Register frozen Parquet files for DuckLake perf

[bill-ph]

Summary

• replace the memory/disk-heavy INSERT … SELECT backfill with ducklake_add_data_files registration of the frozen events and persons Parquet objects
• create the DuckLake tables before registration, then apply the existing production partition specifications; the fixture keys are flat rather than Hive-partitioned
• validate exact file-list registration and schema metadata without scanning the corpus before the perf step

Why

The first real mw-dev run failed during setup_posthog_tables: rewriting the production-sized fixture exhausted worker ephemeral storage. Registration reads Parquet footers and writes DuckLake catalog metadata only, so raw-view and DuckLake-table perf queries use the same immutable S3 objects.

Validation

• go test ./tests/mw-dev/scenario ./tests/mw-dev/scenario/sql
• local smoke test with the pinned DuckLake extension: register a Parquet file, apply partitioning afterward, query it successfully, and confirm no replacement Parquet files were written
• git diff --check

Next: dispatch posthog_frozen_perf from this branch to validate mw-dev setup and performance.

[github-actions]

Test Impact Plan

Deterministic summary of how this PR changes tests, CI runners, and coverage-risk signals.

Summary

Area	Added	Changed	Deleted
Test files	0	4	0
E2E/journey files	0	0	0
Workflow files	0	0	0

Signals

• Test cases: +1 / -1
• Assertions: +4 / -3
• Skips or known failures added: 0
• Workflow continue-on-error added: 0
• Workflow path filters added: 0
• Test commands removed from justfile: 0
• E2E/journey retry lines added: 0

Coverage risk: neutral or increased

No coverage-reduction warnings detected.