not always getting the optional requirement for the jsonschema uri format checker (rfc3987).
The requirement has been added to the requirements_for_test file.
I changed the tests to validate the response schema from the post_notifications tests, this way we can tell if we are breaking the contract.
This showed that the email_from was not returning the entire email address but just the username, that has been corrected here.
Removed the response schema validation tests since they are not being testing in the post notification tests.
Removed the tests for trial mode service for the scheduled tasks and the process job.
Having the validation in the POST notification and create job endpoint is enough.
Updated the test_service_whitelist test because the order of the array is not gaurenteed.
On the frontend, we’re letting users register with international phone
numbers. So we shouldn’t block users from doing this on the API side.
Same thing for the whitelist, where we’re also allowing international
phone numbers now.
previously we didn't do this because the tests all used the same DB
(test_notifications_api), however @minglis shared a snippet that simply
creates one test db per thread.
This was only ever a spike into what it might look like to document
Notify’s API with Swagger (see
7c3d25a87a).
It’s no longer updated, and only talks about version 1 of the public
API.
Keeping it around now is just a liability, and gives us additional Pyup
upgrades to deal with.
specifically, all of the performance platform specific data layout now
happens in performance_platform_client.py - stuff like setting the
_timestamp, period etc, and the perf platform-specific nomenclature is
all handled there.
so that it doesn't appear generic when it's actually specific to
sending the daily notification totals. To do this, split it out into a
separate performance_platform directory, containing the business logic,
and make the performance_platform_client incredibly thin - all it
handles is adding ids to payloads, and sending stats.
Also, some changes to the config (not all done yet) since there is one
token per endpoint, not one for the whole platform as we'd previously
coded
