The only way to generate OpenApi 3.0 from RapidML specs seems to be a two stage process of
You're correct, we don't have a direct RAPID-ML > OpenAPI v3 GenTemplate yet. The RAPID-XChange Contract (RXC) GenTemplate still produces OpenAPI v2, not v3 at this time.
We are encouraging our customers to move to OpenAPI v3, and would like to have this update in the RAPID-ML code generation pipeline as well. But those priorities are driven by our customers, who have yet to bump this up to our high priority list. So it remains as a roadmap item for the moment.
That said, there may be some ways to accelerate this. I will reach out separately to discuss.
Thanks for the suggestion. This one is on the roadmap!
In general, OpenAPI 3.0 support is the primary goal we are focused on now, including code generation, Normalizer and mock service. An OpenAPI v3 version of the RAPID-XChange GenTemplates will follow.
Oh, that's awesome! Thank you.
I'd be nice to have a documentation describing difference between those two gentemplates. I presume that XChange is either the newest one or the one that is meant for automatic post-processing, but I have never seen any confirmation or documentation on that.
Right, it looks like "RAPID-XChange" is a brand name we pulled out of thin air.
In fact it is a specification that formalizes a mapping from RAPID-ML, which is an abstract model of the API, to a concrete wire format that can be described in OpenAPI 2.0 or 3.0 using one of two modes: "Contract" (RXC) or "DTO" (RXDTO).
The RAPID-XChange specification document is internal, pending finalization.
The RXC GenTemplate is complete, and is currently the only supported RAPID-ML --> OpenAPI conversion path.
The RXDTO GenTemplate is currently labeled as "RAPID-XChange Interop," and is not completely implemented.
And the differences between RAPID-XChange Contract and RAPID-XChange DTO still need to be documented properly.
I think you can expect all of this to be completed and clarified with the RAPID-XChange v3 GenTemplates.
I'm currently searching around for this feature - I notice your comment Ted was a year past and there's a big advert about v3.0 support on the homepage...
I notice you have a v3 editor now but I cannot find a RapidML > v3 gentemplate.
Could you please confirm?
"We are encouraging our customers to move to OpenAPI v3"
Just wondering what the workflow is then without RapidML - handcraft the api in the editor? Then there's no usp I may as well use Swagger Editor or something similar?
What are your customers using RepreZen for then, the code gen part?
I was really enjoying the RapidML part as well :(