Is your feature request related to a problem? Please describe.

I'm struggling to write a schema file to use as input to datamodel-codegen that "combines" multiple schema files using $refs. Is something like this possible?

Input JSON schema files (downloaded to local folder):

• https://raw.githubusercontent.com/CycloneDX/specification/master/schema/bom-1.3.schema.json
• https://raw.githubusercontent.com/CycloneDX/specification/master/schema/bom-1.4.schema.json
• https://raw.githubusercontent.com/CycloneDX/specification/master/schema/bom-1.5.schema.json
• https://raw.githubusercontent.com/CycloneDX/specification/master/schema/jsf-0.82.schema.json
• https://raw.githubusercontent.com/CycloneDX/specification/master/schema/spdx.schema.json

I've tried something along the lines of the following:

{
  "allOf": [
    { "$ref": "specification-1.5/schema/bom-1.3.schema.json" },
    { "$ref": "specification-1.5/schema/bom-1.4.schema.json" },
    { "$ref": "specification-1.5/schema/bom-1.5.schema.json" }
  ]
}

as well as anyOf and oneOf.

With the --use-title-as-name flag, it seems to dump everything into an __init__.py file (with repetitive class names like ComponentType, ComponentType1, ComponentType2) along with an additional spdx.py file.

Without the flag, it generates the folder structure

output_models/
├── __init__.py
├── bom_1
│   ├── __init__.py
│   ├── field_3.py
│   ├── field_4.py
│   └── field_5.py
└── spdx.py

Describe the solution you'd like

Desired generated model directory structure:

output_models
├── cyclonedx_1_3.py
├── cyclonedx_1_4.py
├── cyclonedx_1_5.py
├── jsf_0_82.py
└── spdx.py

Describe alternatives you've considered

Current hacky workaround

Additional context

pyproject.toml settings