Hi,

In v1.4.5, numerical enums convert to a union of string literals. However, Zod officially says that numerical enums should be a union of numeric literals. (ref: https://github.com/colinhacks/zod/issues/338) And an unnecessary .int() is added if the type is an integer.

openapi: 3.0.0
info:
  version: 1.0.0
  title: Numerical enums
paths:
  /sample:
    get:
      parameters:
        - in: query
          name: foo
          schema:
            type: integer
            enum:
              - 1
              - -2
              - 3
        - in: query
          name: bar
          schema:
            type: number
            enum:
              - 1.2
              - 34
              - -56.789
      responses:
        "200":
          description: resoponse

parameters: [
    {
        name: "foo",
        type: "Query",
        schema: z
            .union([z.literal("1"), z.literal("-2"), z.literal("3")])
            .int() // should remove
            .optional(),
    },
    {
        name: "bar",
        type: "Query",
        schema: z.union([z.literal("1.2"), z.literal("34"), z.literal("-56.789")]).optional(),
    },
],

I expected as following:

parameters: [
    {
        name: "foo",
        type: "Query",
        schema: z.union([z.literal(1), z.literal(-2), z.literal(3)]).optional(),
    },
    {
        name: "bar",
        type: "Query",
        schema: z.union([z.literal(1.2), z.literal(34), z.literal(-56.789)]).optional(),
    },
],

Hello Alex,

With this pull request i'm adding support for default errors. indeed zodios also has default errors.

This also seems to fix a bug where only other than '200' ok code would not emit response type.

now the /samples/3.0/petstore.yaml convertion will output correct zodios api definition.

I added a handlebar helper to handle this use case. tell me if it's ok for you.

Is there a way to generate all OpenApi component schemas and export them (without hashed names)?

In the Pet Store example, there are 8 component schemas. https://github.com/astahmer/openapi-zod-client/blob/main/example/petstore.yaml

Only 4 of them are added to export const schemas {...}. https://github.com/astahmer/openapi-zod-client/blob/main/example/petstore-schemas.ts

Huhu,

getting the following error on openapi-zod-client <file> -a --export-schemas --group-strategy 'tag'. (any group-strategy seems to fail)

if (schemaName.startsWith("z.")) return;
                             ^

TypeError: Cannot read properties of undefined (reading 'startsWith')
    at addDependencyIfNeeded (node_modules/openapi-zod-client/dist/generateZodClientFromOpenAPI-8e9ca210.cjs.dev.js:1967:24)
    at node_modules/openapi-zod-client/dist/generateZodClientFromOpenAPI-8e9ca210.cjs.dev.js:1972:16
    at Array.forEach (<anonymous>)

for the same payload as here

When i run: pnpx openapi-zod-client 'https://petstore.swagger.io/v2/swagger.json' -o './src/schemas/test.ts'

It gives { method: 'get', path: '/pet/:petId', description: Returns a single pet, requestFormat: 'json', parameters: [ { name: 'petId', type: 'Path', schema: z.unknown(), }, ], response: z.void(), errors: [ { status: 400, description: Invalid ID supplied, schema: z.void(), }, { status: 404, description: Pet not found, schema: z.void(), }, ], },

Why is response: z.void() ?

related to #49 ref https://github.com/astahmer/openapi-zod-client/issues/49#issuecomment-1345370089

Targets characters are as follows:

• TAB, LF and CR
  • \t\n\r
• C0 Control Codes except for TAB, LF, CR, and Space
  • \x00-\x08\x0B\x0C\x0E-\x1F\x7F
• C1 Control Codes
  • \x80-\x9F
• Non-character
  • \uFFFE\uFFFF

Hi,

just tried to used it on a existing service and got the error

Error: Invalid ref: #/components/schemas/NameResolutionType2

{
  ref: '#/components/schemas/NameResolutionType2',
  fallbackName: 'nameResolutionType',
  result: 'NameResolutionType2.optional().default("AssetName")'
}

but don't get it, what is wrong.

minimal reproduction example:

{
  "x-generator": "NSwag v13.17.0.0 (NJsonSchema v10.8.0.0 (Newtonsoft.Json v13.0.0.0))",
  "openapi": "3.0.0",
  "info": {
    "title": "title",
    "version": "1.5.0.0"
  },
  "paths": {
    "/api/v2/name-resolution/via-id/{assetId}": {
      "get": {
        "tags": [
          "NameResolutionApi"
        ],
        "operationId": "api_v2_name-resolution_via-id_{assetid}_get_ResolveNamesForAssetV2",
        "parameters": [
          {
            "name": "nameResolutionType",
            "in": "query",
            "schema": {
              "default": "AssetName",
              "oneOf": [
                {
                  "$ref": "#/components/schemas/NameResolutionType2"
                }
              ]
            }
          }
        ],
        "responses": {
          "200": {
            "description": "",
            "content": {
              "application/json": {
                "schema": {
                  "type": "array",
                  "items": {
                    "type": "string"
                  }
                }
              }
            }
          },
          "400": {
            "description": ""
          },
          "401": {
            "description": "Unauthorized"
          },
          "403": {
            "description": "Forbidden"
          },
          "404": {
            "description": ""
          },
          "500": {
            "description": ""
          }
        }
      }
    }
  },
  "components": {
    "schemas": {
      "NameResolutionType2": {
        "type": "string",
        "description": "",
        "x-enumNames": [
          "AssetName",
          "FullName"
        ],
        "enum": [
          "AssetName",
          "FullName"
        ]
      }
    }
  }
}

Hello alex,

I'm ready to help you. First thing, this PR is just to add github actions to automate your daily process :

• tests
• build
• publish

Note that publishing requires you to add a github secret named NPM_TOKEN else it wont work. But it's really helpfull to automate this process. You will gain a lot of time by doing this.

Automatic publish work by just adding a new tag. so just doing :

npm version <major|minor|patch>
git push --follow-tags

will create a new release automatically if the CI succeed.

Another thing, i propose to add (not this PR) is register a renovate bot. this will do all the maintainance for you. A lot of time gained from it. But it as some consequences, like being spammed to update libs when new versions comes out. So it's your choice

this doesn't include errors 400/401/500 cause no MediaTypeObject is set in the ContentObject this should be included nonetheless as if there was one matching mediaType with a schema: z.void()

paths:
  /v1/stores/{store_id}:
    get:
      tags:
        - store_warehouse
      summary: Get the warehouse linked to a store.
      parameters:
        - name: store_id
          in: path
          description: Please fill the store id
          required: true
          schema:
            type: string
      responses:
        "200":
          description: OK
          content:
            application/json:
              schema:
                $ref: "#/components/schemas/d"
        "400":
          description: Ressource Not Found
          content: {}
        "401":
          description: Not Authorized
          content: {}
        "500":
          description: Methods Not Implemented
          content: {}

zod implementation: https://github.com/colinhacks/zod#dates

To write a schema that accepts either a Date or a date string, use z.preprocess.

const dateSchema = z.preprocess((arg) => {
  if (typeof arg == "string" || arg instanceof Date) return new Date(arg);
}, z.date());
type DateSchema = z.infer<typeof dateSchema>;
// type DateSchema = Date

dateSchema.safeParse(new Date("1/12/22")); // success: true
dateSchema.safeParse("2022-01-12T00:00:00.000Z"); // success: true

open api spec description: https://github.com/OAI/OpenAPI-Specification/blob/main/versions/3.0.3.md#data-types

Hi, am I doing anything wrong here?

$ yarn openapi-zod-client plugin-redoc-0.yaml

yarn run v1.22.19

Retrieving OpenAPI document from plugin-redoc-0.yaml
node:internal/process/promises:288
            triggerUncaughtException(err, true /* fromPromise */);
            ^

[Error: ENOENT: no such file or directory, open './src/template.hbs'] {
  errno: -2,
  code: 'ENOENT',
  syscall: 'open',
  path: './src/template.hbs'
}

Node.js v18.0.0
error Command failed with exit code 1.
info Visit https://yarnpkg.com/en/docs/cli/run for documentation about this command.

Hi! I was wondering if there was a way to whitelist / blacklist some endpoints / endpoint aliases using regular expressions, and if it made sense in this project. My use case would be the following:

Say you have to build a small webpage that connects to a really big, old, monolithic API with hundreds of endpoints, but you only had to consume a small subsection of those, say, auth* and analytics. If we had the ability to whitelist and/blacklist, we could run:

yarn run openapi-zod-client --w "^\/api\/v1\/auth.*$" --w "^\/api\/v1\/analytics.*$"
yarn run openapi-zod-client --wa "^apiV1Auth.*$" --w "^apiV1Analytics.*$"

(Just a proposal of syntax, regex expresiones may not work)

To generate a 1000 lines of code instead of 50000 lines, thus making the typescript language server run way faster and reducing out bundle size

Hi! I've just recently started to work with this library, and I'm loving it so far. I have a question about enums.

I have an endpoint with two GET params:

      - in: query
        name: type
        schema:
          type: string
          enum:
          - One
          - Two
          - Three
      - in: query
        name: o
        schema:
          type: array
          items:
            type: string
            enum:
            - (A - Z)
            - (Z - A)

The type param comes from a Django model with an enum that gets exported to the schema as well:

    TypeEnum:
      enum:
      - ONE
      - TWO
      - THREE
      type: string

As you can see, the enum in typed with the DB values, which are all caps, while the values the API will accept are capitalized. I'm trying to set up a mostly code generated codebase, so I'd love to somehow get api.thatEndPoint.parameters.type.schema.options, but AFAIK this is not a trivial matter.

I've been somewhat successful .finding my way around api.api to get to the type.schema.options, but this is not typed, so my beautifully typed requests complain of [string, ...string[]] not being of type One | Two | Three, so this is a no go.

My question is: Is it possible to somehow get the inline enum schema defined in the makeAPi function while retaining typings? And if not, would it be possible to generate and export a zod type for inline enums in the schema?

Thanks!

https://github.com/astahmer/openapi-zod-client/issues/49#issuecomment-1345644640

EDIT: I misunderstood the problem and solved another instead, idk if this is needed so i'll leave this PR as it is for now until someone asks for it

• generating a changelog on new version (new tag pushed) is not working
• idk what happened but the README is gone when visiting the npm page 🥲 since the last release using the CI

i'm a bit in a rush atm, havent had the time to had test for this and tried to deliver fixes asap for these:

https://github.com/astahmer/openapi-zod-client/commit/3675691cc7f3f5ac70299f6be1b60c8b7ba88598 https://github.com/astahmer/openapi-zod-client/commit/17a7091842220091a11d3586cec92299a5d5fbcd