Can graphql-codegen-client-preset eliminate unused generated documents across apps in a monorepo?

[khuongtp]

Hi! 👋

I’m working with a Turborepo monorepo that contains:

Two separate apps (Next.js + Angular)

One shared package for reusable GraphQL operations (queries, mutations, fragments, etc.)

My issue: when I build one app, the generated gql.ts file includes all operations from both apps, even though only a subset is actually used.

For example, the generated output looks like this:

const documents: Documents = {
    "\n      query GetSomething($id: ID!) {\n        id(id: $id) {\n          ...CommonFields\n        }\n      }\n    ": types.GetSomethingDocument,
    "\n  mutation Login($email: String!, $password: String!) {\n    login(email: $email, password: $password)\n  }\n": types.LoginDocument,
};

In this case, Login is only used in the second app, but it still shows up when building the first app.

👉 My question: Is graphql-codegen-client-preset capable of tree-shaking or eliminating unused documents automatically?

[kdy1]

This is not a correct place to ask