Use `alias_generator=to_camel` on generated Pydantic models instead of per-field aliases

[vdusek]

Background

We're on Pydantic ≥2.11. src/apify_client/_models.py is generated by datamodel-codegen: ~190 models with ~606 explicit Field(alias='camelCase') entries, e.g.:

source_type: Annotated[VersionSourceType | None, Field(alias='sourceType')] = None
is_public: Annotated[bool, Field(alias='isPublic', examples=[False])] = None

All models already share ConfigDict(extra='allow', populate_by_name=True).

Proposal

Drive the codegen so each model's config carries alias_generator=to_camel, and emit an explicit alias only where the conversion is irregular:

from pydantic.alias_generators import to_camel

model_config = ConfigDict(extra='allow', populate_by_name=True, alias_generator=to_camel)

Where it makes sense

Because this file is generated, the work lives in the generation + postprocess pipeline rather than in hand edits:

• Irregular conversions still need an explicit alias. The existing override in datamodel_codegen_aliases.json (gitHubGistUrl → github_gist_url) is a good example: to_camel('github_gist_url') yields githubGistUrl, not gitHubGistUrl, so such overrides must win over the generator.
• scripts/postprocess_generated_models.py builds its alias map by AST-walking Field(alias=...) to synthesize the camelCase *Dict TypedDict variants. If aliases stop being emitted per field, that step must instead derive them via to_camel plus the override map. This is the main thing to verify.

Scope

Refactor only, smaller generated output. The wire format must stay identical, guarded by tests.