Config for purescript-graphql-schema-gen

Hasura roles

To generate schemas for each of your Hasura roles, create a .yaml file and add a ROLES_YAML env var. The yaml should be made up of an array of strings:

- Admin
- User
- Submitter

Outside types

To replace default hasura types with Purescript types, create one or more .yaml files and add them - comma separated - to the OUTSIDE_TYPES_YAML env var.

outside_types:

This section is where you define the schema objects that you want to replace values for.

For example if you wanted to change the types on the user object, you could add the following:

outside_types:
  user:
    with: common
    id: id=UserId
    email: EmailAddress, Data.EmailAddress, emails

This would replace the 'id' field with the 'UserId' type (see 'types' section below). It would also replace the 'email' field with an inline type of the format: [type name], [module name], [package name].

The with key is an optional special key that allows you to define a common set of types to use across multiple objects. For example, if you wanted to use a 'common' set of types on both the 'user' and 'post' objects, you would define 'common' types in the 'templates' section and then add a with: common key to the 'user' and 'post' objects.

types:

This section is where you define shorthand type templates modules containing multiple types. An example for a couple of Id modules is:

types:
  id: $, Data.Id.$, oa-ids
  drId: $, Data.Id.DelegateRegistration.$, oa-ids

The '$' symbol is a placeholder for the type name that the type is called with in templates or outside_types. For example in the 'outside_types' section above, we use id=UserId which is translated into UserId, Data.Id.UserId, oa-ids.

templates:

This section is where you define the common types that you want to use across multiple objects via the with key. For example:

templates
  common:
    author_id: id=AuthorId
    stage_id: id=StageId
    user_id: id=UserId
    created_by: id=UserId
    client_id: id=ClientId
    event_id: id=EventId
  event_types:
    event_title: HTML, Data.HTML
    frequency: Frequency, Data.Frequency
    created_by: id=UserId
    client_id: id=ClientId
    event_id: id=EventId

The keys don't have to exist on the object you call the template for, but any keys that do match will be replaced with the template value.

Type name casing

By default GraphQL type names only get their first letter uppercased to make them valid PureScript type names (users_insert_input -> Users_insert_input). Set pascal_case_types: true to fully PascalCase them instead (users_insert_input -> UsersInsertInput).

pascal_case_types: true

Note that the SHARED_ENUM_SUFFIXES env var is matched against the converted type name, so its entries need to use the same casing (e.g. Enum with pascal_case_types: true, _enum without).

Skipping unused schema plumbing

skip_types takes a list of regexes matched against GraphQL type names: matching types are not generated, matching field return types drop the field, and matching argument types drop the argument. skip_keys does the same for field names. Prefix a pattern with ! to negate it; the last matching rule wins. Types left unreachable from the schema roots after skipping are pruned automatically, along with imports nothing uses.

skip_types:
  - "_stddev"
  - "_var_pop"
  - "_var_samp"
  - "_variance"
  - "_stream_cursor"
skip_keys:
  - "_stream$"

Regeneration is also incremental: files are only written when their content changes (unchanged files keep their mtimes, so purs skips them cheaply), and files from previous runs that are no longer generated are swept afterwards.