Extract `GotScrapingHttpClient` to a separate package and replace with `FetchHttpClient`

Currently `got-scraping` is a direct dependency of `@crawlee/core`. Extracting the `GotScrapingHttpClient` to a separate package (similar to `@crawlee/impit-client`) and replacing it with a simple `fetch`-based `BaseHttpClient` implementation would help with reducing the dependency tree size. The `fetch`-based client could further serve as a reference HttpClient implementation.

Child of #3275 