Expand MongoDB analytics operator parity

[midagedev]

Summary

• add MongoDB-compatible $min/$max update semantics across update, upsert, find-and-modify, and dotted paths
• support expression-based update pipelines with MongoDB's update stage allowlist
• add analytics expressions and accumulators: $dateTrunc, percentile/median, set/object merge, and N-family accumulators
• add a deterministic $setWindowFields subset with shift and ranking operators
• centralize BSON value ordering, expand regression coverage, and prepare the 0.1.10 release line

Why

The in-memory server rejected commands that production MongoDB accepts, blocking Spring Data integration tests for analytics rollups and forcing application-side workarounds for time bucketing, percentiles, top-N, and event-sequence queries.

Impact

Spring Data MongoDB tests can use the same update and aggregation shapes as production for the documented subset. Unsupported timezone and advanced window forms continue to fail explicitly.

Verification

• gradle test — complete project suite passed
• update/upsert command E2E coverage added
• aggregation expression, accumulator, and window coverage added
• git diff --check
• consumer composite-build smoke compiled jongodb and platform-server sources; the requested test filter is absent in the current consumer checkout, so Gradle reported no matching tests

Closes #480