Commit f70c17f
Address PR #371 AI review R2: scenario 15 metadata accuracy
P3 fix: scenario 15's `params.n_groups` was 80 (the switcher cohort
allocator input fed into `gen_reversible(n_groups=N_GOLDEN, ...)`)
while the realized panel actually contains 120 groups (80 switchers +
20 never-treated + 20 always-treated, appended by `gen_reversible`'s
default cohort additions at line 64).
Replace with two explicit fields:
- `n_switcher_groups = 80`: the load-bearing DGP allocator input
- `n_realized_groups = 120`: the actual unique-group count in the
serialized data
The parity test reads the `data` block directly, not `params`, so it
is unaffected by this metadata change. Resolves the misleading-metadata
finding the reviewer flagged on R2.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>1 parent e405aaf commit f70c17f
2 files changed
Lines changed: 13 additions & 2 deletions
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
683 | 683 | | |
684 | 684 | | |
685 | 685 | | |
686 | | - | |
| 686 | + | |
| 687 | + | |
| 688 | + | |
| 689 | + | |
| 690 | + | |
| 691 | + | |
| 692 | + | |
| 693 | + | |
| 694 | + | |
| 695 | + | |
| 696 | + | |
687 | 697 | | |
688 | 698 | | |
689 | 699 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
766 | 766 | | |
767 | 767 | | |
768 | 768 | | |
769 | | - | |
| 769 | + | |
| 770 | + | |
770 | 771 | | |
771 | 772 | | |
772 | 773 | | |
| |||
0 commit comments