feat: accept camelCase keys in input TypedDicts (#793)

## Summary

Synthesize a `<Name>CamelDict` sibling for every input TypedDict so
users can pass API-shaped dicts (e.g. `{'uniqueKey': ..., 'retryCount':
...}`) and still satisfy the type checker. The Pydantic alias map
extracted from `_models.py` is the source of truth for the camelCase
names; nested TypedDict references in cloned annotations are rewired to
their camel variants.

Resource-client signatures in `_resource_clients/request_queue.py` and
the `WebhooksList` union in `_types.py` are widened to accept either
casing alongside the Pydantic model.

Runtime behavior is unchanged — Pydantic's `populate_by_name=True`
already validates both casings.

Closes #756.