RTK-Query Codegen: Camel case applied incorrectly in some situations

[sbward]

In my API I use a variation of camel casing that preserves uppercase in abbreviations. For example, userId becomes userID. Normally this is interpreted as camel case and I would not expect any transformations to be applied to the string.

The issue is that RTK-Query Codegen uses the lodash camelCase function, which converts userID to userId. Lodash's behavior in this regard has been controversial (see lodash/lodash#5045) and I think it would be better to use an alternative (like hump) which has the expected behavior.

Relevant code:

redux-toolkit/packages/rtk-query-codegen-openapi/src/generate.ts

Lines 218 to 221 in 2bb44b9

	const isPureSnakeCase = /^[a-zA-Z][a-zA-Z0-9_]*$/.test(param.name);
	const camelCaseName = camelCase(param.name);
	const name = isPureSnakeCase && !allNames.includes(camelCaseName) ? camelCaseName : param.name;

[jenniferrethi]

This is happening for us as well. Our generated types are all in camelCase but our models are in PascalCase so this is making it hard to actually use the generated types.

[issy]

I have opened a PR (#5174) to allow users to "opt out" of the camelCase behaviour and use exact operation IDs instead. I hope this will be a satisfactory solution both for people affected by this issue and for the Redux maintainers