refactor!: Introduce fully typed clients

.github/workflows/_tests.yaml

@@ -20,7 +20,7 @@ jobs:
       operating_systems: '["ubuntu-latest", "windows-latest"]'
       python_version_for_codecov: "3.14"
       operating_system_for_codecov: ubuntu-latest
-      tests_concurrency: "1"
+      tests_concurrency: "16"
 
   integration_tests:
     name: Integration tests
@@ -36,4 +36,4 @@ jobs:
       operating_systems: '["ubuntu-latest"]'
       python_version_for_codecov: "3.14"
       operating_system_for_codecov: ubuntu-latest
-      tests_concurrency: "1"
+      tests_concurrency: "16"

datamodel_codegen_aliases.json

@@ -0,0 +1,3 @@
+{
+    "gitHubGistUrl": "github_gist_url"
+}

docs/01_overview/code/01_usage_async.py

@@ -16,6 +16,6 @@ async def main() -> None:
         return
 
     # Fetch results from the Actor run's default dataset.
-    dataset_client = apify_client.dataset(call_result['defaultDatasetId'])
+    dataset_client = apify_client.dataset(call_result.default_dataset_id)
     list_items_result = await dataset_client.list_items()
     print(f'Dataset: {list_items_result}')

docs/01_overview/code/01_usage_sync.py

@@ -16,6 +16,6 @@ def main() -> None:
         return
 
     # Fetch results from the Actor run's default dataset.
-    dataset_client = apify_client.dataset(call_result['defaultDatasetId'])
+    dataset_client = apify_client.dataset(call_result.default_dataset_id)
     list_items_result = dataset_client.list_items()
     print(f'Dataset: {list_items_result}')

docs/02_concepts/code/01_async_support.py

@@ -11,7 +11,7 @@ async def main() -> None:
 
     # Start the Actor and get the run ID
     run_result = await actor_client.start()
-    run_client = apify_client.run(run_result['id'])
+    run_client = apify_client.run(run_result.id)
     log_client = run_client.log()
 
     # Stream the logs

docs/02_concepts/code/05_retries_async.py

@@ -1,3 +1,5 @@
+from datetime import timedelta
+
 from apify_client import ApifyClientAsync
 
 TOKEN = 'MY-APIFY-TOKEN'
@@ -7,6 +9,6 @@ async def main() -> None:
     apify_client = ApifyClientAsync(
         token=TOKEN,
         max_retries=8,
-        min_delay_between_retries_millis=500,  # 0.5s
-        timeout_secs=360,  # 6 mins
+        min_delay_between_retries=timedelta(milliseconds=500),
+        timeout=timedelta(seconds=360),
     )

docs/02_concepts/code/05_retries_sync.py

@@ -1,12 +1,14 @@
+from datetime import timedelta
+
 from apify_client import ApifyClient
 
 TOKEN = 'MY-APIFY-TOKEN'
 
 
-async def main() -> None:
+def main() -> None:
     apify_client = ApifyClient(
         token=TOKEN,
         max_retries=8,
-        min_delay_between_retries_millis=500,  # 0.5s
-        timeout_secs=360,  # 6 mins
+        min_delay_between_retries=timedelta(milliseconds=500),
+        timeout=timedelta(seconds=360),
     )

docs/03_examples/code/01_input_async.py

@@ -1,4 +1,5 @@
 import asyncio
+from datetime import timedelta
 
 from apify_client import ApifyClientAsync
 
@@ -16,7 +17,9 @@ async def main() -> None:
 
     # Run the Actor and wait for it to finish up to 60 seconds.
     # Input is not persisted for next runs.
-    run_result = await actor_client.call(run_input=input_data, timeout_secs=60)
+    run_result = await actor_client.call(
+        run_input=input_data, timeout=timedelta(seconds=60)
+    )
 
 
 if __name__ == '__main__':

docs/03_examples/code/01_input_sync.py

@@ -1,3 +1,5 @@
+from datetime import timedelta
+
 from apify_client import ApifyClient
 
 TOKEN = 'MY-APIFY-TOKEN'
@@ -14,7 +16,7 @@ def main() -> None:
 
     # Run the Actor and wait for it to finish up to 60 seconds.
     # Input is not persisted for next runs.
-    run_result = actor_client.call(run_input=input_data, timeout_secs=60)
+    run_result = actor_client.call(run_input=input_data, timeout=timedelta(seconds=60))
 
 
 if __name__ == '__main__':

docs/03_examples/code/02_tasks_async.py

@@ -1,22 +1,16 @@
 import asyncio
 
 from apify_client import ApifyClientAsync
-from apify_client.clients.resource_clients import TaskClientAsync
 
 TOKEN = 'MY-APIFY-TOKEN'
 HASHTAGS = ['zebra', 'lion', 'hippo']
 
 
-async def run_apify_task(client: TaskClientAsync) -> dict:
-    result = await client.call()
-    return result or {}
-
-
 async def main() -> None:
     apify_client = ApifyClientAsync(token=TOKEN)
 
     # Create Apify tasks
-    apify_tasks = list[dict]()
+    apify_tasks = []
     apify_tasks_client = apify_client.tasks()
 
     for hashtag in HASHTAGS:
@@ -31,20 +25,19 @@ async def main() -> None:
     print('Tasks created:', apify_tasks)
 
     # Create Apify task clients
-    apify_task_clients = list[TaskClientAsync]()
-
-    for apify_task in apify_tasks:
-        task_id = apify_task['id']
-        apify_task_client = apify_client.task(task_id)
-        apify_task_clients.append(apify_task_client)
+    apify_task_clients = [apify_client.task(task.id) for task in apify_tasks]
 
     print('Task clients created:', apify_task_clients)
 
     # Execute Apify tasks
-    run_apify_tasks = [run_apify_task(client) for client in apify_task_clients]
-    task_run_results = await asyncio.gather(*run_apify_tasks)
+    task_run_results = await asyncio.gather(
+        *[client.call() for client in apify_task_clients]
+    )
+
+    # Filter out None results (tasks that failed to return a run)
+    successful_runs = [run for run in task_run_results if run is not None]
 
-    print('Task results:', task_run_results)
+    print('Task results:', successful_runs)
 
 
 if __name__ == '__main__':

docs/03_examples/code/02_tasks_sync.py

@@ -1,20 +1,14 @@
 from apify_client import ApifyClient
-from apify_client.clients.resource_clients import TaskClient
 
 TOKEN = 'MY-APIFY-TOKEN'
 HASHTAGS = ['zebra', 'lion', 'hippo']
 
 
-def run_apify_task(client: TaskClient) -> dict:
-    result = client.call()
-    return result or {}
-
-
 def main() -> None:
     apify_client = ApifyClient(token=TOKEN)
 
     # Create Apify tasks
-    apify_tasks = list[dict]()
+    apify_tasks = []
     apify_tasks_client = apify_client.tasks()
 
     for hashtag in HASHTAGS:
@@ -29,23 +23,17 @@ def main() -> None:
     print('Tasks created:', apify_tasks)
 
     # Create Apify task clients
-    apify_task_clients = list[TaskClient]()
-
-    for apify_task in apify_tasks:
-        task_id = apify_task['id']
-        apify_task_client = apify_client.task(task_id)
-        apify_task_clients.append(apify_task_client)
+    apify_task_clients = [apify_client.task(task.id) for task in apify_tasks]
 
     print('Task clients created:', apify_task_clients)
 
     # Execute Apify tasks
-    task_run_results = list[dict]()
+    task_run_results = [client.call() for client in apify_task_clients]
 
-    for client in apify_task_clients:
-        result = run_apify_task(client)
-        task_run_results.append(result)
+    # Filter out None results (tasks that failed to return a run)
+    successful_runs = [run for run in task_run_results if run is not None]
 
-    print('Task results:', task_run_results)
+    print('Task results:', successful_runs)
 
 
 if __name__ == '__main__':

docs/03_examples/code/03_retrieve_async.py

@@ -19,11 +19,11 @@ async def main() -> None:
 
     for dataset_item in actor_datasets.items:
         # Dataset items can be handled here. Dataset items can be paginated
-        dataset_client = apify_client.dataset(dataset_item['id'])
+        dataset_client = apify_client.dataset(dataset_item.id)
         dataset_items = await dataset_client.list_items(limit=1000)
 
         # Items can be pushed to single dataset
-        merging_dataset_client = apify_client.dataset(merging_dataset['id'])
+        merging_dataset_client = apify_client.dataset(merging_dataset.id)
         await merging_dataset_client.push_items(dataset_items.items)
 
         # ...

docs/03_examples/code/03_retrieve_sync.py

@@ -17,11 +17,11 @@ def main() -> None:
 
     for dataset_item in actor_datasets.items:
         # Dataset items can be handled here. Dataset items can be paginated
-        dataset_client = apify_client.dataset(dataset_item['id'])
+        dataset_client = apify_client.dataset(dataset_item.id)
         dataset_items = dataset_client.list_items(limit=1000)
 
         # Items can be pushed to single dataset
-        merging_dataset_client = apify_client.dataset(merging_dataset['id'])
+        merging_dataset_client = apify_client.dataset(merging_dataset.id)
         merging_dataset_client.push_items(dataset_items.items)
 
         # ...

pyproject.toml

@@ -25,10 +25,10 @@ classifiers = [
 ]
 keywords = ["apify", "api", "client", "automation", "crawling", "scraping"]
 dependencies = [
-    "apify-shared>=2.1.0,<3.0.0",
     "colorama>=0.4.0",
     "impit>=0.9.2",
     "more_itertools>=10.0.0",
+    "pydantic[email]>=2.11.0",
 ]
 
 [project.urls]
@@ -47,22 +47,23 @@ dev = [
     # See https://github.com/apify/apify-client-python/pull/582/ for more details.
     # We explicitly constrain black>=24.3.0 to override the transitive dependency.
     "black>=24.3.0",
+    "datamodel-code-generator[http,ruff]<1.0.0",
     "dycw-pytest-only<3.0.0",
     "griffe",
     "poethepoet<1.0.0",
     "pre-commit<5.0.0",
     "pydoc-markdown<5.0.0",
     "pytest-asyncio<2.0.0",
     "pytest-cov<8.0.0",
+    "pytest-httpserver<2.0.0",
     "pytest-timeout<3.0.0",
     "pytest-xdist<4.0.0",
     "pytest<9.0.0",
-    "pytest-httpserver<2.0.0",
     "redbaron<1.0.0",
     "ruff~=0.15.0",
     "setuptools", # setuptools are used by pytest but not explicitly required
-    "ty~=0.0.0",
     "types-colorama<0.5.0",
+    "ty~=0.0.0",
     "werkzeug<4.0.0", # Werkzeug is used by pytest-httpserver
 ]
 
@@ -144,6 +145,12 @@ indent-style = "space"
     "N999",   # Invalid module name
     "T201",   # print found
 ]
+"src/apify_client/_models.py" = [
+    "D",      # Everything from the pydocstyle
+    "E501",   # Line too long
+    "ERA001", # Commented-out code
+    "TC003",  # Move standard library import into a type-checking block
+]
 
 [tool.ruff.lint.flake8-quotes]
 docstring-quotes = "double"
@@ -171,10 +178,7 @@ python-version = "3.10"
 include = ["src", "tests", "scripts", "docs", "website"]
 
 [[tool.ty.overrides]]
-include = [
-    "docs/**/*.py",
-    "website/**/*.py",
-]
+include = ["docs/**/*.py", "website/**/*.py"]
 
 [tool.ty.overrides.rules]
 unresolved-import = "ignore"
@@ -185,6 +189,27 @@ exclude_lines = ["pragma: no cover", "if TYPE_CHECKING:", "assert_never()"]
 [tool.ipdb]
 context = 7
 
+# https://koxudaxi.github.io/datamodel-code-generator/
+[tool.datamodel-codegen]
+url = "https://docs.apify.com/api/openapi.json"
+input_file_type = "openapi"
+output = "src/apify_client/_models.py"
+target_python_version = "3.10"
+output_model_type = "pydantic_v2.BaseModel"
+use_schema_description = true
+use_field_description = true
+use_union_operator = true
+capitalise_enum_members = true
+collapse_root_models = true
+set_default_enum_member = true
+use_annotated = true
+wrap_string_literal = true
+snake_case_field = true
+use_subclass_enum = true
+extra_fields = "allow"
+aliases = "datamodel_codegen_aliases.json"
+formatters = ["ruff-check", "ruff-format"]
+
 # Run tasks with: uv run poe <task>
 [tool.poe.tasks]
 clean = "rm -rf .coverage .pytest_cache .ruff_cache .ty_cache build dist htmlcov"
@@ -220,3 +245,6 @@ cwd = "website"
 [tool.poe.tasks.run-docs]
 shell = "./build_api_reference.sh && corepack enable && yarn && uv run yarn start"
 cwd = "website"
+
+[tool.poe.tasks.generate-models]
+shell = "uv run datamodel-codegen"

scripts/fix_async_docstrings.py

@@ -17,7 +17,9 @@
 
         # Find all classes which end with "ClientAsync" (there should be at most 1 per file)
         async_class = red.find('ClassNode', name=re.compile('.*ClientAsync$'))
-        if not async_class:
+
+        if async_class is None:
+            # No async client class in this file, nothing to fix
             continue
 
         # Find the corresponding sync classes (same name, but without -Async)
@@ -32,6 +34,10 @@
             if len(async_method.decorators) and str(async_method.decorators[0].value) == 'ignore_docs':
                 continue
 
+            # Skip methods that don't exist in the sync class
+            if sync_method is None:
+                continue
+
             # If the sync method has a docstring, copy it to the async method (with adjustments)
             if isinstance(sync_method.value[0].value, str):
                 sync_docstring = sync_method.value[0].value

scripts/utils.py

@@ -25,7 +25,7 @@ def get_current_package_version() -> str:
 # It replaces the version number on the line with the format `version = "1.2.3"`
 def set_current_package_version(version: str) -> None:
     with open(PYPROJECT_TOML_FILE_PATH, 'r+', encoding='utf-8') as pyproject_toml_file:
-        updated_pyproject_toml_file_lines = []
+        updated_pyproject_toml_file_lines = list[str]()
         version_string_found = False
         for line in pyproject_toml_file:
             line_processed = line
@@ -45,7 +45,15 @@ def set_current_package_version(version: str) -> None:
 # Generate convert a docstring from a sync resource client method
 # into a doctring for its async resource client analogue
 def sync_to_async_docstring(docstring: str) -> str:
-    substitutions = [(r'Client', r'ClientAsync')]
+    substitutions = [
+        (r'Client', r'ClientAsync'),
+        (r'\bsynchronously\b', r'asynchronously'),
+        (r'\bSynchronously\b', r'Asynchronously'),
+        (r'\bsynchronous\b', r'asynchronous'),
+        (r'\bSynchronous\b', r'Asynchronous'),
+        (r'Retry a function', r'Retry an async function'),
+        (r'Function to retry', r'Async function to retry'),
+    ]
     res = docstring
     for pattern, replacement in substitutions:
         res = re.sub(pattern, replacement, res, flags=re.MULTILINE)
@@ -59,8 +67,8 @@ def get_published_package_versions() -> list:
         package_data = json.load(urlopen(package_info_url))  # noqa: S310
         published_versions = list(package_data['releases'].keys())
     # If the URL returns 404, it means the package has no releases yet (which is okay in our case)
-    except HTTPError as e:
-        if e.code != 404:
+    except HTTPError as exc:
+        if exc.code != 404:
             raise
         published_versions = []
     return published_versions