Description

• All relevant parts of Actor are initialized in async init, not in __init__.
• Actor is considered finalized after Actor.init was run. This also means that the same configuration used by the Actor is set in the global service_locator.
• There are three valid scenarios for setting up the configuration.
  • Setting global configuration in service_locator before the Actor.init
  • Having no configuration set in service_locator and set it through Actor.(configuration=...) and running Actor.init()
  • Having no configuration set in service_locator and no configuration passed to Actor will create and set implicit default configuration
• Properly set ApifyFileSystemStorageClient as local client to support pre-existing input file.
• Depends on refactor!: Refactor storage creation and caching, configuration and services crawlee-python#1386
• Enable caching of ApifyStorageClient based on token and api_public_url and update NDU storage handling.

Issues

Rated to: #513, #590

Testing

• Added many new initialization tests that show possible and prohibited use cases https://github.com/apify/apify-sdk-python/pull/576/files#diff-d64e1d346cc84a225ace3eb1d1ca826ff1e25c77064c9b1e0145552845fa7b41
• Running benchmark actor based on this and the related Crawlee branch