Commit Graph

800 Commits

Author SHA1 Message Date
Rebecca Law
7fa42c2cc5 When SES callback is for a complaint save that to the new complaints table.
When handling the complaint we don't want to throw an exception if the message is missing fields. Only log an exception if we are unable to tie a complaint to a notification.
2018-06-04 17:29:58 +01:00
Katie Smith
c6c118fea1 Update created_at column of ft_billing to be non-nullable
`created_at` was added previously and made nullable temporarily. This
commit now populates the column, ensures that it will always have a
value, and makes `created_at` non-nullable.
2018-05-23 11:09:13 +01:00
Rebecca Law
b405b131f6 Merge pull request #1886 from alphagov/ft_billing-migration-fixes
Fix how the letter rates are selected
2018-05-23 11:07:59 +01:00
Rebecca Law
1567e9bfe5 In order to correctly select the rate for letters you need the billable_unit or page count to use to match on the letter_rates.
A new column is added to the query to populate ft_billing which is only populated for letter.
2018-05-22 17:35:21 +01:00
Katie Smith
0d4fb81235 Delete ProviderStatistics model
This is not being used anywhere, so can be deleted safely.
2018-05-21 15:03:39 +01:00
Katie Smith
71796311fa Delete unused aggregate_statistics endpoint
`@service_blueprint.route('/<uuid:service_id>/fragment/aggregate_statistics')`
is not being used anywhere, so has been removed. The `provider_statistics_dao`
can also be removed, since the function this contained was only used for
the endpoint.
2018-05-21 15:03:39 +01:00
Rebecca Law
e5584348ef Add rate to the primary key of ft_billing.
Use the primary key constraint in the upsert statement (easier to read than listing all the columns)
2018-05-21 14:38:25 +01:00
Rebecca Law
c2b888a6af Make sure the date does not include time 2018-05-21 11:51:06 +01:00
Rebecca Law
4d8255b8d9 Merge branch 'master' into fix-ft-billing-migration 2018-05-21 11:06:31 +01:00
Rebecca Law
9fad623d91 Update the ft_billing query to only include billable notification status types.
Update test_provider_statistics dao - this is really irrelevant since the endpoint using the query is not being used. We have a PR coming to delete the unused code.
Update rate_multiplier to always be an integer
2018-05-21 10:56:16 +01:00
Rebecca Law
3c90c0ffcd Fix code style 2018-05-16 13:18:36 +01:00
Rebecca Law
eb083e30ed - Only rebuild current month for monthly_billing if today is in the current year.
- Change the usage queries to a union so that billing_units is correct for all notification types. Removing the business logic from the schema.
- Added tests for different fragment counts, rates and sheet counts.
2018-05-16 12:21:59 +01:00
Rebecca Law
271ce6d76e Changed the update/insert to a postgres upsert to avoid concurrency issues. 2018-05-15 11:21:10 +01:00
Rebecca Law
3615f3d00f In order to re-run the create_nightly_billing for dates in the past, we added a condition on which table is used.
This will allow us to re-run nightly billing for those 2 days where it failed.
For the majority of time the query will run on Notiifcations.
2018-05-14 16:21:16 +01:00
Rebecca Law
a640605a7c Merge pull request #1867 from alphagov/use-ft-billing-for-usage
Update ft_billing table to use integer data types
2018-05-14 13:04:52 +01:00
Chris Waszczuk
0e9fca09fb Merge pull request #1855 from alphagov/inbound-sms-max-7-days
Only get inbound messages that are a maximum of 7 days old
2018-05-14 11:15:52 +01:00
Rebecca Law
d98581cfe6 Added a new endpoint for yearly usage totals using ft_billing. 2018-05-11 16:25:16 +01:00
Rebecca Law
8028f6cc28 We found that the reporting task failed twice because of integrity constraints.
This was because the rate_multiplier was being added as 1 and 1.0 which was not resolving to the same.
This updates the table to use Integrer.
Also changed the logging for the task.
2018-05-10 15:35:58 +01:00
chrisw
2abbb590ab Only get inbound messages that are a maximum of 7 days old 2018-05-10 14:00:55 +01:00
Rebecca Law
3e3b885bdc Realised that it's best to cast the Month as date. 2018-05-08 13:53:44 +01:00
Rebecca Law
ea3523199a New endpoint to get monthly billing usage from the ft_billing table.
New command to compare the results of monthly billing to ft_billing.
2018-05-04 13:09:14 +01:00
Rebecca Law
18c2b9a56d Use better date function to get the first of each month.
Build the json object to return for the new endpoint.
2018-05-02 09:36:13 +01:00
Rebecca Law
0fb9c1d318 Add notification_type to query 2018-05-02 09:36:13 +01:00
Rebecca Law
a9c5ba7674 Remove the truncate date function for the query. 2018-05-01 16:48:34 +01:00
Rebecca Law
93c7ab6251 Because dates are confusing we need to fix this query once more.
The template statistics are returning 7 days inclusive, however the big numbers on the dashboard are for 8 days.
This PR fixes that.
2018-05-01 10:47:47 +01:00
Katie Smith
f2d4bc795e Add DAO function and endpoint for archiving letter contact blocks
Added a new DAO function which archives letter contact blocks by
setting archived to True. This raises an ArchiveValidationError if
trying to archive the default letter block for a service or the default
letter contact block for a template.

Added a new endpoint for archiving letter contact blocks.
2018-04-30 15:25:17 +01:00
Katie Smith
472b86f3f4 Add DAO function and endpoint for archiving SMS senders
Added a new DAO function which archives SMS senders by setting
archived to True. This raises an ArchiveValidationError, if
trying to archive a default SMS sender or an inbound number.

Added a new endpoint for archiving SMS senders.
2018-04-30 15:25:17 +01:00
Katie Smith
5f43fe23a7 Add DAO function and endpoint for archiving email reply_to addresses
Added a new DAO function which archives email reply_to addresses by
setting archived to True. This raises a new type of error, an
ArchiveValidationError, if trying to archive a default reply_to address.

Added a new endpoint for archiving email reply_to addresses.
2018-04-30 15:25:17 +01:00
Katie Smith
663021e494 Only return active SMS senders
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.
2018-04-30 15:25:17 +01:00
Katie Smith
f810daa3c5 Only return non-archived letter contact blocks
Changed DAO functions which return one letter contact block and all
letter contact blocks for a service to only return non-archived ones.
2018-04-30 15:25:17 +01:00
Katie Smith
a57b2f50c2 Only return non-archived email reply_to addresses
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.
2018-04-30 15:25:17 +01:00
Leo Hemsted
0efa223fb2 rename days_ago to midnight_n_days_ago
also add some more timezone boundary tests and minor code cleanup
2018-04-30 11:50:56 +01:00
Leo Hemsted
310b8eda4c refactor dao_get_template_usage now that limit_days no longer used
doesnt need to have any conditional stuff now - always requires a date
2018-04-30 11:13:21 +01:00
Leo Hemsted
7dc34fc3a4 set 0'd dict in redis if no data exists in redis
to get the data for a day can be reasonably slow (a few hundred
milliseconds), and if someone's viewing a service with no activity we
don't want to do that query seven times every two seconds. So if there
is no data in redis, when we get the data out of the database, we
should put it in redis so we can just grab it from there next time.

This'll happen in two cases:
* redis data is deleted
* the service sent no messages that day

additionally, make sure that we convert nicely from redis' return
values (ascii strings) to unicode keys and integer counts.
2018-04-30 11:13:21 +01:00
Leo Hemsted
9e8b6fd00d refactor template stats endpoint to read from new redis keys
New redis keys are partitioned per service per day. New process is as
follows:

* require a count of days to filter by. Currently admin always gives 7.
* for each day, check and see if there's anything in redis. There won't
  be if either a) redis is/was down or b) the service didn't send any
  notifications that day
  - if there isn't, go to the database and get a count out.
* combine all these stats together
* get the names/template types etc out of the DB at the end.
2018-04-30 11:13:21 +01:00
Leo Hemsted
5e702449cb move days_ago to utils and make it tz aware
it's used in a few places - it should definitely know what timezones
are and return datetimes rather than dates, which are hard to work with
in terms of figuring out how tz aware they are.
2018-04-30 11:13:21 +01:00
Leo Hemsted
67019fc5a1 add day filter to template usage dao, and remove NotificationHistory
we should be very careful with when we get data from
NotificationHistory - this should probably only be from scheduled
tasks. `dao_get_template_usage` is only called from the template
statistics rest endpoint, so shouldn't ever hit template history.

also, moved tests out to new file to break up the 2k test file a bit
2018-04-30 11:13:21 +01:00
Chris Waszczuk
cc6e4fabe1 Merge pull request #1839 from alphagov/make-service-stats-7-days-max
Adjust get_service_stats to get counts for a maximum of 7 days ago
2018-04-26 15:10:40 +01:00
Rebecca Law
51af6b27a0 Update name of method to clarify the meaning.
Update group by to use label.
Update test.
2018-04-25 17:00:19 +01:00
Rebecca Law
cff731c52d Set filter on query 2018-04-25 16:17:56 +01:00
Rebecca Law
d00614205e Adding test for fetch_biling_data_for_day. 2018-04-25 14:45:51 +01:00
Rebecca Law
16ef133aa5 Refactor reporting task so that methods can be used for the annual usage page.
Still a work in progress, tests are coming.
2018-04-25 14:45:51 +01:00
Rebecca Law
bd91aac763 Write some pseudo code for ft billing 2018-04-25 14:45:51 +01:00
Rebecca Law
8cd4e1be29 Query ft_billing aggregate for a month 2018-04-25 14:45:51 +01:00
chrisw
c157de82c4 Adjust get_service_stats to get counts for a maximum of 7 days ago 2018-04-24 16:05:43 +01:00
chrisw
0f2ddd8cfa added new paginated inbound endpoint 2018-04-04 17:10:41 +01:00
Rebecca Law
e47b9364d9 Now that we are in British Summer Time the delete notification scheduled tasks were running at 23:40, but the code expects that the task runs the following day.
Which resulted in the notifications being around for 8 days.
This PR fixes that.
2018-04-04 12:20:36 +01:00
Rebecca Law
f596d17bf2 If a sms or email has not been sent after 4 hours and 15 minutes then put it on the delivery queue. 2018-03-23 15:38:35 +00:00
chrisw
ac50250ad7 update paginated inbound sms method 2018-03-23 11:09:51 +00:00
Rebecca Law
0798154fa2 Optimize the query used to return the services and todays notification totals.
By changing the created_at filter to use a specific date range I found a significant improvement to the queries performance.
The unit test needed to change because now were are returning todays date as BST the local timezone.

Query plan before

Merge Left Join  (cost=1226133.76..1226143.77 rows=1753 width=70) (actual time=5800.160..5801.657 rows=1849 loops=1)
Merge Cond: (services.id = anon_1.service_id)
->  Sort  (cost=152.99..157.37 rows=1753 width=46) (actual time=2.205..2.631 rows=1762 loops=1)
Sort Key: services.id
Sort Method: quicksort  Memory: 224kB
->  Seq Scan on services  (cost=0.00..58.54 rows=1753 width=46) (actual time=0.011..1.156 rows=1762 loops=1)
Filter: active
Rows Removed by Filter: 101
->  Sort  (cost=1225980.77..1225980.99 rows=86 width=40) (actual time=5797.949..5797.984 rows=198 loops=1)
Sort Key: anon_1.service_id
Sort Method: quicksort  Memory: 40kB
->  Subquery Scan on anon_1  (cost=1225976.29..1225978.01 rows=86 width=40) (actual time=5797.682..5797.823 rows=198 loops=1)
->  HashAggregate  (cost=1225976.29..1225977.15 rows=86 width=48) (actual time=5797.681..5797.747 rows=198 loops=1)
Group Key: notifications.notification_type, notifications.notification_status, notifications.service_id
->  Seq Scan on notifications  (cost=0.00..1220610.86 rows=536543 width=48) (actual time=0.064..5482.975 rows=643799 loops=1)
Filter: (((key_type)::text <> 'TEST'::text) AND (date(created_at) = '2018-03-20'::date))
Rows Removed by Filter: 6804774
Planning time: 1.106 ms
Execution time: 5802.130 ms

Query plan after
Merge Left Join  (cost=953378.30..953388.30 rows=1753 width=70) (actual time=2380.144..2382.499 rows=1852 loops=1)
Merge Cond: (services.id = anon_1.service_id)
->  Sort  (cost=152.99..157.37 rows=1753 width=46) (actual time=2.944..3.570 rows=1762 loops=1)
Sort Key: services.id
Sort Method: quicksort  Memory: 224kB
->  Seq Scan on services  (cost=0.00..58.54 rows=1753 width=46) (actual time=0.006..1.294 rows=1762 loops=1)
Filter: active
Rows Removed by Filter: 101
->  Sort  (cost=953225.31..953225.53 rows=86 width=40) (actual time=2377.194..2377.262 rows=201 loops=1)
Sort Key: anon_1.service_id
Sort Method: quicksort  Memory: 40kB
->  Subquery Scan on anon_1  (cost=953220.83..953222.55 rows=86 width=40) (actual time=2376.797..2377.034 rows=201 loops=1)
->  HashAggregate  (cost=953220.83..953221.69 rows=86 width=48) (actual time=2376.795..2376.905 rows=201 loops=1)
Group Key: notifications.notification_type, notifications.notification_status, notifications.service_id
->  Bitmap Heap Scan on notifications  (cost=29883.14..947856.24 rows=536459 width=48) (actual time=270.061..1887.754 rows=644735 loops=1)
Recheck Cond: ((created_at >= '2018-03-20 00:00:00'::timestamp without time zone) AND (created_at < '2018-03-21 00:00:00'::timestamp without time zone))
Rows Removed by Index Recheck: 947427
Filter: ((key_type)::text <> 'TEST'::text)
Heap Blocks: exact=40882 lossy=186483
->  Bitmap Index Scan on ix_notifications_created_at  (cost=0.00..29749.02 rows=536459 width=0) (actual time=258.631..258.631 rows=644849 loops=1)
Index Cond: ((created_at >= '2018-03-20 00:00:00'::timestamp without time zone) AND (created_at < '2018-03-21 00:00:00'::timestamp without time zone))
Planning time: 0.548 ms
Execution time: 2383.485 ms
2018-03-20 15:48:32 +00:00