Invalid type passed for a POST API

[sambk17]

Description

• When the user installs the generated api_client package (in this case mlp_client), the user will want to DescribeModel or RegisterModel which are both POST APIs
  • Note: Both of the components' schemas (RegisterModelRequest & DescribeModelRequest) are objects
• For this issue, I am sharing just the spec for DescribeModel, but can share the spec for RegisterModel if required
• When the user follows the generated instructions (docs), they will use the following Python code:

import mlp_client
from mlp_client.configurations import api_configuration
from mlp_client.apis.tags import model_api
from mlp_client.components.schema import describe_model_request

used_configuration = api_configuration.ApiConfiguration()
api_client = mlp_client.ApiClient(used_configuration)
api_instance = model_api.ModelApi(api_client)

body = describe_model_request.DescribeModelRequest({
        "id": "MG:mlp:test-mlp:v1"
    })

api_response = api_instance.describe_model(
            body=body,
        )

However, this is the response:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/mlp_client/paths/v1_models_describe/post/operation.py", line 92, in _describe_model
    fields, serialized_body = self._get_fields_and_body(
                              ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/mlp_client/api_client.py", line 1255, in _get_fields_and_body
    serialized_data = request_body.serialize(body, content_type, configuration=self.api_client.schema_configuration)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/mlp_client/api_client.py", line 1398, in serialize
    cast_in_data = schema.validate_base(in_data, configuration=used_configuration)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/mlp_client/schemas/schema.py", line 549, in validate_base
    cast_arg = cast_to_allowed_types(
               ^^^^^^^^^^^^^^^^^^^^^^
  File "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/mlp_client/schemas/schema.py", line 277, in cast_to_allowed_types
    raise exceptions.ApiTypeError('Invalid type passed in got input={} type={}'.format(arg, type(arg)))
mlp_client.exceptions.ApiTypeError: Invalid type passed in got input=DescribeModelRequest(types={'id': 'MG:mlp:mlp-sdk-charles-demo:v1'}, properties={'id': <class 'mlp_client.schemas.schemas.StrSchema'>, 'name': <class 'mlp_client.schemas.schemas.StrSchema'>, 'variant': <class 'mlp_client.schemas.schemas.StrSchema'>, 'owner_team': <class 'mlp_client.schemas.schemas.StrSchema'>, 'omd_business_service': <class 'mlp_client.schemas.schemas.StrSchema'>}, type_to_output_cls={<class 'mlp_client.schemas.original_immutabledict.immutabledict'>: <class 'mlp_client.components.schema.describe_model_request.DescribeModelRequestDict'>}) type=<class 'mlp_client.components.schema.describe_model_request.DescribeModelRequest'>

This is what the spec looks like for DescribeModel:

openapi: 3.0.0
info:
  title: MLP [ML Platform SDK] - OpenAPI 3.0
  description: "This SDK produces Model Registration based on the OpenAPI 3.0 specification"
  termsOfService: http://swagger.io/terms/
  contact:
    email: wbd_ml_platform@wbd.com
  license:
    name: Apache 2.0
    url: http://www.apache.org/licenses/LICENSE-2.0.html
  version: 0.0.22
servers:
  - url: http://127.0.0.1:4011
    description: Local test server (using Prism - https://github.com/stoplightio/prism)
tags:
  - name: model
    description: API for Model Describe
paths:
  /v1/models/describe:
    post:
      tags:
        - model
      summary: Describe a registered model
      description: Return a model based on id or name, variant, owner_team & omd_business_service
      operationId: DescribeModel
      requestBody:
        description: Describe a model in the model server
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/DescribeModelRequest'
      responses:
        '201':
          description: Successful operation
        '400':
          description: Invalid ID supplied
components:
  schemas:
    DescribeModelRequest:
      type: object
      properties:
        id:
          type: string
          example: 'MG:mlp:test-mlp:v1'
        name:
          type: string
          example: 'test-mlp'
        variant:
          type: string
          example: 'v1'
        owner_team:
          type: string
          example: 'mlp'
        omd_business_service:
          type: string
          example: 'ml-and-growth'

  securitySchemes:
    api_key:
      type: apiKey
      name: api_key
      in: header

Question: Is this an issue related to immutabledict? Is there something I will need to modify in order to pass in the object as the request?

openapi-generator version

v3.0.0

Generation Details

executed:

java -jar openapi-json-schema-generator-cli.jar generate \
    -i openapi.yaml \
    -g python \
    -o mlp_client \
    --package-name mlp_client

Steps to reproduce

Run the generator, add an empty __init__.py file under mlp_client/ output, pip install the package on my local machine, run the Python code and return the described error above ⬆️ (no pun intended).

[spacether]

Did you look at the documentation for that endpoint?
Hmm I need to fix the documentation.

You are instantiating the body incorrectly. The model class that you are instantiating is the schema class not the output payload class.
Please update your code to be
body = describe_model_request.DescribeModelRequest.validate({ "id": "MG:mlp:test-mlp:v1" })

[sambk17]

@spacether - I have validated that the code works when I add .validate(). Thank you 😄

Since this solution works, should I close this issue? Or should it remain open until the documentation is fixed?

[spacether]

Let's keep it open until I merge the documentation fix