feat: fold-atomic-statements and flatten-history examples

[pyramation]

Summary

Rows 4–5 of the matrix in constructive-planning#1344 — the direction Dan asked about: atomic statement soup in, grouped module out.

• examples/fold-atomic-statements: input is a migration-log-style file (empty CREATE TABLE () + one ADD COLUMN/ADD CONSTRAINT per statement, FKs last); pgpm import folds each table into one fully packed CREATE TABLE (columns + PK + UNIQUE + CHECK inline, FKs as late ALTER TABLE ONLY) with graph-derived requires. 16 statements → 3 changes, 0 warnings.
• examples/flatten-history: input has real churn (column added then dropped, late SET NOT NULL); import folds the NOT NULL into the column def and preserves the add→drop pair semantically, warning that its revert isn't derivable. 14 statements → 3 changes, 1 warning — pins down exactly how much history gets flattened vs preserved.

New generic acceptance lane (3c): for each soup example, deploy the module and load the raw input fresh, assert identical catalogs; revert lane asserts clean DBs. Full suite passes locally.

Stacked on #6 (the examples/ restructure).

Link to Devin session: https://app.devin.ai/sessions/a0d6ae8694504c5bb66911f35d87a817
Requested by: @pyramation

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring