feat: import pg-aggregates plugin with Constructive adaptations

[pyramation]

Summary

Import pg-aggregates v5 into the Constructive monorepo with full upstream git history preserved (181 commits), adapted to Constructive conventions (pnpm/makage, single quotes, our connection-filter fork).

Disabled by default — PgAggregatesPreset is NOT included in ConstructivePreset. Users opt in by adding it to their own preset's extends array. This supports future per-database feature toggling (e.g. pro tier).

Commit structure:

1. Merge commit — git subtree add (no --squash) preserving all upstream history
2. Adaptation commit — removes upstream tooling (yarn, eslint, prettier, jest), adapts source to Constructive conventions
3. Test commit — 35 tests written from scratch using our graphile-test framework
4. Merge from main — brings in latest changes
5. Disable by default — removes PgAggregatesPreset from ConstructivePreset, updates snapshots

Key adaptation: FilterRelationalAggregatesPlugin rewritten to use $where._manyRelation (our fork) instead of upstream's $where.extensions.pgFilterRelation.

Review & Testing Checklist for Human

• Verify FilterRelationalAggregatesPlugin adaptation — this is the riskiest change. Check src/FilterRelationalAggregatesPlugin.ts lines 40-53 against your connection-filter fork's ConnectionFilterBackwardRelationsPlugin.ts to confirm the _manyRelation shape matches.
• Test relational aggregate filtering on a real schema — run a query like allTeams(where: { playersByTeamId: { aggregates: { sum: { goals: { greaterThan: "20" } } } } }) against a production-like schema after opting in.
• Verify git history preserved — run git log --follow graphile/graphile-pg-aggregates/src/index.ts and confirm all 181 upstream commits are visible.
• Schema snapshot diff — verify the snapshot deletions only removed aggregate-related types (since the preset is now disabled by default).

Notes

• To enable aggregates for a database, add PgAggregatesPreset to the preset extends: import { PgAggregatesPreset } from 'graphile-pg-aggregates'
• 35 tests cover: basic aggregates, groupBy, having, relational aggregates, schema introspection, edge cases (NULL handling, pagination), where pre-filter, multi-key groupBy, and behavior opt-out
• sum(int) returns bigint in PostgreSQL, which GraphQL represents as a string — filter values must be strings (e.g., greaterThan: "20")
• Per-table opt-out is supported via @behavior -aggregates smart tag

Link to Devin session: https://app.devin.ai/sessions/a233256611174c5fa1e38e35807cd661
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 and CI monitoring