Skip to main content

Upgrading to v2

Copy for LLM

This page summarizes the breaking changes between Apify Python SDK v1.x and v2.0.

Python version support

Support for Python 3.8 has been dropped. The Apify Python SDK v2.x now requires Python 3.9 or later. Make sure your environment is running a compatible version before upgrading.

Storages

  • The SDK now uses crawlee for local storage emulation. This change should not affect intended usage (working with Dataset, KeyValueStore and RequestQueue classes from the apify.storages module or using the shortcuts exposed by the Actor class) in any way.
  • There is a difference in the RequestQueue.add_request method: it accepts an apify.Request object instead of a free-form dictionary.
    • A quick way to migrate from dict-based arguments is to wrap it with a Request.model_validate() call.
    • The preferred way is using the Request.from_url helper which prefills the unique_key and id attributes, or instantiating it directly, e.g., Request(url='https://example.tld', ...).
    • For simple use cases, add_request also accepts plain strings that contain an URL, e.g. queue.add_request('https://example.tld').
  • Removing the StorageClientManager class is a significant change. If you need to change the storage client, use crawlee.service_container instead.

Configuration

The apify.Configuration class now uses pydantic_settings to load configuration from environment variables. This eliminates the need for the helper functions which handled environment variables in apify._utils.

Attributes suffixed with _millis were renamed to remove said suffix and have the datetime.timedelta type now.

Actor

  • The Actor.main method has been removed as it brings no benefits compared to using async with Actor.
  • The Actor.add_webhook, Actor.start, Actor.call and Actor.start_task methods now accept instances of the apify.Webhook model instead of an untyped dict.
  • Actor.start, Actor.call, Actor.start_task, Actor.set_status_message and Actor.abort return instances of the ActorRun model instead of an untyped dict.
  • Upon entering the context manager (async with Actor), the Actor puts the default logging configuration in place. This can be disabled using the configure_logging parameter.
  • The config parameter of Actor has been renamed to configuration.
  • Event handlers registered via Actor.on will now receive Pydantic objects instead of untyped dicts. For example, where you would do event['isMigrating'], you should now use event.is_migrating

Scrapy integration

The apify.scrapy.utils.open_queue_with_custom_client function is not necessary anymore and has been removed.

Subpackage visibility

The following modules were made private:

  • apify.proxy_configuration (ProxyConfiguration is still exported from apify)
  • apify.config (Configuration is still exported from apify)
  • apify.actor (Actor is still exported from apify)
  • apify.event_manager
  • apify.consts