On the uploads page we only show jobs which are within a service’s data
retention.
This commit does the same for when we’re listing the jobs for a contact
list. This matches the UI, which says a contact list has been ‘used
`<count_of_jobs>` in the last <data_retention> days’
Because we’re be grouping jobs under their parent contact lists it’s
good to have some information ‘scent’ to help people find their jobs,
ie by clicking into a contact list. It also lets you see which list have
been used more than others, maybe because the update hasn’t been sent
to that group of people yet.
The hint text under uploads always says when they were used. For contact
lists this is a bit more complicated, since they can:
- never have been used
- been used multiple times
This commit makes use of the new fields being returned by the API to say
determine when these messages are relevant. They also let us
differentiate between a contact list that’s never been used, and one
that has been used, but not recently enough to show any jobs against it.
It’s a bit unintuitive that starting a job from a contact list makes a
copy of the file, which has no relationship to the list it was copied
from. This is more of an implementation detail, rather than something
that comes from people’s mental models of what is going on. Or at least
that’s what I hypothesise.
I think it’s clearer to show jobs that come from contact lists within
the lists that they were created from. By naming the jobs by template
this gives a clearer view of what messages have been sent to the group
over time.
The view layer shouldn’t be having to deal with converting dates as
strings. This is an artefact of how we send data from the API to the
admin app. The model layer should be responsible for turning JSON into
richer types, where it can.
The responses we get to paginated queries from the API are fairly
consistent, so we should be able to reuse the code that takes JSON
from the API and turns it into Python objects. This commits factors out
that code so that it is reusable (by inheriting from it).
If a subclass of `JSONModel` defines a property then we shouldn’t try
to override it with the value from the underlying dictionary.
Rather than silently fail we should raise an exception because it will
help keep our list of `ALLOWED_PROPERTIES` nice and tidy.
For emails and text messages we sort by the time the user (or API) sent
them.
This makes sense for broadcasts too, since most users will receive the
alert within seconds of it being broadcast.
For alerts that haven’t started yet we can sort by `updated_at`, which
is when the user preparing the broadcast submitted it for approval.
We wrote custom `__getattr__` and `__getitem__` implementation to make
it easier to find code that was accessing fields in the dictionary that
we weren’t aware of. See this commit for details:
b48305c50d
Now that no view-layer code accesses the service dictionary directly we
don’t need to support this behaviour any more. By removing it we can
make the model code simpler, and closer to the `SerialisedModel` it
inherits from.
It still needs some custom implementation because:
- a lot of our test fixtures are lazy and don’t set up all the expected
fields, so we need to account for fields sometimes being present in
the underlying dictionary and sometimes not
- we often implement a property that has the same name as one of the
fields in the JSON, so we have to be careful not to try to override
this property with the value from the underlying JSON
Our style for areas is pale blue background with black keylines or bold
black text.
This commit makes the display of area names on the dashboard consistent
with that visual style.
This also means that we’re not truncating the list of areas, which is
appropriate because no one area is more important than any of the
others.
Broadcast services only have broadcast templates. But we show the
template type under the name of the template. This is redundant. It
would be better to preview the content of the template instead.
This then makes the templates page consistent with the dashboard.
Depends on:
- [ ] https://github.com/alphagov/notifications-api/pull/2996
We’ve had some feedback from user research that difference between
‘will get alert’ and ‘likely to get alert’ is not clear, and it’s hard
to tell if the latter is inclusive of the former. This leads people to
question the validity of these numbers, which is important, because an
the estimate should give you some idea of the impact of what you’re
about to do.
This commit reformats the number as a range, for example 1,000 to 2,000
phones.
If the range is small, eg 40,000,000 to 40,800,000 then this suggests
a false level of accuracy. So instead we just give one number and say
it’s an estimate, eg ‘40,000,000 phones estimated’
we want to keep track of all broadcast services across govt easily. As
such, when broadcasting is enabled for a service, we've decided we're
going to add the service to a special broadcasting organisation.
This organisation is defined in the config file. It's hard coded for
production, if you want to test locally, you should set
BROADCAST_ORGANISATION_ID in your local environment.
If you’re adding another area to your broadcast it’s likely to be close
to one of the areas you’ve already added.
But we make you start by choosing a library, then you have to find the
local authority again from the long list. This is clunky, and it
interrupts the task the user is trying to complete.
We thought about redirecting you somewhere deep into the hierarchy,
perhaps by sending you to either:
- the parent of the last area you’d chosen
- the common ancestor of all the areas you’d chosen
This approach would however mean you’d need a way to navigate back up
the hierarchy if we’d dropped you in the wrong place. And we don’t have
a pattern for that at the moment.
So instead this commit adds some ‘shortcuts’ to the chose library page,
giving you a choice of all the parents of the areas you’ve currently
selected. In most cases this will be one (unitary authority) or two
(county and district) choices, but it will scale to adding areas from
multiple different authorities.
It does mean an extra click compared to the redirect approach, but this
is still fewer, easier clicks compared to now.
This meant a couple of under-the-hood changes:
- making `BroadcastArea`s hashable so it’s possible to do
`set([BroadcastArea(…), BroadcastArea(…), BroadcastArea(…)])`
- making `BroadcastArea`s aware of which library they live in, so we can
link to the correct _Choose area_ page
At the moment there are some areas which have:
- a `count_of_phones` value of `None`
- no sub-areas
This is wrong, but until we fix the data the phone counting code needs
to handle this.
This commit:
- adds the `or 0` in the right place (where it will catch these areas
with missing data)
- adds a test which checks these areas, and compares them to other kinds
of areas
We need to give people a better feel for the consequences of
broadcasting an alert. We’ve seen in research that some users will
assume it is subscription based, or opt-in, rather than going to every
phone in the area.
I reckon that the most effective way to communicate this is to put some
numbers next to the areas, to give people an idea of how many people
will get alerted.
We can estimate how many phones are in an area by:
- taking the population of all electoral wards in that area
- multiplying it by the percentage of people who own an internet
connected phone[1]
The Office for National Statistics publish both these datasets.
The number of people who own an intenet connected phone varies a lot by
age. Since the population data for each ward is broken down by age we
can factor this in. Simplified, the calculation looks like this:
- take the _Abbey_ ward of _Barking and Dagenham_
- in this ward there are 26 people aged 80
- 40% of people over 65 have an internet-connected phone
- therefore 10 of these 80-year-olds would be likely to receive a
broadcast
- (repeat for all other ages)
These numbers won’t be exact, but should be enough to give people a feel
for the severity of what they’re about to do. We can see if they acheive
this aim in user research.
1. This is a proxy for the number of people who are likely to have a 4G
capable phone, because only 4G capable phones will be receiving
broadcasts to begin with
When creating broadcast message, or updating it.
We want to send simple polygons to API so we can
later relay them to the broadcast provider.
Test that update broadcast message sends polygons correctly
We have a bunch of stuff for doing lat/long transformation in the
`BroadcastMessage` class. This is not a good separation of concerns, now
that we have a separate class for dealing with polygons and coordinates.
This commit does two things:
- uses our new polygon-simplifying library to process the polygons
before storing them, rather than processing them in real time
- stores only the polygons in the database, rather than the whole
GeoJSON feature, because we don’t need any of the other information
about the feature
Simplifying polygons means reducing the number of points used to render
them. This commit implements simplification such that, for any given
input polygons, the combined point count of the simplified polygons is
less than 100.
When simplifying the polygons we are trying to get the smallest number
of points while meeting these two rules:
1. No part of the area the user has chosen can be cut off
2. The area of the simplified polygon should be as small as possible
This commit introduces two techniques we weren’t using before:
1. Dilating and eroding the area to fill in concave details of the
shape, like inlets and harbours[1]
2. Making the simplification threshold proportionate to the perimeter of
all polygons, so bigger and crinklier polygons get more
simplification applied
It also shows the estimated bleed as a separate polygon. This lets us
make it bigger (so it’s more closer the the approximate bleed) without
having to send a bigger area to the CBC and compounding the amount of
actual bleed.
1. Inspired by this blog post about ‘removing the crinkley bits’ from
Vancouver Island:
http://blog.cleverelephant.ca/2010/11/removing-complexities.html
Since we added the end time picker:
- we have discovered that broadcasts can’t be longer than 24h
- we have observed that most users confuse picking the end time for
scheduling the message, or don’t understand exactly what it means for
the broadcast to ‘end’
- we’ve developed the concept of ‘training mode’, which you should be
going through before sending a real broadcast
We also think that, for most scenarios, you won’t necessarily know when
a broadcast should end at the time of starting it because the cause of
the danger is not within your control. So giving you control of the
end time before the broadcast has even been approved is a confusing
distraction.
Having to pick a time at all also makes the whole process feel more
planned and less immediate. Whereas in reality all the phones in the
area will be getting the message in seconds. It’s only people coming
into the area later to whom the ‘ongoing’ aspect of the broadcast
applies.
The best place to explain what’s happening with the phones is at the
approval stage and once you’ve sent your first (training mode)
broadcast. It’s easier to explain what’s happened if it’s in direct
response to something you’ve just done.
Later on we should add some kind of email reminder after 12 hours to
make sure you still want the broadcast live, again after 18 hours, etc.
We could let you schedule an end time once the broadcast is live, but
don’t think there’s a strong need. Knowing enough that you want to
cancel is one thing, but knowing enough to want to cancel but wanting to
wait a bit… nah.
This shows the green banner with a tick when cancelling a user's
invitation to a service or organisation. The accessibility audit noted
that 'When cancelling an invite a new page loads, however, there is no
immediate indication that the invite has been cancelled.'
In order to display the invited user's email address as part of the
flash message, this adds new methods to the api clients for invites to get
a single invite.
Broadcasting is not a precise technology, because:
- cell towers are directional
- their range varies depending on whether they are 2, 3, 4, or 5G
(the higher the bandwidth the shorter the range)
- in urban areas the towers are more densely packed, so a phone is
likely to have a greater choice of tower to connect to, and will
favour a closer one (which has a stronger signal)
- topography and even weather can affect the range of a tower
So it’s good for us to visually indicate that the broadcast is not as
precise as the boundaries of the area, because it gives the person
sending the message an indication of how the technology works.
At the same time we have a restriction on the number of polygons we
think and area can have, so we’ve done some work to make versions of
polygons which are simplified and buffered (see
https://github.com/alphagov/notifications-utils/pull/769 for context).
Serendipitously, the simplified and buffered polygons are larger and
smoother than the detailed polygons we’ve got from the GeoJSON files. So
they naturally give the impression of covering an area which is wider
and less precise.
So this commit takes those simple polygons and uses them to render the
blue fill. This makes the blue fill extend outside the black stroke,
which is still using the detailed polygons direct from the GeoJSON.
At the moment they will get a ‘technical difficulties’ error if they
try.
We probably want to do something around letting people self-approve
broadcasts in trial mode, but for now just telling them they can’t is a
better experience than ‘technical difficulties’ (and will probably be
close to what they should see on a live service as well).
Different emergencies will need broadcasts to last for a variable amount
of time. We give users some control over this by letting them stop a
broadcast early. But we should also let them set a maximum broadcast
time, for:
- when the duration of the danger is known
- when the broadcast has been live long enough to alert everyone who
needs to know about it
This code re-uses the pattern for scheduling jobs, which has some
constraints that are probably OK for now:
- end time is limited to an hour
- longest duration is 3 whole days (eg if you start broadcasting Friday
you have the choice of Saturday, Sunday and all of Monday, up to
midnight)
If a broadcast definitely shouldn’t go out (for example because it has a
spelling mistake or is going to the wrong areas) then we should have a
way of removing it. Once it’s removed no-one else can approve it, and it
isn’t cluttering up the dashboard.
This is a link (because it’s a secondary action) and red (because it’s
destructive, in that it’s throwing away someone’s work).
Since new broadcasts will go into `pending-approval`, we now need a way
of approving them.
This commit adds a button to this page to start (or approve) the
broadcast. This button is wrapped in a bordered box, to emphasise that
it’s something consequential.
We don’t want one person going full yolo and start broadcasting without
any oversight. This commit changes the flow so that the button on the
‘preview’ page puts the broadcast into `pending-approval`, rather than
directly into `broadcasting`.
There was a lot of duplicate code here for updating the status and other
properties of a broadcast.
This commit abstracts them into reusable methods. They’re named with an
underscore to suggest that they shouldn’t be used externally.
This was broken because current_service doesn’t update itself after
calling the `update` method of the API. So we thought we were changing
the permissions like this:
```
{'email', 'sms', 'letter'}
{'email', 'sms', 'letter', 'broadcast'}
{'sms', 'letter', 'broadcast'}
{'letter', 'broadcast'}
{'broadcast'}
```
But actually we were doing this:
```
{'email', 'sms', 'letter'}
{'email', 'sms', 'letter', 'broadcast'}
{'sms', 'letter'}
{'email', 'letter'}
{'email', 'sms'}
```
This commit changes the code to update the permissions like this:
```
{'email', 'sms', 'letter'}
{'broadcast'}
```
It does so by adding a new method to the service model which changes all
the permissions in one API call, and updates the tests to mock the
underlying API call, not the method on the model.
For services with the broadcast permission this hides:
- the ‘View dashboard’ permission (and defaults it to _checked_) because
all users of broadcast services will need to see the dashboard
- the ‘Manage API keys’ permission (and defaults it to _not checked_)
because we don’t offer an API integration for broadcast services yet
– if we do we won’t want existing users to automatically get the
permission
It relabels:
- the ‘Send’ permission to ‘Prepare and approve’ to match the current,
slightly clunky language on the templates page
- the ‘Manage settings’ label to not refer to ‘usage’ because broadcast
services won’t incur cost
