merge dev to main #205
merge dev to main #205
Conversation
* fix: misc code review issues * fix lint
* refactor: re-implement name mapper * update * update
WalkthroughThe PR bumps versions across many packages and samples, adds enum aggregation support via a new type parameter in CRUD types, adjusts groupBy validation for empty non-relational fields, refactors query name-mapping to a scope-based approach, and overhauls/expands tests with a new schema-driven harness and generated test schema files. Changes
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
Poem
Tip 🔌 Remote MCP (Model Context Protocol) integration is now available!Pro plan users can now connect to remote MCP servers from the Integrations page. Connect with popular remote MCPs such as Notion and Linear to add more context to your reviews and chats. ✨ Finishing Touches
🧪 Generate unit tests
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. CodeRabbit Commands (Invoked using PR/Issue comments)Type Other keywords and placeholders
Status, Documentation and Community
|
|
Claude finished @ymc9's task —— View job Code Review for PR #205: merge dev to mainReviewing changes for:
Key Changes Identified:
Review SummaryThis PR contains a substantial refactoring of the name mapping functionality, which appears to be a critical component for handling database schema to model name translations. The changes look well-implemented with comprehensive test coverage. ✅ StrengthsCode Quality & Architecture:
Name Mapping Implementation (
Test Coverage (
🔧 Technical AnalysisPerformance Considerations:
Security:
Robustness:
📝 Minor Observations
🧪 Test QualityThe test suite is comprehensive and covers:
🚀 RecommendationAPPROVE - This is a high-quality refactoring that significantly improves the name mapping functionality. The code is well-architected, thoroughly tested, and handles complex SQL transformation scenarios properly. The implementation follows best practices and maintains type safety throughout. The refactoring appears to address limitations of the previous implementation while maintaining backward compatibility and adding robust support for complex query scenarios. |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (2)
packages/runtime/src/client/crud/validator.ts (1)
1306-1318: Bug: orderBy refinement assumes object but schema allows array; valid inputs may be rejected.orderBy is defined as union of object | object[]. The refinement treats it as object and calls Object.keys directly. If an array is provided, keys like "0"/"1" are checked against bys, causing false negatives.
Apply this refactor to normalize orderBy to an array before validation:
- schema = schema.refine((value) => { - const bys = typeof value.by === 'string' ? [value.by] : value.by; - if ( - value.orderBy && - Object.keys(value.orderBy) - .filter((f) => !AGGREGATE_OPERATORS.includes(f as AGGREGATE_OPERATORS)) - .some((key) => !bys.includes(key)) - ) { - return false; - } else { - return true; - } - }, 'fields in "orderBy" must be in "by"'); + schema = schema.refine((value) => { + const bys = typeof value.by === 'string' ? [value.by] : value.by; + const orderByEntries = value.orderBy + ? (Array.isArray(value.orderBy) ? value.orderBy : [value.orderBy]) + : []; + const violates = orderByEntries.some((o: Record<string, unknown>) => + Object.keys(o) + .filter((f) => !AGGREGATE_OPERATORS.includes(f as AGGREGATE_OPERATORS)) + .some((key) => !bys.includes(key)), + ); + return !violates; + }, 'fields in "orderBy" must be in "by"');packages/runtime/src/client/crud-types.ts (1)
205-243: Propagate WithAggregations through boolean combinators (AND/OR/NOT).
WhereInputgainedWithAggregations, butAND,OR, andNOTstill instantiateWhereInputwithout forwarding it. This disables aggregated predicates under combinators (e.g., inhaving).Apply:
} & { $expr?: (eb: ExpressionBuilder<ToKyselySchema<Schema>, Model>) => OperandExpression<SqlBool>; } & { - AND?: OrArray<WhereInput<Schema, Model, ScalarOnly>>; - OR?: WhereInput<Schema, Model, ScalarOnly>[]; - NOT?: OrArray<WhereInput<Schema, Model, ScalarOnly>>; + AND?: OrArray<WhereInput<Schema, Model, ScalarOnly, WithAggregations>>; + OR?: WhereInput<Schema, Model, ScalarOnly, WithAggregations>[]; + NOT?: OrArray<WhereInput<Schema, Model, ScalarOnly, WithAggregations>>; };
🧹 Nitpick comments (16)
packages/dialects/sql.js/package.json (1)
3-3: Version bump acknowledged; consider explicit publishConfig to avoid accidental restricted publish.Scoped packages default to "restricted" access on npm. If this package is intended to be public, add publishConfig to avoid publish-time surprises.
You can add:
"license": "MIT", + "publishConfig": { "access": "public" }, "files": [packages/zod/package.json (1)
31-38: Align zod devDependencies and peerDependencies across the monorepoQuick check confirms that both packages/zod and packages/runtime pin
zodto~3.25.0under devDependencies but usecatalog:under peerDependencies, which can lead to dual installs if the catalog alias shifts versions.Please align these in both packages—either:
- Use
catalog:for both dev and peer, so you always reference the workspace zod package- Or pin both to the same semver range (
~3.25.0) to guarantee only one version is installedLocations to update:
- packages/zod/package.json (lines 34–38)
- packages/runtime/package.json (lines 82–103)
packages/runtime/src/client/crud/validator.ts (1)
1259-1263: Good hardening: avoid constructing z.enum([]) for models without non-relational fields.Conditionally computing bySchema prevents runtime errors when there are no scalar fields to group by. One small UX improvement: z.never() yields a generic error. Consider emitting a clearer message so users know why groupBy is invalid for the model.
Here’s a minimally invasive tweak that keeps the same behavior but improves the error:
- const bySchema = - nonRelationFields.length > 0 - ? this.orArray(z.enum(nonRelationFields as [string, ...string[]]), true) - : z.never(); + const bySchema = + nonRelationFields.length > 0 + ? this.orArray(z.enum(nonRelationFields as [string, ...string[]]), true) + : z.any().refine(() => false, { + message: `Model "${model}" has no non-relational scalar fields to group by`, + });Also applies to: 1267-1267
packages/runtime/test/utils.ts (1)
136-137: Invariant on workDir is a solid guard; consider gating the console log behind a debug flag.The assert eliminates downstream non-null assertions. The unconditional console.log can be noisy in CI.
- console.log(`Work directory: ${workDir}`); + if (process.env.ZENSTACK_TEST_DEBUG) { + console.log(`Work directory: ${workDir}`); + }packages/runtime/src/client/executor/name-mapper.ts (6)
88-94: Future-proof ReturningNode handling: preserve all properties from super.Current override rebuilds
ReturningNodewith onlykindandselections. If Kysely adds fields later, they’d be dropped. Safer to spread the transformed base node and only replaceselections.Apply:
- protected override transformReturning(node: ReturningNode) { - return { - kind: node.kind, - // map column names in returning selections (include returningAll) - selections: this.processSelections(node.selections), - }; - } + protected override transformReturning(node: ReturningNode) { + const base = super.transformReturning(node) as ReturningNode; + return { + ...base, + // map column names in returning selections (include returningAll) + selections: this.processSelections(node.selections), + }; + }
77-86: Be explicit about scope flags for insert.Minor: the scope created for
transformInsertQueryleavesnamesMappedundefined. Make it explicit (false) for readability/consistency with other call sites.- return this.withScope( - { model: node.into.table.identifier.name }, + return this.withScope( + { model: node.into.table.identifier.name, namesMapped: false }, () => ({ ...super.transformInsertQuery(node), // map table name into: this.processTableRef(node.into!), }) satisfies InsertQueryNode, );
53-69: Ensurefromis transformed exactly once.You run
super.transformSelectQuery(node)and then overwrite.fromwith the result ofprocessFrom(node.from!), which itself callssuper.transformFrom. This is fine, but it means thefromsubtree is not the one returned bysuper.transformSelectQuery. Consider capturing the base result and only replacing itsfromto keep a single transformation pass over children.- return this.withScopes(scopes, () => { - return { - ...super.transformSelectQuery(node), - // convert "from" to nested query as needed - from: this.processFrom(node.from!), - }; - }); + return this.withScopes(scopes, () => { + const base = super.transformSelectQuery(node) as SelectQueryNode; + return { + ...base, + from: this.processFrom(node.from!), + }; + });
323-345: Consistency: prefersuper.transformNodeorsuper.transformFrom, notthis.transformNode.In
processFrom, the else branches mixsuper.transformNode(from)andthis.transformNode(from). Usingsuperavoids re-entering your overrides unexpectedly and keeps the intent clear.- return super.transformNode(from); + return super.transformNode(from); } - return this.transformNode(from); + return super.transformNode(from);Note: the first occurrence is already
super; the second should match it.
303-306: Performance nit:hasMappedColumnsscans the entire map each time.For hot paths, consider building a
Set<string>of models with mapped fields during construction to make this O(1). Not critical, just a micro-optimization.
400-416: Correlation: good handling, but add a test for nested scopes with aliases.
processSelectAlltakes the top scope; this works under yourwithScope(s)usage. To guard against regressions, add a test where a correlated subquery uses an inner alias different from the outer model name (e.g.,selectFrom('User as u')inside aselectFrom('Post')) and verifieswhereRef('u.id','=','Post.authorId').I can draft an additional vitest case targeting this scenario if you want it included in this PR.
packages/runtime/test/client-api/name-mapping.test.ts (3)
10-16: Test setup is clean; consider shaving runtime if CI is tight.Using
{ usePrismaPush: true }rebuilds the DB per test case viabeforeEach, which is thorough but slower. If acceptable, moving tobeforeAll/afterAllor relying on$pushSchema()for SQLite can make the suite faster while keeping isolation.
146-154: Great correlated subquery assertion; add an aliasing variant.You already test a correlated subquery. Add a variant with
selectFrom('User as u')andwhereRef('u.id','=','Post.authorId')to lock in alias behavior.
217-223: Potentially brittle assumption on matching IDs.
deleteFrom('Post') ... returning(['id', 'title'])expects{ id: user.id }. This holds because both are1in a fresh DB, but it ties two independent sequences together. Consider asserting onlytitleor capturing the created post ID instead.Apply:
- ).resolves.toMatchObject({ - id: user.id, - title: 'Post1', - }); + ).resolves.toMatchObject({ + title: 'Post1', + });packages/runtime/test/schemas/name-mapping/input.ts (3)
8-10: Minor: consolidate type-only imports from the same moduleYou can reduce duplication by merging the two type-only imports from "@zenstackhq/runtime" into one. Not critical since this is generated, but it slightly improves readability.
Apply this diff locally if you plan to tweak the generator:
-import type { FindManyArgs as $FindManyArgs, FindUniqueArgs as $FindUniqueArgs, FindFirstArgs as $FindFirstArgs, CreateArgs as $CreateArgs, CreateManyArgs as $CreateManyArgs, CreateManyAndReturnArgs as $CreateManyAndReturnArgs, UpdateArgs as $UpdateArgs, UpdateManyArgs as $UpdateManyArgs, UpdateManyAndReturnArgs as $UpdateManyAndReturnArgs, UpsertArgs as $UpsertArgs, DeleteArgs as $DeleteArgs, DeleteManyArgs as $DeleteManyArgs, CountArgs as $CountArgs, AggregateArgs as $AggregateArgs, GroupByArgs as $GroupByArgs, WhereInput as $WhereInput, SelectInput as $SelectInput, IncludeInput as $IncludeInput, OmitInput as $OmitInput } from "@zenstackhq/runtime"; -import type { SimplifiedModelResult as $SimplifiedModelResult, SelectIncludeOmit as $SelectIncludeOmit } from "@zenstackhq/runtime"; +import type { + FindManyArgs as $FindManyArgs, + FindUniqueArgs as $FindUniqueArgs, + FindFirstArgs as $FindFirstArgs, + CreateArgs as $CreateArgs, + CreateManyArgs as $CreateManyArgs, + CreateManyAndReturnArgs as $CreateManyAndReturnArgs, + UpdateArgs as $UpdateArgs, + UpdateManyArgs as $UpdateManyArgs, + UpdateManyAndReturnArgs as $UpdateManyAndReturnArgs, + UpsertArgs as $UpsertArgs, + DeleteArgs as $DeleteArgs, + DeleteManyArgs as $DeleteManyArgs, + CountArgs as $CountArgs, + AggregateArgs as $AggregateArgs, + GroupByArgs as $GroupByArgs, + WhereInput as $WhereInput, + SelectInput as $SelectInput, + IncludeInput as $IncludeInput, + OmitInput as $OmitInput, + SimplifiedModelResult as $SimplifiedModelResult, + SelectIncludeOmit as $SelectIncludeOmit, +} from "@zenstackhq/runtime";
30-30: Optional: consider providing a convenience default for GetPayloadIf consumers frequently use “default payload” without select/include/omit, adding a helper alias (e.g., UserGetDefaultPayload) could reduce boilerplate in tests. Only do this if it matches runtime semantics for “no SIO args”.
For example (placed after current exports):
// Helper for default payload (no select/include/omit) // Ensure "{}" matches your SimplifiedModelResult semantics before adopting. export type UserGetDefaultPayload = $SimplifiedModelResult<$Schema, "User", {}>;
50-50: Same optional helper consideration for Post GetPayloadMirroring any User default payload helper for Post keeps the surface consistent.
Example:
export type PostGetDefaultPayload = $SimplifiedModelResult<$Schema, "Post", {}>;
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
💡 Knowledge Base configuration:
- MCP integration is disabled by default for public repositories
- Jira integration is disabled by default for public repositories
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (26)
package.json(1 hunks)packages/cli/package.json(1 hunks)packages/common-helpers/package.json(1 hunks)packages/create-zenstack/package.json(1 hunks)packages/dialects/sql.js/package.json(1 hunks)packages/eslint-config/package.json(1 hunks)packages/ide/vscode/package.json(1 hunks)packages/language/package.json(1 hunks)packages/runtime/package.json(1 hunks)packages/runtime/src/client/crud-types.ts(2 hunks)packages/runtime/src/client/crud/validator.ts(1 hunks)packages/runtime/src/client/executor/name-mapper.ts(4 hunks)packages/runtime/test/client-api/name-mapping.test.ts(1 hunks)packages/runtime/test/schemas/name-mapping/input.ts(1 hunks)packages/runtime/test/schemas/name-mapping/models.ts(1 hunks)packages/runtime/test/schemas/name-mapping/schema.ts(1 hunks)packages/runtime/test/schemas/name-mapping/schema.zmodel(1 hunks)packages/runtime/test/utils.ts(2 hunks)packages/sdk/package.json(1 hunks)packages/tanstack-query/package.json(1 hunks)packages/testtools/package.json(1 hunks)packages/typescript-config/package.json(1 hunks)packages/vitest-config/package.json(1 hunks)packages/zod/package.json(1 hunks)samples/blog/package.json(1 hunks)tests/e2e/package.json(1 hunks)
🧰 Additional context used
📓 Path-based instructions (3)
{packages,samples,tests}/**
📄 CodeRabbit inference engine (CLAUDE.md)
Packages are located in
packages/,samples/, andtests/
Files:
packages/common-helpers/package.jsontests/e2e/package.jsonpackages/cli/package.jsonpackages/testtools/package.jsonsamples/blog/package.jsonpackages/vitest-config/package.jsonpackages/dialects/sql.js/package.jsonpackages/zod/package.jsonpackages/runtime/package.jsonpackages/typescript-config/package.jsonpackages/tanstack-query/package.jsonpackages/sdk/package.jsonpackages/create-zenstack/package.jsonpackages/runtime/test/schemas/name-mapping/models.tspackages/eslint-config/package.jsonpackages/language/package.jsonpackages/ide/vscode/package.jsonpackages/runtime/src/client/crud/validator.tspackages/runtime/test/schemas/name-mapping/schema.tspackages/runtime/test/utils.tspackages/runtime/test/schemas/name-mapping/schema.zmodelpackages/runtime/test/schemas/name-mapping/input.tspackages/runtime/test/client-api/name-mapping.test.tspackages/runtime/src/client/crud-types.tspackages/runtime/src/client/executor/name-mapper.ts
tests/e2e/**
📄 CodeRabbit inference engine (CLAUDE.md)
E2E tests are in
tests/e2e/directory
Files:
tests/e2e/package.json
**/schema.zmodel
📄 CodeRabbit inference engine (CLAUDE.md)
Always run
zenstack generateafter modifying ZModel schemas
Files:
packages/runtime/test/schemas/name-mapping/schema.zmodel
🧠 Learnings (3)
📚 Learning: 2025-08-04T08:43:33.161Z
Learnt from: CR
PR: zenstackhq/zenstack-v3#0
File: CLAUDE.md:0-0
Timestamp: 2025-08-04T08:43:33.161Z
Learning: `zenstack generate` compiles ZModel to TypeScript schema (`schema.ts`)
Applied to files:
packages/runtime/test/schemas/name-mapping/models.tspackages/language/package.jsonpackages/runtime/test/schemas/name-mapping/schema.ts
📚 Learning: 2025-08-04T08:43:33.161Z
Learnt from: CR
PR: zenstackhq/zenstack-v3#0
File: CLAUDE.md:0-0
Timestamp: 2025-08-04T08:43:33.161Z
Learning: ZModel schema (`schema.zmodel`) defines database structure and policies
Applied to files:
packages/runtime/test/schemas/name-mapping/schema.zmodel
📚 Learning: 2025-08-04T08:43:33.161Z
Learnt from: CR
PR: zenstackhq/zenstack-v3#0
File: CLAUDE.md:0-0
Timestamp: 2025-08-04T08:43:33.161Z
Learning: Schema-first approach with ZModel DSL extension of Prisma schema language
Applied to files:
packages/runtime/test/schemas/name-mapping/schema.zmodel
🧬 Code graph analysis (6)
packages/runtime/src/client/crud/validator.ts (2)
samples/blog/zenstack/schema.ts (1)
schema(9-230)packages/runtime/test/schemas/basic/schema.ts (1)
schema(9-284)
packages/runtime/test/schemas/name-mapping/schema.ts (2)
packages/runtime/src/schema/expression.ts (1)
ExpressionUtils(16-112)packages/sdk/src/schema/schema.ts (1)
SchemaDef(10-18)
packages/runtime/test/utils.ts (3)
packages/common-helpers/src/tiny-invariant.ts (1)
invariant(4-14)packages/language/src/index.ts (1)
loadDocument(21-132)packages/sdk/src/prisma/prisma-schema-generator.ts (1)
PrismaSchemaGenerator(64-476)
packages/runtime/test/client-api/name-mapping.test.ts (3)
packages/runtime/src/client/contract.ts (1)
ClientContract(50-159)packages/runtime/test/schemas/name-mapping/schema.ts (2)
SchemaType(88-88)schema(9-87)packages/runtime/test/utils.ts (1)
createTestClient(78-194)
packages/runtime/src/client/crud-types.ts (2)
packages/sdk/src/schema/schema.ts (5)
GetModelFieldType(151-155)ModelFieldIsOptional(217-221)SchemaDef(10-18)GetEnums(123-123)GetEnum(125-125)packages/runtime/src/utils/type-utils.ts (1)
NullableIf(5-5)
packages/runtime/src/client/executor/name-mapper.ts (3)
packages/runtime/src/client/query-utils.ts (2)
getModel(16-18)requireModel(20-26)packages/common-helpers/src/tiny-invariant.ts (1)
invariant(4-14)packages/sdk/src/schema/schema.ts (1)
ModelDef(20-36)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (2)
- GitHub Check: claude-review
- GitHub Check: build-test (20.x)
🔇 Additional comments (31)
tests/e2e/package.json (1)
3-3: LGTM: e2e package version bumped consistentlyVersion updated to 3.0.0-alpha.31; matches the monorepo bump intent. No functional impact here.
packages/cli/package.json (1)
6-6: LGTM: CLI package version bumpLooks good. As a quick follow-up, ensure the published CLI reports the same version via
zen --versionin CI.You can validate across the repo with the script in the runtime package comment below.
packages/common-helpers/package.json (1)
3-3: LGTM: common-helpers version bumpNo other changes. This keeps internal workspace consumers aligned.
packages/runtime/package.json (1)
3-3: All package versions verified at 3.0.0-alpha.31Script confirms that every package.json version field matches 3.0.0-alpha.31, there are no lingering “alpha.30” references, and all internal dependencies use the workspace protocol. LGTM.
packages/sdk/package.json (1)
3-3: LGTM: SDK package version bumpAligned with the repo-wide release. Nothing else to call out.
packages/create-zenstack/package.json (1)
3-3: Version bump to 3.0.0-alpha.31 looks goodNo functional or publish-surface changes in this file; aligns with the repo-wide release bump.
packages/typescript-config/package.json (1)
3-3: Typescript config package version bumped correctlyPure metadata update; no API surface. All good.
packages/vitest-config/package.json (1)
4-4: Vitest config version bump approvedMatches the monorepo version. Nothing else changed.
package.json (1)
3-3: Version bump verified — no mismatches or stale references foundA full scan of all package.json files under packages/, samples/, and tests/ confirms every version is now set to
3.0.0-alpha.31. A repo-wide search also shows no lingering3.0.0-alpha.30references (excluding CHANGELOGs).You can safely merge.
packages/eslint-config/package.json (1)
3-3: ESLint config package version bump approvedMetadata-only change; consistent with the release train.
packages/testtools/package.json (1)
3-3: I’ve added a script to list all CI workflow files and search for any “testtools” references. This will help confirm whether our CI matrix explicitly includes the correct peer dependency versions (better-sqlite3 ^12.2.0 and pg ^8.13.1). Please review the output to ensure the workflows install and test against those versions.packages/tanstack-query/package.json (1)
3-3: Version bump approved – peerDependency for @tanstack/react-query remains ^5.0.0
- Verified in packages/tanstack-query/package.json (line 35) that the peerDependency
"@tanstack/react-query": "^5.0.0"is unchanged.- Confirmed that packages/tanstack-query/src/react.ts simply re-exports from
@tanstack/react-query.- No local samples or tests import React Query here — please ensure any consuming examples/tests against v5 still run clean after this bump.
packages/ide/vscode/package.json (1)
4-4: tsup config correctly bundles all runtime dependencies;--no-dependenciespublish is safeVerified that
packages/ide/vscode/tsup.config.tsusesnoExternal: [/^(?!vscode$)/], external: ['vscode'],which bundles every dependency except the VS Code API into
dist. Withvsce publish --no-dependencies, the extension is self‐contained and does not requirenode_modulesat runtime—no further changes needed.samples/blog/package.json (1)
3-3: Version bump consistency confirmedAll
package.jsonfiles across the workspace (includingsamples/blog/package.json) have been updated to3.0.0-alpha.31, and thepnpm-lock.yamlcontains no lingeringalpha.30references. Lockfile and workspace are fully in sync—no further action required.packages/zod/package.json (1)
3-3: LGTM on the package version bump.The package version is correctly bumped to 3.0.0-alpha.31 with no other surface changes.
packages/runtime/test/utils.ts (2)
145-153: Non-null assertions on workDir are safe after invariant; flow is correct.Using workDir! for loadDocument and file writes is justified post-invariant. Also good to enforce schema source presence when using prisma db push.
176-179: Sqlite file path uses workDir safely; handles relative dbName.Joining workDir with dbName (default ./test.db) keeps test DBs contained per project.
packages/runtime/test/schemas/name-mapping/schema.zmodel (3)
6-11: User model mapping setup looks correct for name-mapping tests.
- Field-level @Map on email and model-level @@Map to 'users' are consistent for exercising mapper behavior.
- Unique on email ensures a stable lookup surface.
13-19: Post model mapping and relation look consistent; good coverage of both field and model mapping.
- title and authorId mappings plus @@Map('posts') should adequately test scoped name mapping.
- Relation via authorId -> User.id is correctly declared.
1-4: Generated artifacts verified – schema.ts, models.ts, and input.ts are all present alongside schema.zmodel in packages/runtime/test/schemas/name-mapping. No further action needed.packages/runtime/test/schemas/name-mapping/schema.ts (2)
14-45: LGTM: mapping metadata mirrors the ZModel.
@map/@@mapattributes and FK metadata are correctly represented for User.email →user_email, Post.title →post_title, Post.authorId →author_id, and model tablesusers/posts. Should interop cleanly with the new name-mapper.
8-9: Ensuredist/schemaimport is available in CI testsThe generated test at
packages/runtime/test/schemas/name-mapping/schema.tscurrently does:import { type SchemaDef, ExpressionUtils } from "../../../dist/schema";However,
packages/runtime/distisn’t committed and only appears after a build step. If the CI test runner invokes Jest/TS without first compiling, this import will fail.• Confirm that your CI workflow for
packages/runtimeruns the build (e.g.npm run buildortsc) before executing tests—check yourpackage.jsonscripts and.github/workflows/*or equivalent.
• If the build-then-test ordering isn’t in place, update the CI configuration to include the build step prior to testing.
• Alternatively, adjust the code-generator to emitimport … from "../../src/schema"(or use a test-only flag) so it resolves directly against the source files.packages/runtime/src/client/crud-types.ts (2)
245-264: Enum aggregations: shape looks correct and consistent with primitives.Forwarding
WithAggregationsand adding_count/_min/_maxaligns enum filters with other scalar filters. Nice.
1084-1088: Good call exposing aggregations in GroupByHaving via WhereInput<..., true, true>.This unlocks aggregated enum predicates in
having. Once theAND/OR/NOTfix lands, the entire having tree will correctly accept aggregations.packages/runtime/test/schemas/name-mapping/models.ts (1)
8-11: LGTM: concise, generator-friendly type exports.Types
UserandPostcorrectly aliasModelResultover the generatedSchemaType. No runtime code, safe to keep generated.packages/runtime/test/client-api/name-mapping.test.ts (2)
119-129: Nice coverage of unqualified vs qualified refs and self-refs.The trio of queries (
selectAll+ unqualified where, qualifiedUser.email, andwhereRef('email','=','email')) exercises the name-mapper across common select patterns.
130-140: Join mapping through nested select-all is validated well.
innerJoin('User', 'User.id', 'Post.authorId')plus selecting from both sides ensures mapped table/columns are correctly surfaced via aliases.packages/runtime/test/schemas/name-mapping/input.ts (4)
1-7: Generated file header and lint disable look goodHeader clearly marks the file as generated and disables linting appropriately for emitted types.
11-29: User model argument/type exports look consistent and comprehensiveAliases cover the expected CRUD, aggregation, groupBy, and shape helpers. Naming aligns with the rest of the runtime conventions.
31-49: Post model argument/type exports mirror User correctlySymmetric coverage for Post maintains consistency and reduces surprises in the test harness.
8-50: Runtime type exports verified
All referenced types are correctly exported frompackages/runtime/src/client/crud-types.ts, so no further changes are needed.• SelectIncludeOmit (line 453)
• SimplifiedModelResult (line 181)
• FindManyArgs (line 664)
• FindUniqueArgs (line 667)
• AggregateArgs (line 1002)
• GroupByArgs (line 1089)
• WhereInput (line 207)
• SelectInput (line 459)
• IncludeInput (line 487)
• OmitInput (line 449)
Summary by CodeRabbit