Wire new dashboard reports with real data #1039
Conversation
…nd' card with dummy data
… registration', and 'Controlled patients trend' graphs
They need to be conditional so they only insert where the new record is greater than the existing recorded_at for the period
| @@ -5,22 +5,43 @@ class Dashboard::DistrictsController < AdminController | |||
|
|
|||
| EXAMPLE_DATA_FILE = "db/data/example_dashboard_data.json" | |||
|
|
|||
| def index | |||
| authorize :dashboard, :view_my_facilities? | |||
| @districts = FacilityGroup.all.order(:name) | |||
There was a problem hiding this comment.
This should use some sort of policy scope to only show what a user has access to
There was a problem hiding this comment.
I cargo culted a scope from another controller, I think its good for this first PR. I'll need to work with someone w/ more experience w/ Pundit to make sure we have this nailed down later.
| (1..number).inject([self]) do |quarters, number| | ||
| quarters << quarters.last.next_quarter | ||
| end | ||
| end |
* estimated population * last updated date * the state breadcrumb
| @@ -0,0 +1,2 @@ | |||
| Time::DATE_FORMATS[:month_year] = "%b %Y" | |||
There was a problem hiding this comment.
Would this break other dashboards?
I assume only if they are not explicitly applying strftime.
There was a problem hiding this comment.
This adds a new format we can use for to_s with time / date via time.to_s(:month_year). It doesn't change the default to_s behavior. Its a handy Rails thing for custom formats, check out https://github.com/rails/rails/blob/master/activesupport/lib/active_support/core_ext/time/conversions.rb.
| def compile_control_and_registration_data | ||
| beginning_of_period = selected_date.advance(months: -11) | ||
| previous_registrations = 0 | ||
| registrations.each do |(period, count)| |
There was a problem hiding this comment.
This a bit complicated to understand, from what I gather from the logic, could this be simply achieved by the following?
previous_registrations =
PatientRegistrationsPerDayPerFacility
.where(facility: @facilities)
.group(:year, :month)
.where('year < ?', beginning_of_period.year)
.or(PatientRegistrationsPerDayPerFacility
.where('month <= ?', beginning_of_period.month))
.sum(:registration_count)
There was a problem hiding this comment.
Yeah this was bad. I'm going to replace it with a query directly on the patients table that Tim suggested elsewhere.
Btw, we couldn't do that sort of query you suggested, because our year and months are stored as TEXT in the views. We could cast it I suppose and then do the comparison, but then we are doing conversion logic in our SQL, which kinda defeats the purpose of the views.
| def compile_control_and_registration_data | ||
| beginning_of_period = selected_date.advance(months: -11) | ||
| previous_registrations = 0 | ||
| registrations.each do |(period, count)| |
There was a problem hiding this comment.
Yeah this was bad. I'm going to replace it with a query directly on the patients table that Tim suggested elsewhere.
Btw, we couldn't do that sort of query you suggested, because our year and months are stored as TEXT in the views. We could cast it I suppose and then do the comparison, but then we are doing conversion logic in our SQL, which kinda defeats the purpose of the views.
| @@ -0,0 +1,2 @@ | |||
| Time::DATE_FORMATS[:month_year] = "%b %Y" | |||
There was a problem hiding this comment.
This adds a new format we can use for to_s with time / date via time.to_s(:month_year). It doesn't change the default to_s behavior. Its a handy Rails thing for custom formats, check out https://github.com/rails/rails/blob/master/activesupport/lib/active_support/core_ext/time/conversions.rb.
rsanheim
dismissed stale reviews from harimohanraj89 and prabhanshuguptagit
done
This wires up the new district dashboard views with live queries.
I decided to not go with the rollup summaries right now, because I want to get these reports live in production so that folks w/ access can review them to see how they look and feel with real data. I've been finding hard to test these with sandbox / seed data, because those datasets vary quite a bit from production data.
NOTE: I realize the code here isn't super great, but I think its a fair trade off since none of will be publicly available yet. The URLs are all hidden and will only be distributed to select users for early-access testing.
Definitions Doc: https://docs.google.com/document/d/1MFRYAuy7pgXFZSH_p0YMLAIyGUldXB0KFYZg2jt8KiY/
Story card: https://www.pivotaltracker.com/story/show/171110822
Because
We want great reports for CVHOs
This addresses
Wiring up a district level report for metrics and benchmarks for CVHOSs