Avoid pulling pre-2023 data into movedBoxes stat

[pylipp]

https://trello.com/c/EFSdLs3F

This didn't affect data display because in the FE we filter by a date range (earliest 2023-01-01).
It reduces the querying time and request size (and hence latency) for bases with a considerable amount of pre-2023 boxes.

[pylipp]

https://github.com/copilot/c/3f260545-dd88-47cb-89f0-7c910bf30847

[Copilot]

Pull request overview

This PR adjusts the backend SQL used for the movedBoxes statistics so that, in production, all history rows considered for box reconstruction (including “Record created/deleted/undeleted” events) are constrained by the configured “earliest 2023” history.id cutoff, preventing pre-2023 history events from being pulled into the dataset.

Changes:

• Tightened the BoxHistory CTE filter so h.id >= %s applies uniformly (including to “Record created/deleted/undeleted” change types).
• Simplified the boolean logic by grouping the change-type checks under a single h.id >= %s guard.

[sentry]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.83%. Comparing base (1003b9a) to head (db2f4ce).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #2704   +/-   ##
=======================================
  Coverage   76.83%   76.83%           
=======================================
  Files         300      300           
  Lines       22233    22233           
  Branches     2245     2245           
=======================================
  Hits        17082    17082           
  Misses       5104     5104           
  Partials       47       47           

Flag	Coverage Δ
backend	99.65% <ø> (ø)
frontend	69.10% <ø> (ø)
sharedComponents	68.72% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[aerinsol]

@pylipp I assume that %s is a variable - is there a reason the computation date isn't hardcoded in? Or are you setting this in a central module where it's then called by different functions?

[pylipp]

@pylipp I assume that %s is a variable - is there a reason the computation date isn't hardcoded in? Or are you setting this in a central module where it's then called by different functions?

@aerinsol yes, the query is invoked from one place which passes the cut-off history ID. Since that's not compatible for testing, this variable is only set in production.
We could probably switch it for a date-date-threshold and hardcode it but I have to check if the history.changes column is indexed

[pylipp]

We could probably switch it for a date-date-threshold and hardcode it but I have to check if the history.changedate column is indexed

@aerinsol the changedate column is not indexed (only composite index with tablename or record_id), so I stick with the approach to filter for a variable h.id (for testing set to 0; for production, the first ID in 2023), code for reference