fix(rust): oneOf generation for client

[DDtKey]

Closes #17869, #17896, #9497 and also includes unmerged #17898 (fix of serde::rename mappings)

Unfortunately it affects quite a lot of code, but we can see that only client-side models were affected by re-generation.
I tried to split this PR to several, but they're really coupled and hard to create a chain of PRs (e.g #14898 & #17899 and issue with imports).

Other changes:

• imports are aligned between Rust codegens viaAbstractRustCodegen (it used to be error-prone copy/paste with small differences for server & client)
• related improvement is described here: fix(rust): oneOf generation for client #17915 (comment)
• fixed number representation: f32 -> f64

PR checklist

• Read the contribution guidelines.
• Pull Request title clearly describes the work in the pull request and Pull Request description provides details about how to validate the work. Missing information here may result in delayed response from the community.
• Run the following to build the project and update samples:

  ./mvnw clean package 
  ./bin/generate-samples.sh ./bin/configs/*.yaml
  ./bin/utils/export_docs_generators.sh
  

  (For Windows users, please run the script in Git BASH)
  Commit all changed files.
  This is important, as CI jobs will verify all generator outputs of your HEAD commit as it would merge with master.
  These must match the expectations made by your contribution.
  You may regenerate an individual generator by passing the relevant config(s) as an argument to the script, for example ./bin/generate-samples.sh bin/configs/java*.
  IMPORTANT: Do NOT purge/delete any folders/files (e.g. tests) when regenerating the samples as manually written tests may be removed.
• File the PR against the correct branch: master (upcoming 7.1.0 minor release - breaking changes with fallbacks), 8.0.x (breaking changes without fallbacks)
• If your PR is targeting a particular programming language, @mention the technical committee members, so they are more likely to review the pull request.

@frol @farcaller @richardwhiuk @paladinzh @jacob-pro

[DDtKey]

modules/openapi-generator/src/test/resources/3_0/oneOfArrayMapImport.yaml covers issue described in #17869 pretty well

But JFYI, it generates the following enum against example in that PR:

/// Either an account name or an email.
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
#[serde(untagged)]
pub enum AccountOrEmail {
    /// An email address.
    Email(Box<String>),
    /// An account name.
    AccountName(Box<String>),
}

cc @RobbieMcKinstry

[ctrlaltf24]

Thanks for the fixes @DDtKey ! Doesn't look like the samples compile as-is.

cd rust/client/others/rust
cargo check

nor does the cli compile

docker build .

Will look at it in-depth tomorrow to help this along, though we'll still need approval from rust maintainers

[DDtKey]

There were missing changes, sorry

cargo check works on the latest commit

UPD: however I see issues for petstore. Will fix a bit later. I see the reason
UPD 2: fixed

[DDtKey]

A lot of client files contain only such changes - it's now aligned with server-side generators (via AbstractRustCodegen).
Also indentation is fixed in models (used to be too many extra blank lines)

[DDtKey]

It was error-prone because almost copy-paste between rust-generators, so only extracted to AbstractRustCodegen and unified with client .

We can see that only client samples were affected by this PR.

[DDtKey]

PR is ready for review, cleaned-up the code a bit

[RobbieMcKinstry]

modules/openapi-generator/src/test/resources/3_0/oneOfArrayMapImport.yaml covers issue described in #17869 pretty well

But JFYI, it generates the following enum against example in that PR:

/// Either an account name or an email.
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
#[serde(untagged)]
pub enum AccountOrEmail {
    /// An email address.
    Email(Box<String>),
    /// An account name.
    AccountName(Box<String>),
}

cc @RobbieMcKinstry

That looks good to me! It's not obvious to me why the Strings have to be boxed, but that's not even a minor concern for me. 👍🏻

Thank you very much for this PR! ❤️

[DDtKey]

It's not obvious to me why the Strings have to be boxed

Yeah, also find this confusing.
But I don't want to change too much in this PR. I think it can be separate PR to clean-up Box usage. Because this behavior existed before.

[DDtKey]

Just pushed one more improvement:

• prefer mapping name for enum variant if present
• handle the case when mapping has the same ref with different names. E.g:

      "Fruit": {
        "type": "object",
        "oneOf": [
          {
            "$ref": "#/components/schemas/Apple"
          },
          {
            "$ref": "#/components/schemas/Apple"
          }
        ],
        "discriminator": {
          "propertyName": "kind",
          "mapping": {
            "red_apple": "#/components/schemas/Apple",
            "green_apple": "#/components/schemas/Apple"
          }
        }
      },

The result is the following:

#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
#[serde(tag = "kind")]
pub enum Fruit {
  RedApple(Box<models::Apple>),
  GreenApple(Box<models::Apple>),
}

UPD: added test coverage for this

[ctrlaltf24]

concerns addressed, LGTM. Works with my openapi definition that motivated the original change.

Thanks @DDtKey !

[wing328]

thanks for the PR. please review the build failure when you've time, e.g. https://github.com/OpenAPITools/openapi-generator/actions/runs/7996641152/job/21865620537?pr=17915

[DDtKey]

Ouch, sorry, my bad - there is no such method for the interface, but only for a particular type.

Fixed, build works as expected (at least locally)

[DDtKey]

Sorry, @wing328 any chance it will be merged soon?

CI is green and I'd like to update #17931 once this one is merged 🙂

[wing328]

lgtm. let's give it a try

have a nice weekend