By serialising these straight away we can:
- not go back to the database later, potentially closing the connection
sooner
- potentially cache the serialised data, meaning we don’t touch the
database at all
Soon enough every service will have this permission, and they won’t be
able to switch it off. So we should clean up our codebase and make it
so there’s no dependancy on a row existing in the permissions table.
This is the first step of that process for the API. Before we can remove
it, we have to stop checking from it. Next step will be to stop
inserting the permission, then finally remove it from the database.
When a precompiled letter is sent via the admin app, we now pass in the address which can be set in the Notifications.to field.
Once a precompiled letters sent by the API has passed validation we can set the address in Notifications.to field.
The celery tasks to validate precompiled letters sent by the API will be done in another PR.
Now we consistently use the created_at date, so we can always get the right file location and name.
The previous updates to this code were trying to solve the problem if a pdf being created at 17:29, but not ready to upload until 17:31 after the antivirus and validation check.
But in those cases we would have trouble finding the file.
This function checks various permissions, downloads the PDF from the
transient bucket, creates the notification then moves the letter to the
'normal' bucket.
Letters should always have a reference, because that’s what DVLA use to
tell us when they’ve sent a letter.
If a letter has a reference of `None` then DVLA say they’ve sent a
letter with a reference of `'None'`. This means we can never reconcile
the letter, which means it stays in `created`, which means it never
gets billed.
We don’t think this has affected any real letters yet, just ones that
we’ve sent as tests.
This commit modifies the code paths the admin app uses to send one off
emails and text messages to also accept letters.
This mostly worked already, the two changes were:
- making sure that one-off letters are processed by the correct task,
from the correct queue
- one-off letters sent from a service in research mode don’t get put on
a queue and go straight to `delivered` (because we don’t want to send
them for real)
Updated the DAO methods which return a single SMS sender and all SMS senders
to only return the non-archived senders. Changed the error raised in the Admin
interface from a SQLAlchemyError to a BadRequestError.
Updated the DAO methods which return a single email reply_to address and
all reply_to addresses to only return the non-archived addresses.
Changed the type of error that gets raised when using the Admin
interface to be BadRequestError instead of a SQLAlchemyError.
The vast majority of messages that are being sent one-off are
time-sensitive. A typical example is a caseworker on the phone who sends
a message at the end of the call. They normally wait until the message
has been delivered, so all the time they’re waiting is time when they
can’t be helping someone else.
What we don’t want to happen is for the messages they’re sending to get
stuck behind a big lump of GOV.UK Subscription emails or passport
reminder texts. I think the best way to do this is shift them onto the
priority queue.
We’re currently seeing queue sizes of up to 5,000 on the ‘normal’
queues; I don’t think there’s any risk of this change making the
priority queue more heavily-laden than this. Especially since the
traffic patterns of users sending one-off messages won’t be spiky.
The whitelist was built to help developers and designers making
prototypes to do realistic usability testing of them, without having to
go through the whole go live process.
These users are sending messages using the API. The whitelist wasn’t
made available to users uploading spreadsheets. The users sending one
off messages are similar to those uploading spreadsheets, not those
using the API. Therefore they shouldn’t be able to use the whitelist to
expand the range of recipients they can send to.
Passing the argument through three methods doesn’t feel that great, but
can’t think of a better way without major refactoring…
currently, they're made by creating a one-line job, but we want to
reduce task/csv file noise so we're moving them to persist in the
same vein as API usage. However, we can't just call through to that
since there are some differences:
* no api keys
* tighter control over API format
* no scheduling
* no client references
etc.
So, re-using as much of the v2 validation stuff as possible, I've
created this file that just does basic validation, and then calls
through to persist_notification and schedules a task. Woo.
