perf: enable brotli request compression by default

pyproject.toml

@@ -35,7 +35,10 @@ keywords = [
     "scraping",
 ]
 dependencies = [
-    "apify-client>=3.0.0,<4.0.0",
+    # TODO: Switch to `apify-client[brotli]>=3.1.0,<4.0.0` once apify-client v3.1 is released.
+    # Until then, use the latest beta that already ships the `brotli` extra so brotli request body
+    # compression is enabled by default (falling back to gzip when brotli isn't installed).
+    "apify-client[brotli]>=3.0.7b1,<4.0.0",
     "crawlee>=1.8.0,<2.0.0",
     "cachetools>=5.5.0",
     "cryptography>=42.0.0",

src/apify/_actor.py

@@ -536,6 +536,7 @@ def new_client(
         client_kwargs: dict[str, Any] = {
             'token': token or self.configuration.token,
             'api_url': api_url or self.configuration.api_base_url,
+            'compression': 'brotli',
         }
         if max_retries is not None:
             client_kwargs['max_retries'] = max_retries

src/apify/storage_clients/_apify/_api_client_creation.py

@@ -179,4 +179,5 @@ def _create_api_client(configuration: Configuration) -> ApifyClientAsync:
         api_url=configuration.api_base_url,
         api_public_url=configuration.api_public_base_url,
         max_retries=8,
+        compression='brotli',
     )

tests/unit/storage_clients/test_api_client_creation.py

@@ -1,10 +1,23 @@
 from __future__ import annotations
 
+import json
+from typing import TYPE_CHECKING, Any
+
+import brotli
 import pytest
 
+from apify import Actor
 from apify._configuration import Configuration
 from apify.storage_clients._apify._api_client_creation import _create_api_client, create_storage_api_client
 
+if TYPE_CHECKING:
+    from collections.abc import Callable
+
+    from pytest_httpserver import HTTPServer
+    from werkzeug import Request, Response
+
+    from apify_client import ApifyClientAsync
+
 
 def test_create_api_client_without_token() -> None:
     """Test that _create_api_client raises ValueError when no token is set."""
@@ -50,3 +63,48 @@ async def test_create_storage_unknown_type() -> None:
             storage_type='UnknownType',
             configuration=config,
         )
+
+
+def _client_from_configuration(api_url: str) -> ApifyClientAsync:
+    """Build an API client the way storage clients do, via `_create_api_client`."""
+    config = Configuration(token='test-token', api_base_url=api_url, api_public_base_url=api_url)
+    return _create_api_client(config)
+
+
+def _client_from_new_client(api_url: str) -> ApifyClientAsync:
+    """Build an API client the way user code does, via `Actor.new_client`."""
+    return Actor.new_client(token='test-token', api_url=api_url)
+
+
+@pytest.mark.parametrize(
+    'client_factory',
+    [
+        pytest.param(_client_from_configuration, id='storage_api_client'),
+        pytest.param(_client_from_new_client, id='actor_new_client'),
+    ],
+)
+async def test_sdk_client_compresses_request_body_with_brotli(
+    httpserver: HTTPServer,
+    client_factory: Callable[[str], ApifyClientAsync],
+) -> None:
+    """API clients created by the SDK compress request bodies with brotli (Content-Encoding: br)."""
+    api_url = str(httpserver.url_for('/')).removesuffix('/')
+    client = client_factory(api_url)
+
+    captured: dict[str, Any] = {}
+
+    def request_handler(request: Request, response: Response) -> Response:
+        captured['content_encoding'] = request.headers.get('Content-Encoding')
+        captured['body'] = request.get_data()
+        return response
+
+    httpserver.expect_request('/v2/datasets/test-dataset/items', method='POST').with_post_hook(
+        request_handler
+    ).respond_with_json({'data': {}}, status=201)
+
+    items = [{'hello': 'world'}, {'answer': 42}]
+    await client.dataset(dataset_id='test-dataset').push_items(items)
+
+    assert captured['content_encoding'] == 'br'
+    # The body must be genuinely brotli-compressed and round-trip back to the original items.
+    assert json.loads(brotli.decompress(captured['body'])) == items