peterstockings/Stride
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Paginate activity history and aggregate dashboard totals in SQL

Browse Source
This commit is contained in:
Peter Stockings
2026-10-02 14:36:35 +10:00
parent a300ea007d
commit 5830976e31
11 changed files with 386 additions and 65 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/PERFORMANCE.md
+27
View File
@@ -63,3 +63,30 @@ HTTP 200 and `status: ok`; Dokku web and sync worker startup checks succeeded.
The synthetic browser fixture verified dashboard replay, route colour switching
and charts. Its in-memory SQLite server uses one request thread to avoid concurrent
access to the shared SQLite connection.
## Paginated history
Dashboard, shared profiles and the friends feed show 25 activities per page,
ordered by start time descending with activity ID as the tie-breaker. Navigation
preserves profile URLs and returns to the history section. Invalid or empty
out-of-range pages return 404; an empty first page remains valid.
Dashboard and profile totals are SQL aggregates over the entire authorized
history. Eight weekly buckets use independent local-time Monday boundaries
converted to UTC, including daylight-saving changes. The latest GPS run is an
independent indexed query, so paging does not change the map. Friend/profile
filters apply equally to pagination, counts, totals and route selection.
Personal record comparisons stream only ID, title, timestamp, sport, calculation
version and stored efforts in batches of 1,000. They retain at most six current
record holders and effort details for the requested page (or one run's detail).
No full-history Activity ORM objects, raw samples, charts or laps are loaded.
Comparisons remain linear in the number of recorded runs; page rows and ORM
object counts are bounded. Owners are loaded with the page to avoid per-row
athlete queries. Composite indexes support owner/date and shared owner/date lists.
A disposable local SQLite benchmark with 10,000 synthetic activities returned
25 rows and 26 ORM objects on pages 1, 200 and 400. HTML remained about 30 KiB;
measured request times were 148, 122 and 97 ms respectively. These are synthetic
local measurements, not production latency guarantees. Browser checks with 55
synthetic activities verified pages 1–3, including the five-row final page.