Add multiple versions of group annotation types

These are:

 * By the group
 * By the group - bubbled up to courses
 * By the organization

lms/data_tasks/report/create_from_scratch/00_schema/03_create_fdw_schema.sql

@@ -4,21 +4,25 @@ CREATE SCHEMA h;
 
 -- These types are required to successfully import the group table
 DROP TYPE IF EXISTS public.group_joinable_by;
-DROP TYPE IF EXISTS public.group_readable_by;
-DROP TYPE IF EXISTS public.group_writeable_by;
-
 CREATE TYPE public.group_joinable_by AS ENUM (
     'authority'
 );
 
+DROP TYPE IF EXISTS public.group_readable_by;
 CREATE TYPE public.group_readable_by AS ENUM (
     'authority', 'members', 'world'
 );
 
+DROP TYPE IF EXISTS public.group_writeable_by;
 CREATE TYPE public.group_writeable_by AS ENUM (
     'authority', 'members'
 );
 
+DROP TYPE IF EXISTS public.annotation_sub_type;
+CREATE TYPE report.annotation_sub_type AS ENUM (
+    'annotation', 'reply', 'highlight', 'page_note'
+);
+
 IMPORT FOREIGN SCHEMA "public" LIMIT TO (
     user_group,
     "user",
@@ -28,5 +32,6 @@ IMPORT FOREIGN SCHEMA "public" LIMIT TO (
 IMPORT FOREIGN SCHEMA "report" LIMIT TO (
     authorities,
     annotation_group_counts,
+    annotation_type_group_counts,
     annotation_user_counts
 ) FROM SERVER "h_server" INTO h;

lms/data_tasks/report/create_from_scratch/04_activity_counts/01_groups/01_group_activity/02_initial_fill.sql

@@ -5,4 +5,5 @@ REFRESH MATERIALIZED VIEW report.group_activity;
 ANALYSE report.group_activity;
 
 -- A unique index is mandatory for concurrent updates used in the refresh
-CREATE UNIQUE INDEX group_activity_created_week_group_id_idx ON report.group_activity (created_week, group_id);
+CREATE UNIQUE INDEX group_activity_created_week_group_id_idx
+    ON report.group_activity (group_id, created_week);

lms/data_tasks/report/create_from_scratch/04_activity_counts/01_groups/02_group_bubbled_activity/01_create_view.sql

@@ -7,10 +7,10 @@ CREATE MATERIALIZED VIEW report.group_bubbled_activity AS (
     SELECT
         created_week,
         group_to_group.parent_id AS group_id,
-        SUM(annotation_count) as annotation_count,
-        SUM(annotation_shared_count) as annotation_shared_count,
-        SUM(annotation_replies_count) as annotation_replies_count,
-        SUM(launch_count) as launch_count
+        SUM(annotation_count) AS annotation_count,
+        SUM(annotation_shared_count) AS annotation_shared_count,
+        SUM(annotation_replies_count) AS annotation_replies_count,
+        SUM(launch_count) AS launch_count
     FROM report.group_to_group
     -- The join on group to group without a relation includes self,
     -- so the child_id here is all children and ourselves

lms/data_tasks/report/create_from_scratch/04_activity_counts/01_groups/02_group_bubbled_activity/02_initial_fill.sql

@@ -5,4 +5,5 @@ REFRESH MATERIALIZED VIEW report.group_bubbled_activity;
 ANALYSE report.group_bubbled_activity;
 
 -- A unique index is mandatory for concurrent updates used in the refresh
-CREATE UNIQUE INDEX group_bubbled_activity_created_week_group_id_idx ON report.group_bubbled_activity (created_week, group_id);
+CREATE UNIQUE INDEX group_bubbled_activity_created_week_group_id_idx
+    ON report.group_bubbled_activity (group_id, created_week);

lms/data_tasks/report/create_from_scratch/04_activity_counts/01_groups/03_group_bubbled_counts/02_initial_fill.sql

@@ -5,4 +5,5 @@ REFRESH MATERIALIZED VIEW report.group_bubbled_counts;
 ANALYSE report.group_bubbled_counts;
 
 -- A unique index is mandatory for concurrent updates used in the refresh
-CREATE UNIQUE INDEX group_bubbled_counts_group_id_idx ON report.group_bubbled_counts (group_id);
+CREATE UNIQUE INDEX group_bubbled_counts_group_id_idx
+    ON report.group_bubbled_counts (group_id);

lms/data_tasks/report/create_from_scratch/04_activity_counts/01_groups/04_group_type_counts/01_create_view.jinja2.sql

@@ -0,0 +1,16 @@
+DROP MATERIALIZED VIEW IF EXISTS report.group_type_counts CASCADE;
+
+-- A weekly count of group annotation types limited to the correct authority
+
+CREATE MATERIALIZED VIEW report.group_type_counts AS (
+    SELECT
+        created_week,
+        group_id,
+        sub_type,
+        shared,
+        count
+    FROM h.annotation_type_group_counts
+    JOIN h.authorities ON
+        annotation_type_group_counts.authority_id = authorities.id
+        AND authorities.authority = '{{ region.authority }}'
+) WITH NO DATA;

lms/data_tasks/report/create_from_scratch/04_activity_counts/01_groups/04_group_type_counts/02_initial_fill.sql

@@ -0,0 +1,9 @@
+DROP INDEX IF EXISTS report.group_type_counts_created_week_group_id_sub_type_shared_idx;
+
+REFRESH MATERIALIZED VIEW report.group_type_counts;
+
+ANALYSE report.group_type_counts;
+
+-- A unique index is mandatory for concurrent updates used in the refresh
+CREATE UNIQUE INDEX group_type_counts_created_week_group_id_sub_type_shared_idx
+    ON report.group_type_counts (group_id, sub_type, shared, created_week);

lms/data_tasks/report/create_from_scratch/04_activity_counts/01_groups/05_group_bubbled_type_counts/01_create_view.sql

@@ -0,0 +1,20 @@
+DROP VIEW IF EXISTS report.group_bubbled_type_counts CASCADE;
+
+-- A weekly count of annotation type, but with counts summed
+-- up and down between parents and children as appropriate.
+
+CREATE MATERIALIZED VIEW report.group_bubbled_type_counts AS (
+    SELECT
+        created_week,
+        group_to_group.parent_id AS group_id,
+        sub_type,
+        shared,
+        SUM(count) AS count
+    FROM report.group_to_group
+    -- The join on group to group without a relation includes self,
+    -- so the child_id here is all children and ourselves
+    JOIN report.group_type_counts ON
+        group_type_counts.group_id = group_to_group.child_id
+    GROUP BY created_week, group_to_group.parent_id, sub_type, shared
+    ORDER BY created_week, group_to_group.parent_id, sub_type, shared
+) WITH NO DATA;

lms/data_tasks/report/create_from_scratch/04_activity_counts/01_groups/05_group_bubbled_type_counts/02_initial_fill.sql

@@ -0,0 +1,9 @@
+DROP INDEX IF EXISTS report.group_bubbled_type_counts_created_week_group_id_idx;
+
+REFRESH MATERIALIZED VIEW report.group_bubbled_type_counts;
+
+ANALYSE report.group_bubbled_type_counts;
+
+-- A unique index is mandatory for concurrent updates used in the refresh
+CREATE UNIQUE INDEX group_bubbled_type_counts_created_week_group_id_idx
+    ON report.group_bubbled_type_counts (group_id, sub_type, shared, created_week);

lms/data_tasks/report/create_from_scratch/04_activity_counts/03_organization_activity/02_initial_fill.sql → lms/data_tasks/report/create_from_scratch/04_activity_counts/03_organizations/01_organization_activity/02_initial_fill.sql

@@ -6,5 +6,7 @@ REFRESH MATERIALIZED VIEW report.organization_activity;
 ANALYSE report.organization_activity;
 
 -- A unique index is mandatory for concurrent updates used in the refresh
-CREATE UNIQUE INDEX organization_activity_timescale_period_role_organization_id_idx ON report.organization_activity (timescale, period, role, organization_id);
-CREATE INDEX organization_activity_start_date_end_date_idx ON report.organization_activity (start_date, end_date);
+CREATE UNIQUE INDEX organization_activity_timescale_period_role_organization_id_idx
+    ON report.organization_activity (timescale, period, role, organization_id);
+CREATE INDEX organization_activity_start_date_end_date_idx
+    ON report.organization_activity (start_date, end_date);

lms/data_tasks/report/create_from_scratch/04_activity_counts/03_organizations/02_organization_annotation_types/01_create_view.sql

@@ -0,0 +1,57 @@
+DROP MATERIALIZED VIEW IF EXISTS report.organization_annotation_types CASCADE;
+
+-- A multifaceted look at organization annotation types over different
+-- timescales
+
+CREATE MATERIALIZED VIEW report.organization_annotation_types AS (
+    WITH
+        -- In order to get all the combinations of different types of data we
+        -- want in this table without writing an explosion of queries we first
+        -- create a "facets" CTE which represents all the different
+        -- combinations of values we want to be able to filter by.
+
+        -- We could probably do without this, but it makes it more like the
+        -- `organization_activity` query
+        weeks AS (
+            SELECT DISTINCT(created_week) AS timestamp_week
+            FROM report.group_type_counts
+        ),
+
+        timescales AS (
+            SELECT column1 AS timescale FROM (
+                VALUES
+                    ('week'),
+                    ('month'),
+                    ('semester'),
+                    ('academic_year'),
+                    ('trailing_year'),
+                    ('all_time')
+            ) AS data
+        ),
+
+        facets AS (
+            SELECT
+                timestamp_week,
+                timescale::report.academic_timescale,
+                report.multi_truncate(timescale, timestamp_week) AS period
+            FROM weeks
+            CROSS JOIN timescales
+        )
+
+    SELECT
+        timescale,
+        period::DATE AS start_date,
+        (period + report.single_interval(timescale::text))::DATE AS end_date,
+        report.present_date(timescale::text, period) AS period,
+        group_map.organization_id,
+        group_type_counts.sub_type,
+        group_type_counts.shared,
+        SUM(group_type_counts.count) AS count
+    FROM facets
+    JOIN report.group_type_counts ON
+        group_type_counts.created_week = facets.timestamp_week
+    JOIN report.group_map ON
+        group_map.group_id = group_type_counts.group_id
+    GROUP BY period, timescale, sub_type, shared, organization_id
+    ORDER BY period, timescale, sub_type, shared, organization_id
+) WITH NO DATA;

lms/data_tasks/report/create_from_scratch/04_activity_counts/03_organizations/02_organization_annotation_types/02_initial_fill.sql

@@ -0,0 +1,12 @@
+DROP INDEX IF EXISTS report.organization_annotation_types_period_timescale_sub_type_idx;
+DROP INDEX IF EXISTS report.organization_annotation_types_start_date_end_date_idx;
+
+REFRESH MATERIALIZED VIEW report.organization_annotation_types;
+
+ANALYSE report.organization_annotation_types;
+
+-- A unique index is mandatory for concurrent updates used in the refresh
+CREATE UNIQUE INDEX organization_annotation_types_period_timescale_sub_type_idx
+    ON report.organization_annotation_types (organization_id, timescale, sub_type, shared, period);
+CREATE INDEX organization_annotation_types_start_date_end_date_idx
+    ON report.organization_annotation_types (start_date, end_date);

lms/data_tasks/report/create_from_scratch/99_grant_permissions/01_grant_schema.jinja2.sql

@@ -24,8 +24,11 @@
     GRANT SELECT ON report.group_map TO "{{fdw_user}}";
     GRANT SELECT ON report.group_bubbled_activity TO "{{fdw_user}}";
     GRANT SELECT ON report.group_bubbled_counts TO "{{fdw_user}}";
+    GRANT SELECT ON report.group_type_counts TO "{{fdw_user}}";
+    GRANT SELECT ON report.group_bubbled_type_counts TO "{{fdw_user}}";
     GRANT SELECT ON report.group_roles TO "{{fdw_user}}";
     GRANT SELECT ON report.organization_activity TO "{{fdw_user}}";
+    GRANT SELECT ON report.organization_annotation_types TO "{{fdw_user}}";
     GRANT SELECT ON report.organization_assignments TO "{{fdw_user}}";
     GRANT SELECT ON report.organization_roles TO "{{fdw_user}}";
     GRANT SELECT ON report.users TO "{{fdw_user}}";

lms/data_tasks/report/refresh/03_activity_counts_refresh.sql

@@ -7,8 +7,17 @@ ANALYSE report.group_bubbled_activity;
 REFRESH MATERIALIZED VIEW CONCURRENTLY report.group_bubbled_counts;
 ANALYSE report.group_bubbled_counts;
 
+REFRESH MATERIALIZED VIEW CONCURRENTLY report.group_type_counts;
+ANALYSE report.group_type_counts;
+
+REFRESH MATERIALIZED VIEW CONCURRENTLY report.group_bubbled_type_counts;
+ANALYSE report.group_bubbled_type_counts;
+
 REFRESH MATERIALIZED VIEW CONCURRENTLY report.user_activity;
-ANALYSE report.organization_activity;
+ANALYSE report.user_activity;
 
 REFRESH MATERIALIZED VIEW CONCURRENTLY report.organization_activity;
 ANALYSE report.organization_activity;
+
+REFRESH MATERIALIZED VIEW CONCURRENTLY report.organization_annotation_types;
+ANALYSE report.organization_annotation_types;