types: avoid unnecessary MergeType<> if TOverrides not set, clean up statics and insertMany() type issues

[vkarpov15]

Summary

#13529 points out some issues with MergeType, this PR makes it so that we avoid using MergeType in HydratedDocument<> where possible.

However, for a future release, we should remove HydratedDocument, and instead rely on inferring the hydrated document type from the schema. Inferring HydratedDocument type from the raw document type isn't possible in general, and relies on Omit<>, which is super brittle in TypeScript - breaks on private/protected fields, unions with generics, and other cases.

Examples

[hasezoey]

However, for a future release, we should remove HydratedDocument, and instead rely on inferring the hydrated document type from the schema.

to my knowledge HydratedDocument is useful when you want to assign values explicitly and dont have something like typegoose would provide called DocumentType
(ie have a type that is guaranteed to be the correct and full Document, with Require_id, overrides applied and everything that might be necessary in the future)

[hasezoey]

Looks good to me, also mostly compiles fine with current typegoose

currently typegoose defines HydratedDocument<Something, any>, which makes it incompatible with HydratedDocument<Something, {}>, changing to the following helps, but i dont know if this is the best

TOverrides extends Record<string, never> ?
  Document<unknown, TQueryHelpers, DocType> & Require_id<DocType> :
  IfAny<TOverrides, Document<unknown, TQueryHelpers, DocType> & Require_id<DocType>, Document<unknown, TQueryHelpers, DocType> & MergeType<
    Require_id<DocType>,
    IfAny<TOverrides, {}>
  >>

another option (on typegoose's side) would be to remove the , any from the HydratedDocument call

[vkarpov15]

The idea is to make type HydratedDocument just pull a generic from a schema or model, instead of attempting to reconstruct the hydrated document from the raw document. In other words, right now we infer the raw document from the schema, then infer the hydrated document type from the raw type. We should instead infer both the raw document type and the hydrated document type from the schema. That will avoid having to Omit<>, would also clean up our problem with FlattenMaps in #13523

[vkarpov15]

@hasezoey I applied your suggested updates in 8b8f37b, can you please re-review?

[hasezoey]

Looks good to me and compiles and runs without problems in the current latest typegoose

[orgads]

Thank you!