Is automatic code generation LLM
Not at all. In my case, automatic code generation is a process of automated parsing of an existing Ruby on Rails API code plus some machine-readable comments/syntax I created in the RoR codebase. The way this API was built and versioned, no existing Gem could be used to generate docs. The code generation part is a set of C# "templates" and a parser I built. The parser takes the Ruby API code plus my comments, and generates unit and integration tests for nUnit. This is probably the most common use case for automatic code generation. But... doesn't building unit tests based on existing code potentially create a bad unit test? I'm glad you asked!
The API endpoints are vetted and have their own RoR tests. We rebuilt this API in something more performant than Ruby before we moved it to the cloud. I also built generators that output ASP.NET API endpoint stubs with documentation. So the stubs just get filled out and the test suite is already built. Run Swashbuckle on the new code and out comes the OpenAPI spec, which is then used to build our documentation site and SDKs. The SDKs and docs site are updated in lockstep with any changes to the API.
Edit: extra word and spaces
Ah! Smort! Thanks for the guidance.