Skip to content

Phase 2b — schema: port Graphene → Strawberry at byte-identical SDL parity #91

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
chrisbc wants to merge 1 commit into from
Closed

Phase 2b — schema: port Graphene → Strawberry at byte-identical SDL parity #91

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 18 additions & 3 deletions docs/MIGRATION_LOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -115,9 +115,10 @@ Source: code exploration 2026-06-24 on `deploy-test` @ `93f6f62`. Version **0.9.

### Phase 2 — Data layer + schema migration 🟡
- [x] `BinaryLargeObjectModel` (PynamoDB) → `pydantic` + `boto3` CRUD behind the **unchanged** `BinaryLargeObject` wrapper; `migrate()`/`drop_tables()` kept; **`pynamodb` dependency removed**
- [ ] Port ~25 graphene types across the 8 modules to Strawberry at SDL + runtime parity (relay node parity, nullability, `UNSET` args)
- [ ] Keep `cli` / `cli_ab_test` imports working
- [ ] SDL parity gate green
- [x] Port the ~25 graphene types (8 modules) to Strawberry — **SDL byte-identical**; custom `Node` interface (`id: ID!`), nullability + `UNSET`-arg traps, partial style-arg defaults all matched
- [x] `cli` / `cli_ab_test` still import (graphene path untouched)
- [x] SDL parity gate green (`tools/schema_parity.py`); corpus replay re-pointed at the Strawberry schema
- [ ] **Runtime parity for archive-dependent fields → Phase 3** (rupture field computes, pagination, sections geojson/mfd/colour, `node()` dispatch — need the tiny-archive fixtures + test-suite conversion)

### Phase 3 — Tests
- [ ] Convert suite to drive Strawberry (keep moto); parametrize legacy+strawberry where useful
Expand Down Expand Up @@ -173,4 +174,18 @@ Source: code exploration 2026-06-24 on `deploy-test` @ `93f6f62`. Version **0.9.
- Verified: the moto contract `data_store/test/test_model.py` **4/4**; full suite **77 passed / 10 skipped**; ruff + mypy clean. Container entry/Dockerfile untouched, so the Phase 1 build proof still holds.
- **Next:** Phase 2b — port the ~25 graphene types (8 modules) to Strawberry at SDL parity; re-point the parity gate + corpus replay at the new schema.

### 2026-06-24 — Phase 2b: Graphene → Strawberry schema port (SDL byte-identical)
- **`strawberry_schema.py`** now defines the full schema — 29 types — and **`tools/schema_parity.py`** confirms it is **byte-identical** to `schema.legacy.graphql` (order-insensitive normalise + diff). `app.py` already serves it.
- **Parity traps cleared** (runbook + Model feedback in action):
- root type **`QueryRoot`** (not Strawberry's `Query`); **custom `Node` interface** `id: ID!` (Model C1 — NOT `strawberry.relay`/`GlobalID`); `auto_camel_case=False` with relay `PageInfo` fields kept camelCase via `name=`.
- **nullability**: every nullable-with-default input field needs `X | None` or Strawberry emits `!` (a broad version of Model T2).
- **`UNSET` vs `null`** (Model T3): optional args with no SDL default use `strawberry.UNSET`; only `ColorScaleArgsInput.normalisation` keeps `= null` (`= None`).
- **partial style-arg default**: legacy `style = {stroke_color, stroke_width, stroke_opacity}` (3 keys, no fill) reproduced by a default input instance with `fill_*` set to `strawberry.UNSET` (so they neither render nor apply).
- `<pkg>/schema.py` name clash (Model G4) → file is `strawberry_schema.py`, rename at cutover.
- **Compute reused, not rewritten**: resolvers call the existing Graphene-free helpers (`cached.*`, `color_scale.get_colour_scale`, `apply_geojson_style`, `get_location_detail_list`). Light query-root resolvers **verified executing** vs legacy values (about, get_locations, radii as ints, colour-scale matplotlib output, location lists).
- **mypy**: added the `strawberry.ext.mypy_plugin`; new code is clean (TYPE_CHECKING aliases for the `JSONString` scalar + `SetOperationEnum`; 2 targeted `[misc]` ignores for the `id` resolver overriding the interface field — Model also accepted a scalar ignore). The 29 remaining errors are **pre-existing** in the legacy graphene modules (`follow_untyped_imports` config, deleted at cutover), not from this change, and mypy isn't the PR test gate.
- Corpus replay (14 queries) now validates against the Strawberry schema: **15 passed**. Full suite **77 passed / 10 skipped**; ruff clean.
- **Deferred** (benign): `strawberry.scalar()` class-form DeprecationWarning on `JSONString` (same one the Model pilot deferred; changing it risks the scalar's SDL name/description).
- **Next:** Phase 3 — convert the test suite to drive the Strawberry schema (moto + tiny-archive fixtures), implement the archive-dependent resolvers to runtime parity, wire CI.

<!-- Append new dated entries above this line as the migration proceeds. -->
1 change: 1 addition & 0 deletions pyproject.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -90,6 +90,7 @@ select = ["E", "F", "I", "B", "UP"] # G004 (f-string logging) deferred to a fol
quote-style = "preserve"

[tool.mypy]
plugins = ["strawberry.ext.mypy_plugin"]
ignore_missing_imports = true
exclude = "solvis_graphql_api/ab_test/client"
follow_untyped_imports = true
Loading
Loading