metaflow-stubs 2.16.7__py2.py3-none-any.whl → 2.17.0__py2.py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Potentially problematic release.
This version of metaflow-stubs might be problematic. Click here for more details.
- metaflow-stubs/__init__.pyi +692 -692
- metaflow-stubs/cards.pyi +2 -2
- metaflow-stubs/cli.pyi +2 -2
- metaflow-stubs/cli_components/__init__.pyi +2 -2
- metaflow-stubs/cli_components/utils.pyi +2 -2
- metaflow-stubs/client/__init__.pyi +2 -2
- metaflow-stubs/client/core.pyi +5 -5
- metaflow-stubs/client/filecache.pyi +3 -3
- metaflow-stubs/events.pyi +2 -2
- metaflow-stubs/exception.pyi +2 -2
- metaflow-stubs/flowspec.pyi +5 -5
- metaflow-stubs/generated_for.txt +1 -1
- metaflow-stubs/includefile.pyi +4 -4
- metaflow-stubs/meta_files.pyi +2 -2
- metaflow-stubs/metadata_provider/__init__.pyi +2 -2
- metaflow-stubs/metadata_provider/heartbeat.pyi +2 -2
- metaflow-stubs/metadata_provider/metadata.pyi +2 -2
- metaflow-stubs/metadata_provider/util.pyi +2 -2
- metaflow-stubs/metaflow_config.pyi +2 -2
- metaflow-stubs/metaflow_current.pyi +35 -35
- metaflow-stubs/metaflow_git.pyi +2 -2
- metaflow-stubs/multicore_utils.pyi +2 -2
- metaflow-stubs/packaging_sys/__init__.pyi +6 -6
- metaflow-stubs/packaging_sys/backend.pyi +4 -4
- metaflow-stubs/packaging_sys/distribution_support.pyi +4 -4
- metaflow-stubs/packaging_sys/tar_backend.pyi +4 -4
- metaflow-stubs/packaging_sys/utils.pyi +2 -2
- metaflow-stubs/packaging_sys/v1.pyi +3 -3
- metaflow-stubs/parameters.pyi +4 -4
- metaflow-stubs/plugins/__init__.pyi +14 -14
- metaflow-stubs/plugins/airflow/__init__.pyi +2 -2
- metaflow-stubs/plugins/airflow/airflow_decorator.pyi +2 -2
- metaflow-stubs/plugins/airflow/airflow_utils.pyi +2 -2
- metaflow-stubs/plugins/airflow/exception.pyi +2 -2
- metaflow-stubs/plugins/airflow/sensors/__init__.pyi +2 -2
- metaflow-stubs/plugins/airflow/sensors/base_sensor.pyi +2 -2
- metaflow-stubs/plugins/airflow/sensors/external_task_sensor.pyi +2 -2
- metaflow-stubs/plugins/airflow/sensors/s3_sensor.pyi +2 -2
- metaflow-stubs/plugins/argo/__init__.pyi +2 -2
- metaflow-stubs/plugins/argo/argo_client.pyi +2 -2
- metaflow-stubs/plugins/argo/argo_events.pyi +2 -2
- metaflow-stubs/plugins/argo/argo_workflows.pyi +32 -3
- metaflow-stubs/plugins/argo/argo_workflows_decorator.pyi +4 -4
- metaflow-stubs/plugins/argo/argo_workflows_deployer.pyi +3 -3
- metaflow-stubs/plugins/argo/argo_workflows_deployer_objects.pyi +3 -3
- metaflow-stubs/plugins/argo/exit_hooks.pyi +2 -2
- metaflow-stubs/plugins/aws/__init__.pyi +3 -3
- metaflow-stubs/plugins/aws/aws_client.pyi +2 -2
- metaflow-stubs/plugins/aws/aws_utils.pyi +2 -2
- metaflow-stubs/plugins/aws/batch/__init__.pyi +2 -2
- metaflow-stubs/plugins/aws/batch/batch.pyi +2 -2
- metaflow-stubs/plugins/aws/batch/batch_client.pyi +2 -2
- metaflow-stubs/plugins/aws/batch/batch_decorator.pyi +2 -2
- metaflow-stubs/plugins/aws/secrets_manager/__init__.pyi +2 -2
- metaflow-stubs/plugins/aws/secrets_manager/aws_secrets_manager_secrets_provider.pyi +4 -4
- metaflow-stubs/plugins/aws/step_functions/__init__.pyi +2 -2
- metaflow-stubs/plugins/aws/step_functions/dynamo_db_client.pyi +2 -2
- metaflow-stubs/plugins/aws/step_functions/event_bridge_client.pyi +2 -2
- metaflow-stubs/plugins/aws/step_functions/schedule_decorator.pyi +2 -2
- metaflow-stubs/plugins/aws/step_functions/step_functions.pyi +2 -2
- metaflow-stubs/plugins/aws/step_functions/step_functions_client.pyi +2 -2
- metaflow-stubs/plugins/aws/step_functions/step_functions_decorator.pyi +2 -2
- metaflow-stubs/plugins/aws/step_functions/step_functions_deployer.pyi +3 -3
- metaflow-stubs/plugins/aws/step_functions/step_functions_deployer_objects.pyi +3 -3
- metaflow-stubs/plugins/azure/__init__.pyi +2 -2
- metaflow-stubs/plugins/azure/azure_credential.pyi +2 -2
- metaflow-stubs/plugins/azure/azure_exceptions.pyi +2 -2
- metaflow-stubs/plugins/azure/azure_secret_manager_secrets_provider.pyi +4 -4
- metaflow-stubs/plugins/azure/azure_utils.pyi +2 -2
- metaflow-stubs/plugins/azure/blob_service_client_factory.pyi +2 -2
- metaflow-stubs/plugins/azure/includefile_support.pyi +2 -2
- metaflow-stubs/plugins/cards/__init__.pyi +2 -2
- metaflow-stubs/plugins/cards/card_client.pyi +2 -2
- metaflow-stubs/plugins/cards/card_creator.pyi +2 -2
- metaflow-stubs/plugins/cards/card_datastore.pyi +2 -2
- metaflow-stubs/plugins/cards/card_decorator.pyi +2 -2
- metaflow-stubs/plugins/cards/card_modules/__init__.pyi +2 -2
- metaflow-stubs/plugins/cards/card_modules/basic.pyi +2 -2
- metaflow-stubs/plugins/cards/card_modules/card.pyi +2 -2
- metaflow-stubs/plugins/cards/card_modules/components.pyi +3 -3
- metaflow-stubs/plugins/cards/card_modules/convert_to_native_type.pyi +2 -2
- metaflow-stubs/plugins/cards/card_modules/renderer_tools.pyi +2 -2
- metaflow-stubs/plugins/cards/card_modules/test_cards.pyi +2 -2
- metaflow-stubs/plugins/cards/card_resolver.pyi +2 -2
- metaflow-stubs/plugins/cards/component_serializer.pyi +2 -2
- metaflow-stubs/plugins/cards/exception.pyi +2 -2
- metaflow-stubs/plugins/catch_decorator.pyi +3 -3
- metaflow-stubs/plugins/datatools/__init__.pyi +2 -2
- metaflow-stubs/plugins/datatools/local.pyi +2 -2
- metaflow-stubs/plugins/datatools/s3/__init__.pyi +2 -2
- metaflow-stubs/plugins/datatools/s3/s3.pyi +2 -2
- metaflow-stubs/plugins/datatools/s3/s3tail.pyi +2 -2
- metaflow-stubs/plugins/datatools/s3/s3util.pyi +2 -2
- metaflow-stubs/plugins/debug_logger.pyi +2 -2
- metaflow-stubs/plugins/debug_monitor.pyi +2 -2
- metaflow-stubs/plugins/environment_decorator.pyi +2 -2
- metaflow-stubs/plugins/events_decorator.pyi +2 -2
- metaflow-stubs/plugins/exit_hook/__init__.pyi +2 -2
- metaflow-stubs/plugins/exit_hook/exit_hook_decorator.pyi +2 -2
- metaflow-stubs/plugins/frameworks/__init__.pyi +2 -2
- metaflow-stubs/plugins/frameworks/pytorch.pyi +2 -2
- metaflow-stubs/plugins/gcp/__init__.pyi +2 -2
- metaflow-stubs/plugins/gcp/gcp_secret_manager_secrets_provider.pyi +4 -4
- metaflow-stubs/plugins/gcp/gs_exceptions.pyi +2 -2
- metaflow-stubs/plugins/gcp/gs_storage_client_factory.pyi +2 -2
- metaflow-stubs/plugins/gcp/gs_utils.pyi +2 -2
- metaflow-stubs/plugins/gcp/includefile_support.pyi +2 -2
- metaflow-stubs/plugins/kubernetes/__init__.pyi +2 -2
- metaflow-stubs/plugins/kubernetes/kube_utils.pyi +2 -2
- metaflow-stubs/plugins/kubernetes/kubernetes.pyi +2 -2
- metaflow-stubs/plugins/kubernetes/kubernetes_client.pyi +2 -2
- metaflow-stubs/plugins/kubernetes/kubernetes_decorator.pyi +2 -2
- metaflow-stubs/plugins/kubernetes/kubernetes_job.pyi +2 -2
- metaflow-stubs/plugins/kubernetes/kubernetes_jobsets.pyi +3 -2
- metaflow-stubs/plugins/kubernetes/spot_monitor_sidecar.pyi +2 -2
- metaflow-stubs/plugins/parallel_decorator.pyi +2 -2
- metaflow-stubs/plugins/project_decorator.pyi +2 -2
- metaflow-stubs/plugins/pypi/__init__.pyi +2 -2
- metaflow-stubs/plugins/pypi/conda_decorator.pyi +2 -2
- metaflow-stubs/plugins/pypi/conda_environment.pyi +5 -5
- metaflow-stubs/plugins/pypi/parsers.pyi +2 -2
- metaflow-stubs/plugins/pypi/pypi_decorator.pyi +2 -2
- metaflow-stubs/plugins/pypi/pypi_environment.pyi +2 -2
- metaflow-stubs/plugins/pypi/utils.pyi +2 -2
- metaflow-stubs/plugins/resources_decorator.pyi +2 -2
- metaflow-stubs/plugins/retry_decorator.pyi +2 -2
- metaflow-stubs/plugins/secrets/__init__.pyi +3 -3
- metaflow-stubs/plugins/secrets/inline_secrets_provider.pyi +3 -3
- metaflow-stubs/plugins/secrets/secrets_decorator.pyi +2 -2
- metaflow-stubs/plugins/secrets/secrets_func.pyi +2 -2
- metaflow-stubs/plugins/secrets/secrets_spec.pyi +2 -2
- metaflow-stubs/plugins/secrets/utils.pyi +2 -2
- metaflow-stubs/plugins/storage_executor.pyi +2 -2
- metaflow-stubs/plugins/test_unbounded_foreach_decorator.pyi +2 -2
- metaflow-stubs/plugins/timeout_decorator.pyi +3 -3
- metaflow-stubs/plugins/uv/__init__.pyi +2 -2
- metaflow-stubs/plugins/uv/uv_environment.pyi +3 -3
- metaflow-stubs/pylint_wrapper.pyi +2 -2
- metaflow-stubs/runner/__init__.pyi +2 -2
- metaflow-stubs/runner/deployer.pyi +32 -32
- metaflow-stubs/runner/deployer_impl.pyi +3 -3
- metaflow-stubs/runner/metaflow_runner.pyi +3 -3
- metaflow-stubs/runner/nbdeploy.pyi +2 -2
- metaflow-stubs/runner/nbrun.pyi +2 -2
- metaflow-stubs/runner/subprocess_manager.pyi +2 -2
- metaflow-stubs/runner/utils.pyi +4 -4
- metaflow-stubs/system/__init__.pyi +2 -2
- metaflow-stubs/system/system_logger.pyi +3 -3
- metaflow-stubs/system/system_monitor.pyi +2 -2
- metaflow-stubs/tagging_util.pyi +2 -2
- metaflow-stubs/tuple_util.pyi +2 -2
- metaflow-stubs/user_configs/__init__.pyi +2 -2
- metaflow-stubs/user_configs/config_options.pyi +3 -3
- metaflow-stubs/user_configs/config_parameters.pyi +6 -6
- metaflow-stubs/user_decorators/__init__.pyi +2 -2
- metaflow-stubs/user_decorators/common.pyi +2 -2
- metaflow-stubs/user_decorators/mutable_flow.pyi +6 -6
- metaflow-stubs/user_decorators/mutable_step.pyi +5 -5
- metaflow-stubs/user_decorators/user_flow_decorator.pyi +5 -5
- metaflow-stubs/user_decorators/user_step_decorator.pyi +4 -4
- metaflow-stubs/version.pyi +2 -2
- {metaflow_stubs-2.16.7.dist-info → metaflow_stubs-2.17.0.dist-info}/METADATA +2 -2
- metaflow_stubs-2.17.0.dist-info/RECORD +166 -0
- metaflow_stubs-2.16.7.dist-info/RECORD +0 -166
- {metaflow_stubs-2.16.7.dist-info → metaflow_stubs-2.17.0.dist-info}/WHEEL +0 -0
- {metaflow_stubs-2.16.7.dist-info → metaflow_stubs-2.17.0.dist-info}/top_level.txt +0 -0
metaflow-stubs/__init__.pyi
CHANGED
|
@@ -1,15 +1,15 @@
|
|
|
1
1
|
######################################################################################################
|
|
2
2
|
# Auto-generated Metaflow stub file #
|
|
3
|
-
# MF version: 2.
|
|
4
|
-
# Generated on 2025-
|
|
3
|
+
# MF version: 2.17.0 #
|
|
4
|
+
# Generated on 2025-08-06T11:05:04.016056 #
|
|
5
5
|
######################################################################################################
|
|
6
6
|
|
|
7
7
|
from __future__ import annotations
|
|
8
8
|
|
|
9
9
|
import typing
|
|
10
10
|
if typing.TYPE_CHECKING:
|
|
11
|
-
import datetime
|
|
12
11
|
import typing
|
|
12
|
+
import datetime
|
|
13
13
|
FlowSpecDerived = typing.TypeVar("FlowSpecDerived", bound="FlowSpec", contravariant=False, covariant=False)
|
|
14
14
|
StepFlag = typing.NewType("StepFlag", bool)
|
|
15
15
|
|
|
@@ -39,17 +39,17 @@ from .user_decorators.user_step_decorator import UserStepDecorator as UserStepDe
|
|
|
39
39
|
from .user_decorators.user_step_decorator import StepMutator as StepMutator
|
|
40
40
|
from .user_decorators.user_step_decorator import user_step_decorator as user_step_decorator
|
|
41
41
|
from .user_decorators.user_flow_decorator import FlowMutator as FlowMutator
|
|
42
|
+
from . import tuple_util as tuple_util
|
|
42
43
|
from . import metaflow_git as metaflow_git
|
|
43
44
|
from . import events as events
|
|
44
|
-
from . import tuple_util as tuple_util
|
|
45
45
|
from . import runner as runner
|
|
46
46
|
from . import plugins as plugins
|
|
47
47
|
from .plugins.datatools.s3.s3 import S3 as S3
|
|
48
48
|
from . import includefile as includefile
|
|
49
49
|
from .includefile import IncludeFile as IncludeFile
|
|
50
50
|
from .plugins.pypi.parsers import pyproject_toml_parser as pyproject_toml_parser
|
|
51
|
-
from .plugins.pypi.parsers import conda_environment_yml_parser as conda_environment_yml_parser
|
|
52
51
|
from .plugins.pypi.parsers import requirements_txt_parser as requirements_txt_parser
|
|
52
|
+
from .plugins.pypi.parsers import conda_environment_yml_parser as conda_environment_yml_parser
|
|
53
53
|
from . import cards as cards
|
|
54
54
|
from . import client as client
|
|
55
55
|
from .client.core import namespace as namespace
|
|
@@ -153,109 +153,170 @@ def step(f: typing.Union[typing.Callable[[FlowSpecDerived], None], typing.Callab
|
|
|
153
153
|
...
|
|
154
154
|
|
|
155
155
|
@typing.overload
|
|
156
|
-
def
|
|
156
|
+
def resources(*, cpu: int = 1, gpu: typing.Optional[int] = None, disk: typing.Optional[int] = None, memory: int = 4096, shared_memory: typing.Optional[int] = None) -> typing.Callable[[typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]], typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]]:
|
|
157
157
|
"""
|
|
158
|
-
Specifies
|
|
159
|
-
the execution of a step.
|
|
158
|
+
Specifies the resources needed when executing this step.
|
|
160
159
|
|
|
160
|
+
Use `@resources` to specify the resource requirements
|
|
161
|
+
independently of the specific compute layer (`@batch`, `@kubernetes`).
|
|
161
162
|
|
|
162
|
-
|
|
163
|
-
|
|
164
|
-
|
|
165
|
-
|
|
166
|
-
|
|
167
|
-
|
|
168
|
-
|
|
169
|
-
|
|
170
|
-
|
|
171
|
-
|
|
172
|
-
def secrets(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
173
|
-
...
|
|
174
|
-
|
|
175
|
-
@typing.overload
|
|
176
|
-
def secrets(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
177
|
-
...
|
|
178
|
-
|
|
179
|
-
def secrets(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, sources: typing.List[typing.Union[str, typing.Dict[str, typing.Any]]] = [], role: typing.Optional[str] = None):
|
|
180
|
-
"""
|
|
181
|
-
Specifies secrets to be retrieved and injected as environment variables prior to
|
|
182
|
-
the execution of a step.
|
|
163
|
+
You can choose the compute layer on the command line by executing e.g.
|
|
164
|
+
```
|
|
165
|
+
python myflow.py run --with batch
|
|
166
|
+
```
|
|
167
|
+
or
|
|
168
|
+
```
|
|
169
|
+
python myflow.py run --with kubernetes
|
|
170
|
+
```
|
|
171
|
+
which executes the flow on the desired system using the
|
|
172
|
+
requirements specified in `@resources`.
|
|
183
173
|
|
|
184
174
|
|
|
185
175
|
Parameters
|
|
186
176
|
----------
|
|
187
|
-
|
|
188
|
-
|
|
189
|
-
|
|
190
|
-
|
|
177
|
+
cpu : int, default 1
|
|
178
|
+
Number of CPUs required for this step.
|
|
179
|
+
gpu : int, optional, default None
|
|
180
|
+
Number of GPUs required for this step.
|
|
181
|
+
disk : int, optional, default None
|
|
182
|
+
Disk size (in MB) required for this step. Only applies on Kubernetes.
|
|
183
|
+
memory : int, default 4096
|
|
184
|
+
Memory size (in MB) required for this step.
|
|
185
|
+
shared_memory : int, optional, default None
|
|
186
|
+
The value for the size (in MiB) of the /dev/shm volume for this step.
|
|
187
|
+
This parameter maps to the `--shm-size` option in Docker.
|
|
191
188
|
"""
|
|
192
189
|
...
|
|
193
190
|
|
|
194
191
|
@typing.overload
|
|
195
|
-
def
|
|
196
|
-
"""
|
|
197
|
-
Decorator prototype for all step decorators. This function gets specialized
|
|
198
|
-
and imported for all decorators types by _import_plugin_decorators().
|
|
199
|
-
"""
|
|
192
|
+
def resources(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
200
193
|
...
|
|
201
194
|
|
|
202
195
|
@typing.overload
|
|
203
|
-
def
|
|
204
|
-
...
|
|
205
|
-
|
|
206
|
-
def parallel(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None):
|
|
207
|
-
"""
|
|
208
|
-
Decorator prototype for all step decorators. This function gets specialized
|
|
209
|
-
and imported for all decorators types by _import_plugin_decorators().
|
|
210
|
-
"""
|
|
196
|
+
def resources(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
211
197
|
...
|
|
212
198
|
|
|
213
|
-
|
|
214
|
-
def card(*, type: str = 'default', id: typing.Optional[str] = None, options: typing.Dict[str, typing.Any] = {}, timeout: int = 45) -> typing.Callable[[typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]], typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]]:
|
|
199
|
+
def resources(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, cpu: int = 1, gpu: typing.Optional[int] = None, disk: typing.Optional[int] = None, memory: int = 4096, shared_memory: typing.Optional[int] = None):
|
|
215
200
|
"""
|
|
216
|
-
|
|
201
|
+
Specifies the resources needed when executing this step.
|
|
217
202
|
|
|
218
|
-
|
|
203
|
+
Use `@resources` to specify the resource requirements
|
|
204
|
+
independently of the specific compute layer (`@batch`, `@kubernetes`).
|
|
205
|
+
|
|
206
|
+
You can choose the compute layer on the command line by executing e.g.
|
|
207
|
+
```
|
|
208
|
+
python myflow.py run --with batch
|
|
209
|
+
```
|
|
210
|
+
or
|
|
211
|
+
```
|
|
212
|
+
python myflow.py run --with kubernetes
|
|
213
|
+
```
|
|
214
|
+
which executes the flow on the desired system using the
|
|
215
|
+
requirements specified in `@resources`.
|
|
219
216
|
|
|
220
217
|
|
|
221
218
|
Parameters
|
|
222
219
|
----------
|
|
223
|
-
|
|
224
|
-
|
|
225
|
-
|
|
226
|
-
|
|
227
|
-
|
|
228
|
-
|
|
229
|
-
|
|
230
|
-
|
|
220
|
+
cpu : int, default 1
|
|
221
|
+
Number of CPUs required for this step.
|
|
222
|
+
gpu : int, optional, default None
|
|
223
|
+
Number of GPUs required for this step.
|
|
224
|
+
disk : int, optional, default None
|
|
225
|
+
Disk size (in MB) required for this step. Only applies on Kubernetes.
|
|
226
|
+
memory : int, default 4096
|
|
227
|
+
Memory size (in MB) required for this step.
|
|
228
|
+
shared_memory : int, optional, default None
|
|
229
|
+
The value for the size (in MiB) of the /dev/shm volume for this step.
|
|
230
|
+
This parameter maps to the `--shm-size` option in Docker.
|
|
231
231
|
"""
|
|
232
232
|
...
|
|
233
233
|
|
|
234
|
-
|
|
235
|
-
def card(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
236
|
-
...
|
|
237
|
-
|
|
238
|
-
@typing.overload
|
|
239
|
-
def card(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
240
|
-
...
|
|
241
|
-
|
|
242
|
-
def card(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, type: str = 'default', id: typing.Optional[str] = None, options: typing.Dict[str, typing.Any] = {}, timeout: int = 45):
|
|
234
|
+
def kubernetes(*, cpu: int = 1, memory: int = 4096, disk: int = 10240, image: typing.Optional[str] = None, image_pull_policy: str = 'KUBERNETES_IMAGE_PULL_POLICY', image_pull_secrets: typing.List[str] = [], service_account: str = 'METAFLOW_KUBERNETES_SERVICE_ACCOUNT', secrets: typing.Optional[typing.List[str]] = None, node_selector: typing.Union[typing.Dict[str, str], str, None] = None, namespace: str = 'METAFLOW_KUBERNETES_NAMESPACE', gpu: typing.Optional[int] = None, gpu_vendor: str = 'KUBERNETES_GPU_VENDOR', tolerations: typing.List[typing.Dict[str, str]] = [], labels: typing.Dict[str, str] = 'METAFLOW_KUBERNETES_LABELS', annotations: typing.Dict[str, str] = 'METAFLOW_KUBERNETES_ANNOTATIONS', use_tmpfs: bool = False, tmpfs_tempdir: bool = True, tmpfs_size: typing.Optional[int] = None, tmpfs_path: typing.Optional[str] = '/metaflow_temp', persistent_volume_claims: typing.Optional[typing.Dict[str, str]] = None, shared_memory: typing.Optional[int] = None, port: typing.Optional[int] = None, compute_pool: typing.Optional[str] = None, hostname_resolution_timeout: int = 600, qos: str = 'Burstable', security_context: typing.Optional[typing.Dict[str, typing.Any]] = None) -> typing.Callable[[typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]], typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]]:
|
|
243
235
|
"""
|
|
244
|
-
|
|
245
|
-
|
|
246
|
-
Note that you may add multiple `@card` decorators in a step with different parameters.
|
|
236
|
+
Specifies that this step should execute on Kubernetes.
|
|
247
237
|
|
|
248
238
|
|
|
249
239
|
Parameters
|
|
250
240
|
----------
|
|
251
|
-
|
|
252
|
-
|
|
253
|
-
|
|
254
|
-
|
|
255
|
-
|
|
256
|
-
|
|
257
|
-
|
|
258
|
-
|
|
241
|
+
cpu : int, default 1
|
|
242
|
+
Number of CPUs required for this step. If `@resources` is
|
|
243
|
+
also present, the maximum value from all decorators is used.
|
|
244
|
+
memory : int, default 4096
|
|
245
|
+
Memory size (in MB) required for this step. If
|
|
246
|
+
`@resources` is also present, the maximum value from all decorators is
|
|
247
|
+
used.
|
|
248
|
+
disk : int, default 10240
|
|
249
|
+
Disk size (in MB) required for this step. If
|
|
250
|
+
`@resources` is also present, the maximum value from all decorators is
|
|
251
|
+
used.
|
|
252
|
+
image : str, optional, default None
|
|
253
|
+
Docker image to use when launching on Kubernetes. If not specified, and
|
|
254
|
+
METAFLOW_KUBERNETES_CONTAINER_IMAGE is specified, that image is used. If
|
|
255
|
+
not, a default Docker image mapping to the current version of Python is used.
|
|
256
|
+
image_pull_policy: str, default KUBERNETES_IMAGE_PULL_POLICY
|
|
257
|
+
If given, the imagePullPolicy to be applied to the Docker image of the step.
|
|
258
|
+
image_pull_secrets: List[str], default []
|
|
259
|
+
The default is extracted from METAFLOW_KUBERNETES_IMAGE_PULL_SECRETS.
|
|
260
|
+
Kubernetes image pull secrets to use when pulling container images
|
|
261
|
+
in Kubernetes.
|
|
262
|
+
service_account : str, default METAFLOW_KUBERNETES_SERVICE_ACCOUNT
|
|
263
|
+
Kubernetes service account to use when launching pod in Kubernetes.
|
|
264
|
+
secrets : List[str], optional, default None
|
|
265
|
+
Kubernetes secrets to use when launching pod in Kubernetes. These
|
|
266
|
+
secrets are in addition to the ones defined in `METAFLOW_KUBERNETES_SECRETS`
|
|
267
|
+
in Metaflow configuration.
|
|
268
|
+
node_selector: Union[Dict[str,str], str], optional, default None
|
|
269
|
+
Kubernetes node selector(s) to apply to the pod running the task.
|
|
270
|
+
Can be passed in as a comma separated string of values e.g.
|
|
271
|
+
'kubernetes.io/os=linux,kubernetes.io/arch=amd64' or as a dictionary
|
|
272
|
+
{'kubernetes.io/os': 'linux', 'kubernetes.io/arch': 'amd64'}
|
|
273
|
+
namespace : str, default METAFLOW_KUBERNETES_NAMESPACE
|
|
274
|
+
Kubernetes namespace to use when launching pod in Kubernetes.
|
|
275
|
+
gpu : int, optional, default None
|
|
276
|
+
Number of GPUs required for this step. A value of zero implies that
|
|
277
|
+
the scheduled node should not have GPUs.
|
|
278
|
+
gpu_vendor : str, default KUBERNETES_GPU_VENDOR
|
|
279
|
+
The vendor of the GPUs to be used for this step.
|
|
280
|
+
tolerations : List[Dict[str,str]], default []
|
|
281
|
+
The default is extracted from METAFLOW_KUBERNETES_TOLERATIONS.
|
|
282
|
+
Kubernetes tolerations to use when launching pod in Kubernetes.
|
|
283
|
+
labels: Dict[str, str], default: METAFLOW_KUBERNETES_LABELS
|
|
284
|
+
Kubernetes labels to use when launching pod in Kubernetes.
|
|
285
|
+
annotations: Dict[str, str], default: METAFLOW_KUBERNETES_ANNOTATIONS
|
|
286
|
+
Kubernetes annotations to use when launching pod in Kubernetes.
|
|
287
|
+
use_tmpfs : bool, default False
|
|
288
|
+
This enables an explicit tmpfs mount for this step.
|
|
289
|
+
tmpfs_tempdir : bool, default True
|
|
290
|
+
sets METAFLOW_TEMPDIR to tmpfs_path if set for this step.
|
|
291
|
+
tmpfs_size : int, optional, default: None
|
|
292
|
+
The value for the size (in MiB) of the tmpfs mount for this step.
|
|
293
|
+
This parameter maps to the `--tmpfs` option in Docker. Defaults to 50% of the
|
|
294
|
+
memory allocated for this step.
|
|
295
|
+
tmpfs_path : str, optional, default /metaflow_temp
|
|
296
|
+
Path to tmpfs mount for this step.
|
|
297
|
+
persistent_volume_claims : Dict[str, str], optional, default None
|
|
298
|
+
A map (dictionary) of persistent volumes to be mounted to the pod for this step. The map is from persistent
|
|
299
|
+
volumes to the path to which the volume is to be mounted, e.g., `{'pvc-name': '/path/to/mount/on'}`.
|
|
300
|
+
shared_memory: int, optional
|
|
301
|
+
Shared memory size (in MiB) required for this step
|
|
302
|
+
port: int, optional
|
|
303
|
+
Port number to specify in the Kubernetes job object
|
|
304
|
+
compute_pool : str, optional, default None
|
|
305
|
+
Compute pool to be used for for this step.
|
|
306
|
+
If not specified, any accessible compute pool within the perimeter is used.
|
|
307
|
+
hostname_resolution_timeout: int, default 10 * 60
|
|
308
|
+
Timeout in seconds for the workers tasks in the gang scheduled cluster to resolve the hostname of control task.
|
|
309
|
+
Only applicable when @parallel is used.
|
|
310
|
+
qos: str, default: Burstable
|
|
311
|
+
Quality of Service class to assign to the pod. Supported values are: Guaranteed, Burstable, BestEffort
|
|
312
|
+
|
|
313
|
+
security_context: Dict[str, Any], optional, default None
|
|
314
|
+
Container security context. Applies to the task container. Allows the following keys:
|
|
315
|
+
- privileged: bool, optional, default None
|
|
316
|
+
- allow_privilege_escalation: bool, optional, default None
|
|
317
|
+
- run_as_user: int, optional, default None
|
|
318
|
+
- run_as_group: int, optional, default None
|
|
319
|
+
- run_as_non_root: bool, optional, default None
|
|
259
320
|
"""
|
|
260
321
|
...
|
|
261
322
|
|
|
@@ -319,132 +380,200 @@ def timeout(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None],
|
|
|
319
380
|
...
|
|
320
381
|
|
|
321
382
|
@typing.overload
|
|
322
|
-
def
|
|
323
|
-
"""
|
|
324
|
-
Specifies that the step will success under all circumstances.
|
|
325
|
-
|
|
326
|
-
The decorator will create an optional artifact, specified by `var`, which
|
|
327
|
-
contains the exception raised. You can use it to detect the presence
|
|
328
|
-
of errors, indicating that all happy-path artifacts produced by the step
|
|
329
|
-
are missing.
|
|
330
|
-
|
|
331
|
-
|
|
332
|
-
Parameters
|
|
333
|
-
----------
|
|
334
|
-
var : str, optional, default None
|
|
335
|
-
Name of the artifact in which to store the caught exception.
|
|
336
|
-
If not specified, the exception is not stored.
|
|
337
|
-
print_exception : bool, default True
|
|
338
|
-
Determines whether or not the exception is printed to
|
|
339
|
-
stdout when caught.
|
|
340
|
-
"""
|
|
341
|
-
...
|
|
342
|
-
|
|
343
|
-
@typing.overload
|
|
344
|
-
def catch(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
345
|
-
...
|
|
346
|
-
|
|
347
|
-
@typing.overload
|
|
348
|
-
def catch(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
349
|
-
...
|
|
350
|
-
|
|
351
|
-
def catch(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, var: typing.Optional[str] = None, print_exception: bool = True):
|
|
383
|
+
def batch(*, cpu: int = 1, gpu: int = 0, memory: int = 4096, image: typing.Optional[str] = None, queue: str = 'METAFLOW_BATCH_JOB_QUEUE', iam_role: str = 'METAFLOW_ECS_S3_ACCESS_IAM_ROLE', execution_role: str = 'METAFLOW_ECS_FARGATE_EXECUTION_ROLE', shared_memory: typing.Optional[int] = None, max_swap: typing.Optional[int] = None, swappiness: typing.Optional[int] = None, use_tmpfs: bool = False, tmpfs_tempdir: bool = True, tmpfs_size: typing.Optional[int] = None, tmpfs_path: typing.Optional[str] = None, inferentia: int = 0, trainium: int = None, efa: int = 0, ephemeral_storage: int = None, log_driver: typing.Optional[str] = None, log_options: typing.Optional[typing.List[str]] = None) -> typing.Callable[[typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]], typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]]:
|
|
352
384
|
"""
|
|
353
|
-
Specifies that
|
|
354
|
-
|
|
355
|
-
The decorator will create an optional artifact, specified by `var`, which
|
|
356
|
-
contains the exception raised. You can use it to detect the presence
|
|
357
|
-
of errors, indicating that all happy-path artifacts produced by the step
|
|
358
|
-
are missing.
|
|
385
|
+
Specifies that this step should execute on [AWS Batch](https://aws.amazon.com/batch/).
|
|
359
386
|
|
|
360
387
|
|
|
361
388
|
Parameters
|
|
362
389
|
----------
|
|
363
|
-
|
|
364
|
-
|
|
365
|
-
|
|
366
|
-
|
|
367
|
-
|
|
368
|
-
|
|
390
|
+
cpu : int, default 1
|
|
391
|
+
Number of CPUs required for this step. If `@resources` is
|
|
392
|
+
also present, the maximum value from all decorators is used.
|
|
393
|
+
gpu : int, default 0
|
|
394
|
+
Number of GPUs required for this step. If `@resources` is
|
|
395
|
+
also present, the maximum value from all decorators is used.
|
|
396
|
+
memory : int, default 4096
|
|
397
|
+
Memory size (in MB) required for this step. If
|
|
398
|
+
`@resources` is also present, the maximum value from all decorators is
|
|
399
|
+
used.
|
|
400
|
+
image : str, optional, default None
|
|
401
|
+
Docker image to use when launching on AWS Batch. If not specified, and
|
|
402
|
+
METAFLOW_BATCH_CONTAINER_IMAGE is specified, that image is used. If
|
|
403
|
+
not, a default Docker image mapping to the current version of Python is used.
|
|
404
|
+
queue : str, default METAFLOW_BATCH_JOB_QUEUE
|
|
405
|
+
AWS Batch Job Queue to submit the job to.
|
|
406
|
+
iam_role : str, default METAFLOW_ECS_S3_ACCESS_IAM_ROLE
|
|
407
|
+
AWS IAM role that AWS Batch container uses to access AWS cloud resources.
|
|
408
|
+
execution_role : str, default METAFLOW_ECS_FARGATE_EXECUTION_ROLE
|
|
409
|
+
AWS IAM role that AWS Batch can use [to trigger AWS Fargate tasks]
|
|
410
|
+
(https://docs.aws.amazon.com/batch/latest/userguide/execution-IAM-role.html).
|
|
411
|
+
shared_memory : int, optional, default None
|
|
412
|
+
The value for the size (in MiB) of the /dev/shm volume for this step.
|
|
413
|
+
This parameter maps to the `--shm-size` option in Docker.
|
|
414
|
+
max_swap : int, optional, default None
|
|
415
|
+
The total amount of swap memory (in MiB) a container can use for this
|
|
416
|
+
step. This parameter is translated to the `--memory-swap` option in
|
|
417
|
+
Docker where the value is the sum of the container memory plus the
|
|
418
|
+
`max_swap` value.
|
|
419
|
+
swappiness : int, optional, default None
|
|
420
|
+
This allows you to tune memory swappiness behavior for this step.
|
|
421
|
+
A swappiness value of 0 causes swapping not to happen unless absolutely
|
|
422
|
+
necessary. A swappiness value of 100 causes pages to be swapped very
|
|
423
|
+
aggressively. Accepted values are whole numbers between 0 and 100.
|
|
424
|
+
use_tmpfs : bool, default False
|
|
425
|
+
This enables an explicit tmpfs mount for this step. Note that tmpfs is
|
|
426
|
+
not available on Fargate compute environments
|
|
427
|
+
tmpfs_tempdir : bool, default True
|
|
428
|
+
sets METAFLOW_TEMPDIR to tmpfs_path if set for this step.
|
|
429
|
+
tmpfs_size : int, optional, default None
|
|
430
|
+
The value for the size (in MiB) of the tmpfs mount for this step.
|
|
431
|
+
This parameter maps to the `--tmpfs` option in Docker. Defaults to 50% of the
|
|
432
|
+
memory allocated for this step.
|
|
433
|
+
tmpfs_path : str, optional, default None
|
|
434
|
+
Path to tmpfs mount for this step. Defaults to /metaflow_temp.
|
|
435
|
+
inferentia : int, default 0
|
|
436
|
+
Number of Inferentia chips required for this step.
|
|
437
|
+
trainium : int, default None
|
|
438
|
+
Alias for inferentia. Use only one of the two.
|
|
439
|
+
efa : int, default 0
|
|
440
|
+
Number of elastic fabric adapter network devices to attach to container
|
|
441
|
+
ephemeral_storage : int, default None
|
|
442
|
+
The total amount, in GiB, of ephemeral storage to set for the task, 21-200GiB.
|
|
443
|
+
This is only relevant for Fargate compute environments
|
|
444
|
+
log_driver: str, optional, default None
|
|
445
|
+
The log driver to use for the Amazon ECS container.
|
|
446
|
+
log_options: List[str], optional, default None
|
|
447
|
+
List of strings containing options for the chosen log driver. The configurable values
|
|
448
|
+
depend on the `log driver` chosen. Validation of these options is not supported yet.
|
|
449
|
+
Example: [`awslogs-group:aws/batch/job`]
|
|
369
450
|
"""
|
|
370
451
|
...
|
|
371
452
|
|
|
372
453
|
@typing.overload
|
|
373
|
-
def
|
|
454
|
+
def batch(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
455
|
+
...
|
|
456
|
+
|
|
457
|
+
@typing.overload
|
|
458
|
+
def batch(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
459
|
+
...
|
|
460
|
+
|
|
461
|
+
def batch(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, cpu: int = 1, gpu: int = 0, memory: int = 4096, image: typing.Optional[str] = None, queue: str = 'METAFLOW_BATCH_JOB_QUEUE', iam_role: str = 'METAFLOW_ECS_S3_ACCESS_IAM_ROLE', execution_role: str = 'METAFLOW_ECS_FARGATE_EXECUTION_ROLE', shared_memory: typing.Optional[int] = None, max_swap: typing.Optional[int] = None, swappiness: typing.Optional[int] = None, use_tmpfs: bool = False, tmpfs_tempdir: bool = True, tmpfs_size: typing.Optional[int] = None, tmpfs_path: typing.Optional[str] = None, inferentia: int = 0, trainium: int = None, efa: int = 0, ephemeral_storage: int = None, log_driver: typing.Optional[str] = None, log_options: typing.Optional[typing.List[str]] = None):
|
|
374
462
|
"""
|
|
375
|
-
Specifies
|
|
376
|
-
|
|
377
|
-
Use `@resources` to specify the resource requirements
|
|
378
|
-
independently of the specific compute layer (`@batch`, `@kubernetes`).
|
|
379
|
-
|
|
380
|
-
You can choose the compute layer on the command line by executing e.g.
|
|
381
|
-
```
|
|
382
|
-
python myflow.py run --with batch
|
|
383
|
-
```
|
|
384
|
-
or
|
|
385
|
-
```
|
|
386
|
-
python myflow.py run --with kubernetes
|
|
387
|
-
```
|
|
388
|
-
which executes the flow on the desired system using the
|
|
389
|
-
requirements specified in `@resources`.
|
|
463
|
+
Specifies that this step should execute on [AWS Batch](https://aws.amazon.com/batch/).
|
|
390
464
|
|
|
391
465
|
|
|
392
466
|
Parameters
|
|
393
467
|
----------
|
|
394
468
|
cpu : int, default 1
|
|
395
|
-
Number of CPUs required for this step.
|
|
396
|
-
|
|
397
|
-
|
|
398
|
-
|
|
399
|
-
|
|
469
|
+
Number of CPUs required for this step. If `@resources` is
|
|
470
|
+
also present, the maximum value from all decorators is used.
|
|
471
|
+
gpu : int, default 0
|
|
472
|
+
Number of GPUs required for this step. If `@resources` is
|
|
473
|
+
also present, the maximum value from all decorators is used.
|
|
400
474
|
memory : int, default 4096
|
|
401
|
-
Memory size (in MB) required for this step.
|
|
475
|
+
Memory size (in MB) required for this step. If
|
|
476
|
+
`@resources` is also present, the maximum value from all decorators is
|
|
477
|
+
used.
|
|
478
|
+
image : str, optional, default None
|
|
479
|
+
Docker image to use when launching on AWS Batch. If not specified, and
|
|
480
|
+
METAFLOW_BATCH_CONTAINER_IMAGE is specified, that image is used. If
|
|
481
|
+
not, a default Docker image mapping to the current version of Python is used.
|
|
482
|
+
queue : str, default METAFLOW_BATCH_JOB_QUEUE
|
|
483
|
+
AWS Batch Job Queue to submit the job to.
|
|
484
|
+
iam_role : str, default METAFLOW_ECS_S3_ACCESS_IAM_ROLE
|
|
485
|
+
AWS IAM role that AWS Batch container uses to access AWS cloud resources.
|
|
486
|
+
execution_role : str, default METAFLOW_ECS_FARGATE_EXECUTION_ROLE
|
|
487
|
+
AWS IAM role that AWS Batch can use [to trigger AWS Fargate tasks]
|
|
488
|
+
(https://docs.aws.amazon.com/batch/latest/userguide/execution-IAM-role.html).
|
|
402
489
|
shared_memory : int, optional, default None
|
|
403
490
|
The value for the size (in MiB) of the /dev/shm volume for this step.
|
|
404
491
|
This parameter maps to the `--shm-size` option in Docker.
|
|
492
|
+
max_swap : int, optional, default None
|
|
493
|
+
The total amount of swap memory (in MiB) a container can use for this
|
|
494
|
+
step. This parameter is translated to the `--memory-swap` option in
|
|
495
|
+
Docker where the value is the sum of the container memory plus the
|
|
496
|
+
`max_swap` value.
|
|
497
|
+
swappiness : int, optional, default None
|
|
498
|
+
This allows you to tune memory swappiness behavior for this step.
|
|
499
|
+
A swappiness value of 0 causes swapping not to happen unless absolutely
|
|
500
|
+
necessary. A swappiness value of 100 causes pages to be swapped very
|
|
501
|
+
aggressively. Accepted values are whole numbers between 0 and 100.
|
|
502
|
+
use_tmpfs : bool, default False
|
|
503
|
+
This enables an explicit tmpfs mount for this step. Note that tmpfs is
|
|
504
|
+
not available on Fargate compute environments
|
|
505
|
+
tmpfs_tempdir : bool, default True
|
|
506
|
+
sets METAFLOW_TEMPDIR to tmpfs_path if set for this step.
|
|
507
|
+
tmpfs_size : int, optional, default None
|
|
508
|
+
The value for the size (in MiB) of the tmpfs mount for this step.
|
|
509
|
+
This parameter maps to the `--tmpfs` option in Docker. Defaults to 50% of the
|
|
510
|
+
memory allocated for this step.
|
|
511
|
+
tmpfs_path : str, optional, default None
|
|
512
|
+
Path to tmpfs mount for this step. Defaults to /metaflow_temp.
|
|
513
|
+
inferentia : int, default 0
|
|
514
|
+
Number of Inferentia chips required for this step.
|
|
515
|
+
trainium : int, default None
|
|
516
|
+
Alias for inferentia. Use only one of the two.
|
|
517
|
+
efa : int, default 0
|
|
518
|
+
Number of elastic fabric adapter network devices to attach to container
|
|
519
|
+
ephemeral_storage : int, default None
|
|
520
|
+
The total amount, in GiB, of ephemeral storage to set for the task, 21-200GiB.
|
|
521
|
+
This is only relevant for Fargate compute environments
|
|
522
|
+
log_driver: str, optional, default None
|
|
523
|
+
The log driver to use for the Amazon ECS container.
|
|
524
|
+
log_options: List[str], optional, default None
|
|
525
|
+
List of strings containing options for the chosen log driver. The configurable values
|
|
526
|
+
depend on the `log driver` chosen. Validation of these options is not supported yet.
|
|
527
|
+
Example: [`awslogs-group:aws/batch/job`]
|
|
405
528
|
"""
|
|
406
529
|
...
|
|
407
530
|
|
|
408
531
|
@typing.overload
|
|
409
|
-
def
|
|
532
|
+
def card(*, type: str = 'default', id: typing.Optional[str] = None, options: typing.Dict[str, typing.Any] = {}, timeout: int = 45) -> typing.Callable[[typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]], typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]]:
|
|
533
|
+
"""
|
|
534
|
+
Creates a human-readable report, a Metaflow Card, after this step completes.
|
|
535
|
+
|
|
536
|
+
Note that you may add multiple `@card` decorators in a step with different parameters.
|
|
537
|
+
|
|
538
|
+
|
|
539
|
+
Parameters
|
|
540
|
+
----------
|
|
541
|
+
type : str, default 'default'
|
|
542
|
+
Card type.
|
|
543
|
+
id : str, optional, default None
|
|
544
|
+
If multiple cards are present, use this id to identify this card.
|
|
545
|
+
options : Dict[str, Any], default {}
|
|
546
|
+
Options passed to the card. The contents depend on the card type.
|
|
547
|
+
timeout : int, default 45
|
|
548
|
+
Interrupt reporting if it takes more than this many seconds.
|
|
549
|
+
"""
|
|
410
550
|
...
|
|
411
551
|
|
|
412
552
|
@typing.overload
|
|
413
|
-
def
|
|
553
|
+
def card(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
414
554
|
...
|
|
415
555
|
|
|
416
|
-
|
|
556
|
+
@typing.overload
|
|
557
|
+
def card(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
558
|
+
...
|
|
559
|
+
|
|
560
|
+
def card(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, type: str = 'default', id: typing.Optional[str] = None, options: typing.Dict[str, typing.Any] = {}, timeout: int = 45):
|
|
417
561
|
"""
|
|
418
|
-
|
|
419
|
-
|
|
420
|
-
Use `@resources` to specify the resource requirements
|
|
421
|
-
independently of the specific compute layer (`@batch`, `@kubernetes`).
|
|
562
|
+
Creates a human-readable report, a Metaflow Card, after this step completes.
|
|
422
563
|
|
|
423
|
-
|
|
424
|
-
```
|
|
425
|
-
python myflow.py run --with batch
|
|
426
|
-
```
|
|
427
|
-
or
|
|
428
|
-
```
|
|
429
|
-
python myflow.py run --with kubernetes
|
|
430
|
-
```
|
|
431
|
-
which executes the flow on the desired system using the
|
|
432
|
-
requirements specified in `@resources`.
|
|
564
|
+
Note that you may add multiple `@card` decorators in a step with different parameters.
|
|
433
565
|
|
|
434
566
|
|
|
435
567
|
Parameters
|
|
436
568
|
----------
|
|
437
|
-
|
|
438
|
-
|
|
439
|
-
|
|
440
|
-
|
|
441
|
-
|
|
442
|
-
|
|
443
|
-
|
|
444
|
-
|
|
445
|
-
shared_memory : int, optional, default None
|
|
446
|
-
The value for the size (in MiB) of the /dev/shm volume for this step.
|
|
447
|
-
This parameter maps to the `--shm-size` option in Docker.
|
|
569
|
+
type : str, default 'default'
|
|
570
|
+
Card type.
|
|
571
|
+
id : str, optional, default None
|
|
572
|
+
If multiple cards are present, use this id to identify this card.
|
|
573
|
+
options : Dict[str, Any], default {}
|
|
574
|
+
Options passed to the card. The contents depend on the card type.
|
|
575
|
+
timeout : int, default 45
|
|
576
|
+
Interrupt reporting if it takes more than this many seconds.
|
|
448
577
|
"""
|
|
449
578
|
...
|
|
450
579
|
|
|
@@ -508,41 +637,132 @@ def conda(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], ty
|
|
|
508
637
|
...
|
|
509
638
|
|
|
510
639
|
@typing.overload
|
|
511
|
-
def
|
|
640
|
+
def parallel(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
512
641
|
"""
|
|
513
|
-
|
|
514
|
-
|
|
515
|
-
|
|
516
|
-
|
|
517
|
-
|
|
518
|
-
|
|
519
|
-
|
|
520
|
-
|
|
521
|
-
|
|
522
|
-
|
|
642
|
+
Decorator prototype for all step decorators. This function gets specialized
|
|
643
|
+
and imported for all decorators types by _import_plugin_decorators().
|
|
644
|
+
"""
|
|
645
|
+
...
|
|
646
|
+
|
|
647
|
+
@typing.overload
|
|
648
|
+
def parallel(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
649
|
+
...
|
|
650
|
+
|
|
651
|
+
def parallel(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None):
|
|
652
|
+
"""
|
|
653
|
+
Decorator prototype for all step decorators. This function gets specialized
|
|
654
|
+
and imported for all decorators types by _import_plugin_decorators().
|
|
655
|
+
"""
|
|
656
|
+
...
|
|
657
|
+
|
|
658
|
+
@typing.overload
|
|
659
|
+
def environment(*, vars: typing.Dict[str, str] = {}) -> typing.Callable[[typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]], typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]]:
|
|
660
|
+
"""
|
|
661
|
+
Specifies environment variables to be set prior to the execution of a step.
|
|
523
662
|
|
|
524
663
|
|
|
525
664
|
Parameters
|
|
526
665
|
----------
|
|
527
|
-
|
|
528
|
-
|
|
529
|
-
minutes_between_retries : int, default 2
|
|
530
|
-
Number of minutes between retries.
|
|
666
|
+
vars : Dict[str, str], default {}
|
|
667
|
+
Dictionary of environment variables to set.
|
|
531
668
|
"""
|
|
532
669
|
...
|
|
533
670
|
|
|
534
671
|
@typing.overload
|
|
535
|
-
def
|
|
672
|
+
def environment(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
536
673
|
...
|
|
537
674
|
|
|
538
675
|
@typing.overload
|
|
539
|
-
def
|
|
676
|
+
def environment(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
540
677
|
...
|
|
541
678
|
|
|
542
|
-
def
|
|
679
|
+
def environment(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, vars: typing.Dict[str, str] = {}):
|
|
543
680
|
"""
|
|
544
|
-
Specifies
|
|
545
|
-
|
|
681
|
+
Specifies environment variables to be set prior to the execution of a step.
|
|
682
|
+
|
|
683
|
+
|
|
684
|
+
Parameters
|
|
685
|
+
----------
|
|
686
|
+
vars : Dict[str, str], default {}
|
|
687
|
+
Dictionary of environment variables to set.
|
|
688
|
+
"""
|
|
689
|
+
...
|
|
690
|
+
|
|
691
|
+
@typing.overload
|
|
692
|
+
def secrets(*, sources: typing.List[typing.Union[str, typing.Dict[str, typing.Any]]] = [], role: typing.Optional[str] = None) -> typing.Callable[[typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]], typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]]:
|
|
693
|
+
"""
|
|
694
|
+
Specifies secrets to be retrieved and injected as environment variables prior to
|
|
695
|
+
the execution of a step.
|
|
696
|
+
|
|
697
|
+
|
|
698
|
+
Parameters
|
|
699
|
+
----------
|
|
700
|
+
sources : List[Union[str, Dict[str, Any]]], default: []
|
|
701
|
+
List of secret specs, defining how the secrets are to be retrieved
|
|
702
|
+
role : str, optional, default: None
|
|
703
|
+
Role to use for fetching secrets
|
|
704
|
+
"""
|
|
705
|
+
...
|
|
706
|
+
|
|
707
|
+
@typing.overload
|
|
708
|
+
def secrets(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
709
|
+
...
|
|
710
|
+
|
|
711
|
+
@typing.overload
|
|
712
|
+
def secrets(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
713
|
+
...
|
|
714
|
+
|
|
715
|
+
def secrets(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, sources: typing.List[typing.Union[str, typing.Dict[str, typing.Any]]] = [], role: typing.Optional[str] = None):
|
|
716
|
+
"""
|
|
717
|
+
Specifies secrets to be retrieved and injected as environment variables prior to
|
|
718
|
+
the execution of a step.
|
|
719
|
+
|
|
720
|
+
|
|
721
|
+
Parameters
|
|
722
|
+
----------
|
|
723
|
+
sources : List[Union[str, Dict[str, Any]]], default: []
|
|
724
|
+
List of secret specs, defining how the secrets are to be retrieved
|
|
725
|
+
role : str, optional, default: None
|
|
726
|
+
Role to use for fetching secrets
|
|
727
|
+
"""
|
|
728
|
+
...
|
|
729
|
+
|
|
730
|
+
@typing.overload
|
|
731
|
+
def retry(*, times: int = 3, minutes_between_retries: int = 2) -> typing.Callable[[typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]], typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]]:
|
|
732
|
+
"""
|
|
733
|
+
Specifies the number of times the task corresponding
|
|
734
|
+
to a step needs to be retried.
|
|
735
|
+
|
|
736
|
+
This decorator is useful for handling transient errors, such as networking issues.
|
|
737
|
+
If your task contains operations that can't be retried safely, e.g. database updates,
|
|
738
|
+
it is advisable to annotate it with `@retry(times=0)`.
|
|
739
|
+
|
|
740
|
+
This can be used in conjunction with the `@catch` decorator. The `@catch`
|
|
741
|
+
decorator will execute a no-op task after all retries have been exhausted,
|
|
742
|
+
ensuring that the flow execution can continue.
|
|
743
|
+
|
|
744
|
+
|
|
745
|
+
Parameters
|
|
746
|
+
----------
|
|
747
|
+
times : int, default 3
|
|
748
|
+
Number of times to retry this task.
|
|
749
|
+
minutes_between_retries : int, default 2
|
|
750
|
+
Number of minutes between retries.
|
|
751
|
+
"""
|
|
752
|
+
...
|
|
753
|
+
|
|
754
|
+
@typing.overload
|
|
755
|
+
def retry(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
756
|
+
...
|
|
757
|
+
|
|
758
|
+
@typing.overload
|
|
759
|
+
def retry(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
760
|
+
...
|
|
761
|
+
|
|
762
|
+
def retry(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, times: int = 3, minutes_between_retries: int = 2):
|
|
763
|
+
"""
|
|
764
|
+
Specifies the number of times the task corresponding
|
|
765
|
+
to a step needs to be retried.
|
|
546
766
|
|
|
547
767
|
This decorator is useful for handling transient errors, such as networking issues.
|
|
548
768
|
If your task contains operations that can't be retried safely, e.g. database updates,
|
|
@@ -614,528 +834,88 @@ def pypi(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typ
|
|
|
614
834
|
...
|
|
615
835
|
|
|
616
836
|
@typing.overload
|
|
617
|
-
def
|
|
837
|
+
def catch(*, var: typing.Optional[str] = None, print_exception: bool = True) -> typing.Callable[[typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]], typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]]:
|
|
618
838
|
"""
|
|
619
|
-
Specifies
|
|
839
|
+
Specifies that the step will success under all circumstances.
|
|
840
|
+
|
|
841
|
+
The decorator will create an optional artifact, specified by `var`, which
|
|
842
|
+
contains the exception raised. You can use it to detect the presence
|
|
843
|
+
of errors, indicating that all happy-path artifacts produced by the step
|
|
844
|
+
are missing.
|
|
620
845
|
|
|
621
846
|
|
|
622
847
|
Parameters
|
|
623
848
|
----------
|
|
624
|
-
|
|
625
|
-
|
|
849
|
+
var : str, optional, default None
|
|
850
|
+
Name of the artifact in which to store the caught exception.
|
|
851
|
+
If not specified, the exception is not stored.
|
|
852
|
+
print_exception : bool, default True
|
|
853
|
+
Determines whether or not the exception is printed to
|
|
854
|
+
stdout when caught.
|
|
626
855
|
"""
|
|
627
856
|
...
|
|
628
857
|
|
|
629
858
|
@typing.overload
|
|
630
|
-
def
|
|
859
|
+
def catch(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
631
860
|
...
|
|
632
861
|
|
|
633
862
|
@typing.overload
|
|
634
|
-
def
|
|
635
|
-
...
|
|
636
|
-
|
|
637
|
-
def environment(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, vars: typing.Dict[str, str] = {}):
|
|
638
|
-
"""
|
|
639
|
-
Specifies environment variables to be set prior to the execution of a step.
|
|
640
|
-
|
|
641
|
-
|
|
642
|
-
Parameters
|
|
643
|
-
----------
|
|
644
|
-
vars : Dict[str, str], default {}
|
|
645
|
-
Dictionary of environment variables to set.
|
|
646
|
-
"""
|
|
863
|
+
def catch(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
647
864
|
...
|
|
648
865
|
|
|
649
|
-
def
|
|
866
|
+
def catch(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, var: typing.Optional[str] = None, print_exception: bool = True):
|
|
650
867
|
"""
|
|
651
|
-
Specifies that
|
|
652
|
-
|
|
653
|
-
|
|
654
|
-
Parameters
|
|
655
|
-
----------
|
|
656
|
-
cpu : int, default 1
|
|
657
|
-
Number of CPUs required for this step. If `@resources` is
|
|
658
|
-
also present, the maximum value from all decorators is used.
|
|
659
|
-
memory : int, default 4096
|
|
660
|
-
Memory size (in MB) required for this step. If
|
|
661
|
-
`@resources` is also present, the maximum value from all decorators is
|
|
662
|
-
used.
|
|
663
|
-
disk : int, default 10240
|
|
664
|
-
Disk size (in MB) required for this step. If
|
|
665
|
-
`@resources` is also present, the maximum value from all decorators is
|
|
666
|
-
used.
|
|
667
|
-
image : str, optional, default None
|
|
668
|
-
Docker image to use when launching on Kubernetes. If not specified, and
|
|
669
|
-
METAFLOW_KUBERNETES_CONTAINER_IMAGE is specified, that image is used. If
|
|
670
|
-
not, a default Docker image mapping to the current version of Python is used.
|
|
671
|
-
image_pull_policy: str, default KUBERNETES_IMAGE_PULL_POLICY
|
|
672
|
-
If given, the imagePullPolicy to be applied to the Docker image of the step.
|
|
673
|
-
image_pull_secrets: List[str], default []
|
|
674
|
-
The default is extracted from METAFLOW_KUBERNETES_IMAGE_PULL_SECRETS.
|
|
675
|
-
Kubernetes image pull secrets to use when pulling container images
|
|
676
|
-
in Kubernetes.
|
|
677
|
-
service_account : str, default METAFLOW_KUBERNETES_SERVICE_ACCOUNT
|
|
678
|
-
Kubernetes service account to use when launching pod in Kubernetes.
|
|
679
|
-
secrets : List[str], optional, default None
|
|
680
|
-
Kubernetes secrets to use when launching pod in Kubernetes. These
|
|
681
|
-
secrets are in addition to the ones defined in `METAFLOW_KUBERNETES_SECRETS`
|
|
682
|
-
in Metaflow configuration.
|
|
683
|
-
node_selector: Union[Dict[str,str], str], optional, default None
|
|
684
|
-
Kubernetes node selector(s) to apply to the pod running the task.
|
|
685
|
-
Can be passed in as a comma separated string of values e.g.
|
|
686
|
-
'kubernetes.io/os=linux,kubernetes.io/arch=amd64' or as a dictionary
|
|
687
|
-
{'kubernetes.io/os': 'linux', 'kubernetes.io/arch': 'amd64'}
|
|
688
|
-
namespace : str, default METAFLOW_KUBERNETES_NAMESPACE
|
|
689
|
-
Kubernetes namespace to use when launching pod in Kubernetes.
|
|
690
|
-
gpu : int, optional, default None
|
|
691
|
-
Number of GPUs required for this step. A value of zero implies that
|
|
692
|
-
the scheduled node should not have GPUs.
|
|
693
|
-
gpu_vendor : str, default KUBERNETES_GPU_VENDOR
|
|
694
|
-
The vendor of the GPUs to be used for this step.
|
|
695
|
-
tolerations : List[Dict[str,str]], default []
|
|
696
|
-
The default is extracted from METAFLOW_KUBERNETES_TOLERATIONS.
|
|
697
|
-
Kubernetes tolerations to use when launching pod in Kubernetes.
|
|
698
|
-
labels: Dict[str, str], default: METAFLOW_KUBERNETES_LABELS
|
|
699
|
-
Kubernetes labels to use when launching pod in Kubernetes.
|
|
700
|
-
annotations: Dict[str, str], default: METAFLOW_KUBERNETES_ANNOTATIONS
|
|
701
|
-
Kubernetes annotations to use when launching pod in Kubernetes.
|
|
702
|
-
use_tmpfs : bool, default False
|
|
703
|
-
This enables an explicit tmpfs mount for this step.
|
|
704
|
-
tmpfs_tempdir : bool, default True
|
|
705
|
-
sets METAFLOW_TEMPDIR to tmpfs_path if set for this step.
|
|
706
|
-
tmpfs_size : int, optional, default: None
|
|
707
|
-
The value for the size (in MiB) of the tmpfs mount for this step.
|
|
708
|
-
This parameter maps to the `--tmpfs` option in Docker. Defaults to 50% of the
|
|
709
|
-
memory allocated for this step.
|
|
710
|
-
tmpfs_path : str, optional, default /metaflow_temp
|
|
711
|
-
Path to tmpfs mount for this step.
|
|
712
|
-
persistent_volume_claims : Dict[str, str], optional, default None
|
|
713
|
-
A map (dictionary) of persistent volumes to be mounted to the pod for this step. The map is from persistent
|
|
714
|
-
volumes to the path to which the volume is to be mounted, e.g., `{'pvc-name': '/path/to/mount/on'}`.
|
|
715
|
-
shared_memory: int, optional
|
|
716
|
-
Shared memory size (in MiB) required for this step
|
|
717
|
-
port: int, optional
|
|
718
|
-
Port number to specify in the Kubernetes job object
|
|
719
|
-
compute_pool : str, optional, default None
|
|
720
|
-
Compute pool to be used for for this step.
|
|
721
|
-
If not specified, any accessible compute pool within the perimeter is used.
|
|
722
|
-
hostname_resolution_timeout: int, default 10 * 60
|
|
723
|
-
Timeout in seconds for the workers tasks in the gang scheduled cluster to resolve the hostname of control task.
|
|
724
|
-
Only applicable when @parallel is used.
|
|
725
|
-
qos: str, default: Burstable
|
|
726
|
-
Quality of Service class to assign to the pod. Supported values are: Guaranteed, Burstable, BestEffort
|
|
868
|
+
Specifies that the step will success under all circumstances.
|
|
727
869
|
|
|
728
|
-
|
|
729
|
-
|
|
730
|
-
|
|
731
|
-
|
|
732
|
-
- run_as_user: int, optional, default None
|
|
733
|
-
- run_as_group: int, optional, default None
|
|
734
|
-
- run_as_non_root: bool, optional, default None
|
|
735
|
-
"""
|
|
736
|
-
...
|
|
737
|
-
|
|
738
|
-
@typing.overload
|
|
739
|
-
def batch(*, cpu: int = 1, gpu: int = 0, memory: int = 4096, image: typing.Optional[str] = None, queue: str = 'METAFLOW_BATCH_JOB_QUEUE', iam_role: str = 'METAFLOW_ECS_S3_ACCESS_IAM_ROLE', execution_role: str = 'METAFLOW_ECS_FARGATE_EXECUTION_ROLE', shared_memory: typing.Optional[int] = None, max_swap: typing.Optional[int] = None, swappiness: typing.Optional[int] = None, use_tmpfs: bool = False, tmpfs_tempdir: bool = True, tmpfs_size: typing.Optional[int] = None, tmpfs_path: typing.Optional[str] = None, inferentia: int = 0, trainium: int = None, efa: int = 0, ephemeral_storage: int = None, log_driver: typing.Optional[str] = None, log_options: typing.Optional[typing.List[str]] = None) -> typing.Callable[[typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]], typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]]]:
|
|
740
|
-
"""
|
|
741
|
-
Specifies that this step should execute on [AWS Batch](https://aws.amazon.com/batch/).
|
|
870
|
+
The decorator will create an optional artifact, specified by `var`, which
|
|
871
|
+
contains the exception raised. You can use it to detect the presence
|
|
872
|
+
of errors, indicating that all happy-path artifacts produced by the step
|
|
873
|
+
are missing.
|
|
742
874
|
|
|
743
875
|
|
|
744
876
|
Parameters
|
|
745
877
|
----------
|
|
746
|
-
|
|
747
|
-
|
|
748
|
-
|
|
749
|
-
|
|
750
|
-
|
|
751
|
-
|
|
752
|
-
|
|
753
|
-
|
|
754
|
-
|
|
755
|
-
|
|
756
|
-
|
|
757
|
-
|
|
758
|
-
|
|
759
|
-
|
|
760
|
-
|
|
761
|
-
|
|
762
|
-
|
|
763
|
-
|
|
764
|
-
|
|
765
|
-
|
|
766
|
-
|
|
767
|
-
|
|
768
|
-
|
|
769
|
-
|
|
770
|
-
|
|
771
|
-
The
|
|
772
|
-
|
|
773
|
-
|
|
774
|
-
`
|
|
775
|
-
|
|
776
|
-
|
|
777
|
-
|
|
778
|
-
|
|
779
|
-
|
|
780
|
-
|
|
781
|
-
|
|
782
|
-
|
|
783
|
-
|
|
784
|
-
|
|
785
|
-
|
|
786
|
-
|
|
787
|
-
This parameter maps to the `--tmpfs` option in Docker. Defaults to 50% of the
|
|
788
|
-
memory allocated for this step.
|
|
789
|
-
tmpfs_path : str, optional, default None
|
|
790
|
-
Path to tmpfs mount for this step. Defaults to /metaflow_temp.
|
|
791
|
-
inferentia : int, default 0
|
|
792
|
-
Number of Inferentia chips required for this step.
|
|
793
|
-
trainium : int, default None
|
|
794
|
-
Alias for inferentia. Use only one of the two.
|
|
795
|
-
efa : int, default 0
|
|
796
|
-
Number of elastic fabric adapter network devices to attach to container
|
|
797
|
-
ephemeral_storage : int, default None
|
|
798
|
-
The total amount, in GiB, of ephemeral storage to set for the task, 21-200GiB.
|
|
799
|
-
This is only relevant for Fargate compute environments
|
|
800
|
-
log_driver: str, optional, default None
|
|
801
|
-
The log driver to use for the Amazon ECS container.
|
|
802
|
-
log_options: List[str], optional, default None
|
|
803
|
-
List of strings containing options for the chosen log driver. The configurable values
|
|
804
|
-
depend on the `log driver` chosen. Validation of these options is not supported yet.
|
|
805
|
-
Example: [`awslogs-group:aws/batch/job`]
|
|
806
|
-
"""
|
|
807
|
-
...
|
|
808
|
-
|
|
809
|
-
@typing.overload
|
|
810
|
-
def batch(f: typing.Callable[[FlowSpecDerived, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, StepFlag], None]:
|
|
811
|
-
...
|
|
812
|
-
|
|
813
|
-
@typing.overload
|
|
814
|
-
def batch(f: typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]) -> typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None]:
|
|
815
|
-
...
|
|
816
|
-
|
|
817
|
-
def batch(f: typing.Union[typing.Callable[[FlowSpecDerived, StepFlag], None], typing.Callable[[FlowSpecDerived, typing.Any, StepFlag], None], None] = None, *, cpu: int = 1, gpu: int = 0, memory: int = 4096, image: typing.Optional[str] = None, queue: str = 'METAFLOW_BATCH_JOB_QUEUE', iam_role: str = 'METAFLOW_ECS_S3_ACCESS_IAM_ROLE', execution_role: str = 'METAFLOW_ECS_FARGATE_EXECUTION_ROLE', shared_memory: typing.Optional[int] = None, max_swap: typing.Optional[int] = None, swappiness: typing.Optional[int] = None, use_tmpfs: bool = False, tmpfs_tempdir: bool = True, tmpfs_size: typing.Optional[int] = None, tmpfs_path: typing.Optional[str] = None, inferentia: int = 0, trainium: int = None, efa: int = 0, ephemeral_storage: int = None, log_driver: typing.Optional[str] = None, log_options: typing.Optional[typing.List[str]] = None):
|
|
818
|
-
"""
|
|
819
|
-
Specifies that this step should execute on [AWS Batch](https://aws.amazon.com/batch/).
|
|
820
|
-
|
|
821
|
-
|
|
822
|
-
Parameters
|
|
823
|
-
----------
|
|
824
|
-
cpu : int, default 1
|
|
825
|
-
Number of CPUs required for this step. If `@resources` is
|
|
826
|
-
also present, the maximum value from all decorators is used.
|
|
827
|
-
gpu : int, default 0
|
|
828
|
-
Number of GPUs required for this step. If `@resources` is
|
|
829
|
-
also present, the maximum value from all decorators is used.
|
|
830
|
-
memory : int, default 4096
|
|
831
|
-
Memory size (in MB) required for this step. If
|
|
832
|
-
`@resources` is also present, the maximum value from all decorators is
|
|
833
|
-
used.
|
|
834
|
-
image : str, optional, default None
|
|
835
|
-
Docker image to use when launching on AWS Batch. If not specified, and
|
|
836
|
-
METAFLOW_BATCH_CONTAINER_IMAGE is specified, that image is used. If
|
|
837
|
-
not, a default Docker image mapping to the current version of Python is used.
|
|
838
|
-
queue : str, default METAFLOW_BATCH_JOB_QUEUE
|
|
839
|
-
AWS Batch Job Queue to submit the job to.
|
|
840
|
-
iam_role : str, default METAFLOW_ECS_S3_ACCESS_IAM_ROLE
|
|
841
|
-
AWS IAM role that AWS Batch container uses to access AWS cloud resources.
|
|
842
|
-
execution_role : str, default METAFLOW_ECS_FARGATE_EXECUTION_ROLE
|
|
843
|
-
AWS IAM role that AWS Batch can use [to trigger AWS Fargate tasks]
|
|
844
|
-
(https://docs.aws.amazon.com/batch/latest/userguide/execution-IAM-role.html).
|
|
845
|
-
shared_memory : int, optional, default None
|
|
846
|
-
The value for the size (in MiB) of the /dev/shm volume for this step.
|
|
847
|
-
This parameter maps to the `--shm-size` option in Docker.
|
|
848
|
-
max_swap : int, optional, default None
|
|
849
|
-
The total amount of swap memory (in MiB) a container can use for this
|
|
850
|
-
step. This parameter is translated to the `--memory-swap` option in
|
|
851
|
-
Docker where the value is the sum of the container memory plus the
|
|
852
|
-
`max_swap` value.
|
|
853
|
-
swappiness : int, optional, default None
|
|
854
|
-
This allows you to tune memory swappiness behavior for this step.
|
|
855
|
-
A swappiness value of 0 causes swapping not to happen unless absolutely
|
|
856
|
-
necessary. A swappiness value of 100 causes pages to be swapped very
|
|
857
|
-
aggressively. Accepted values are whole numbers between 0 and 100.
|
|
858
|
-
use_tmpfs : bool, default False
|
|
859
|
-
This enables an explicit tmpfs mount for this step. Note that tmpfs is
|
|
860
|
-
not available on Fargate compute environments
|
|
861
|
-
tmpfs_tempdir : bool, default True
|
|
862
|
-
sets METAFLOW_TEMPDIR to tmpfs_path if set for this step.
|
|
863
|
-
tmpfs_size : int, optional, default None
|
|
864
|
-
The value for the size (in MiB) of the tmpfs mount for this step.
|
|
865
|
-
This parameter maps to the `--tmpfs` option in Docker. Defaults to 50% of the
|
|
866
|
-
memory allocated for this step.
|
|
867
|
-
tmpfs_path : str, optional, default None
|
|
868
|
-
Path to tmpfs mount for this step. Defaults to /metaflow_temp.
|
|
869
|
-
inferentia : int, default 0
|
|
870
|
-
Number of Inferentia chips required for this step.
|
|
871
|
-
trainium : int, default None
|
|
872
|
-
Alias for inferentia. Use only one of the two.
|
|
873
|
-
efa : int, default 0
|
|
874
|
-
Number of elastic fabric adapter network devices to attach to container
|
|
875
|
-
ephemeral_storage : int, default None
|
|
876
|
-
The total amount, in GiB, of ephemeral storage to set for the task, 21-200GiB.
|
|
877
|
-
This is only relevant for Fargate compute environments
|
|
878
|
-
log_driver: str, optional, default None
|
|
879
|
-
The log driver to use for the Amazon ECS container.
|
|
880
|
-
log_options: List[str], optional, default None
|
|
881
|
-
List of strings containing options for the chosen log driver. The configurable values
|
|
882
|
-
depend on the `log driver` chosen. Validation of these options is not supported yet.
|
|
883
|
-
Example: [`awslogs-group:aws/batch/job`]
|
|
884
|
-
"""
|
|
885
|
-
...
|
|
886
|
-
|
|
887
|
-
@typing.overload
|
|
888
|
-
def trigger(*, event: typing.Union[str, typing.Dict[str, typing.Any], None] = None, events: typing.List[typing.Union[str, typing.Dict[str, typing.Any]]] = [], options: typing.Dict[str, typing.Any] = {}) -> typing.Callable[[typing.Type[FlowSpecDerived]], typing.Type[FlowSpecDerived]]:
|
|
889
|
-
"""
|
|
890
|
-
Specifies the event(s) that this flow depends on.
|
|
891
|
-
|
|
892
|
-
```
|
|
893
|
-
@trigger(event='foo')
|
|
894
|
-
```
|
|
895
|
-
or
|
|
896
|
-
```
|
|
897
|
-
@trigger(events=['foo', 'bar'])
|
|
898
|
-
```
|
|
899
|
-
|
|
900
|
-
Additionally, you can specify the parameter mappings
|
|
901
|
-
to map event payload to Metaflow parameters for the flow.
|
|
902
|
-
```
|
|
903
|
-
@trigger(event={'name':'foo', 'parameters':{'flow_param': 'event_field'}})
|
|
904
|
-
```
|
|
905
|
-
or
|
|
906
|
-
```
|
|
907
|
-
@trigger(events=[{'name':'foo', 'parameters':{'flow_param_1': 'event_field_1'},
|
|
908
|
-
{'name':'bar', 'parameters':{'flow_param_2': 'event_field_2'}])
|
|
909
|
-
```
|
|
910
|
-
|
|
911
|
-
'parameters' can also be a list of strings and tuples like so:
|
|
912
|
-
```
|
|
913
|
-
@trigger(event={'name':'foo', 'parameters':['common_name', ('flow_param', 'event_field')]})
|
|
914
|
-
```
|
|
915
|
-
This is equivalent to:
|
|
916
|
-
```
|
|
917
|
-
@trigger(event={'name':'foo', 'parameters':{'common_name': 'common_name', 'flow_param': 'event_field'}})
|
|
918
|
-
```
|
|
919
|
-
|
|
920
|
-
|
|
921
|
-
Parameters
|
|
922
|
-
----------
|
|
923
|
-
event : Union[str, Dict[str, Any]], optional, default None
|
|
924
|
-
Event dependency for this flow.
|
|
925
|
-
events : List[Union[str, Dict[str, Any]]], default []
|
|
926
|
-
Events dependency for this flow.
|
|
927
|
-
options : Dict[str, Any], default {}
|
|
928
|
-
Backend-specific configuration for tuning eventing behavior.
|
|
929
|
-
"""
|
|
930
|
-
...
|
|
931
|
-
|
|
932
|
-
@typing.overload
|
|
933
|
-
def trigger(f: typing.Type[FlowSpecDerived]) -> typing.Type[FlowSpecDerived]:
|
|
934
|
-
...
|
|
935
|
-
|
|
936
|
-
def trigger(f: typing.Optional[typing.Type[FlowSpecDerived]] = None, *, event: typing.Union[str, typing.Dict[str, typing.Any], None] = None, events: typing.List[typing.Union[str, typing.Dict[str, typing.Any]]] = [], options: typing.Dict[str, typing.Any] = {}):
|
|
937
|
-
"""
|
|
938
|
-
Specifies the event(s) that this flow depends on.
|
|
939
|
-
|
|
940
|
-
```
|
|
941
|
-
@trigger(event='foo')
|
|
942
|
-
```
|
|
943
|
-
or
|
|
944
|
-
```
|
|
945
|
-
@trigger(events=['foo', 'bar'])
|
|
946
|
-
```
|
|
947
|
-
|
|
948
|
-
Additionally, you can specify the parameter mappings
|
|
949
|
-
to map event payload to Metaflow parameters for the flow.
|
|
950
|
-
```
|
|
951
|
-
@trigger(event={'name':'foo', 'parameters':{'flow_param': 'event_field'}})
|
|
952
|
-
```
|
|
953
|
-
or
|
|
954
|
-
```
|
|
955
|
-
@trigger(events=[{'name':'foo', 'parameters':{'flow_param_1': 'event_field_1'},
|
|
956
|
-
{'name':'bar', 'parameters':{'flow_param_2': 'event_field_2'}])
|
|
957
|
-
```
|
|
958
|
-
|
|
959
|
-
'parameters' can also be a list of strings and tuples like so:
|
|
960
|
-
```
|
|
961
|
-
@trigger(event={'name':'foo', 'parameters':['common_name', ('flow_param', 'event_field')]})
|
|
962
|
-
```
|
|
963
|
-
This is equivalent to:
|
|
964
|
-
```
|
|
965
|
-
@trigger(event={'name':'foo', 'parameters':{'common_name': 'common_name', 'flow_param': 'event_field'}})
|
|
966
|
-
```
|
|
967
|
-
|
|
968
|
-
|
|
969
|
-
Parameters
|
|
970
|
-
----------
|
|
971
|
-
event : Union[str, Dict[str, Any]], optional, default None
|
|
972
|
-
Event dependency for this flow.
|
|
973
|
-
events : List[Union[str, Dict[str, Any]]], default []
|
|
974
|
-
Events dependency for this flow.
|
|
975
|
-
options : Dict[str, Any], default {}
|
|
976
|
-
Backend-specific configuration for tuning eventing behavior.
|
|
977
|
-
"""
|
|
978
|
-
...
|
|
979
|
-
|
|
980
|
-
def project(*, name: str, branch: typing.Optional[str] = None, production: bool = False) -> typing.Callable[[typing.Type[FlowSpecDerived]], typing.Type[FlowSpecDerived]]:
|
|
981
|
-
"""
|
|
982
|
-
Specifies what flows belong to the same project.
|
|
983
|
-
|
|
984
|
-
A project-specific namespace is created for all flows that
|
|
985
|
-
use the same `@project(name)`.
|
|
986
|
-
|
|
987
|
-
|
|
988
|
-
Parameters
|
|
989
|
-
----------
|
|
990
|
-
name : str
|
|
991
|
-
Project name. Make sure that the name is unique amongst all
|
|
992
|
-
projects that use the same production scheduler. The name may
|
|
993
|
-
contain only lowercase alphanumeric characters and underscores.
|
|
994
|
-
|
|
995
|
-
branch : Optional[str], default None
|
|
996
|
-
The branch to use. If not specified, the branch is set to
|
|
997
|
-
`user.<username>` unless `production` is set to `True`. This can
|
|
998
|
-
also be set on the command line using `--branch` as a top-level option.
|
|
999
|
-
It is an error to specify `branch` in the decorator and on the command line.
|
|
1000
|
-
|
|
1001
|
-
production : bool, default False
|
|
1002
|
-
Whether or not the branch is the production branch. This can also be set on the
|
|
1003
|
-
command line using `--production` as a top-level option. It is an error to specify
|
|
1004
|
-
`production` in the decorator and on the command line.
|
|
1005
|
-
The project branch name will be:
|
|
1006
|
-
- if `branch` is specified:
|
|
1007
|
-
- if `production` is True: `prod.<branch>`
|
|
1008
|
-
- if `production` is False: `test.<branch>`
|
|
1009
|
-
- if `branch` is not specified:
|
|
1010
|
-
- if `production` is True: `prod`
|
|
1011
|
-
- if `production` is False: `user.<username>`
|
|
1012
|
-
"""
|
|
1013
|
-
...
|
|
1014
|
-
|
|
1015
|
-
def airflow_s3_key_sensor(*, timeout: int, poke_interval: int, mode: str, exponential_backoff: bool, pool: str, soft_fail: bool, name: str, description: str, bucket_key: typing.Union[str, typing.List[str]], bucket_name: str, wildcard_match: bool, aws_conn_id: str, verify: bool) -> typing.Callable[[typing.Type[FlowSpecDerived]], typing.Type[FlowSpecDerived]]:
|
|
1016
|
-
"""
|
|
1017
|
-
The `@airflow_s3_key_sensor` decorator attaches a Airflow [S3KeySensor](https://airflow.apache.org/docs/apache-airflow-providers-amazon/stable/_api/airflow/providers/amazon/aws/sensors/s3/index.html#airflow.providers.amazon.aws.sensors.s3.S3KeySensor)
|
|
1018
|
-
before the start step of the flow. This decorator only works when a flow is scheduled on Airflow
|
|
1019
|
-
and is compiled using `airflow create`. More than one `@airflow_s3_key_sensor` can be
|
|
1020
|
-
added as a flow decorators. Adding more than one decorator will ensure that `start` step
|
|
1021
|
-
starts only after all sensors finish.
|
|
1022
|
-
|
|
1023
|
-
|
|
1024
|
-
Parameters
|
|
1025
|
-
----------
|
|
1026
|
-
timeout : int
|
|
1027
|
-
Time, in seconds before the task times out and fails. (Default: 3600)
|
|
1028
|
-
poke_interval : int
|
|
1029
|
-
Time in seconds that the job should wait in between each try. (Default: 60)
|
|
1030
|
-
mode : str
|
|
1031
|
-
How the sensor operates. Options are: { poke | reschedule }. (Default: "poke")
|
|
1032
|
-
exponential_backoff : bool
|
|
1033
|
-
allow progressive longer waits between pokes by using exponential backoff algorithm. (Default: True)
|
|
1034
|
-
pool : str
|
|
1035
|
-
the slot pool this task should run in,
|
|
1036
|
-
slot pools are a way to limit concurrency for certain tasks. (Default:None)
|
|
1037
|
-
soft_fail : bool
|
|
1038
|
-
Set to true to mark the task as SKIPPED on failure. (Default: False)
|
|
1039
|
-
name : str
|
|
1040
|
-
Name of the sensor on Airflow
|
|
1041
|
-
description : str
|
|
1042
|
-
Description of sensor in the Airflow UI
|
|
1043
|
-
bucket_key : Union[str, List[str]]
|
|
1044
|
-
The key(s) being waited on. Supports full s3:// style url or relative path from root level.
|
|
1045
|
-
When it's specified as a full s3:// url, please leave `bucket_name` as None
|
|
1046
|
-
bucket_name : str
|
|
1047
|
-
Name of the S3 bucket. Only needed when bucket_key is not provided as a full s3:// url.
|
|
1048
|
-
When specified, all the keys passed to bucket_key refers to this bucket. (Default:None)
|
|
1049
|
-
wildcard_match : bool
|
|
1050
|
-
whether the bucket_key should be interpreted as a Unix wildcard pattern. (Default: False)
|
|
1051
|
-
aws_conn_id : str
|
|
1052
|
-
a reference to the s3 connection on Airflow. (Default: None)
|
|
1053
|
-
verify : bool
|
|
1054
|
-
Whether or not to verify SSL certificates for S3 connection. (Default: None)
|
|
1055
|
-
"""
|
|
1056
|
-
...
|
|
1057
|
-
|
|
1058
|
-
@typing.overload
|
|
1059
|
-
def pypi_base(*, packages: typing.Dict[str, str] = {}, python: typing.Optional[str] = None) -> typing.Callable[[typing.Type[FlowSpecDerived]], typing.Type[FlowSpecDerived]]:
|
|
1060
|
-
"""
|
|
1061
|
-
Specifies the PyPI packages for all steps of the flow.
|
|
1062
|
-
|
|
1063
|
-
Use `@pypi_base` to set common packages required by all
|
|
1064
|
-
steps and use `@pypi` to specify step-specific overrides.
|
|
1065
|
-
|
|
1066
|
-
Parameters
|
|
1067
|
-
----------
|
|
1068
|
-
packages : Dict[str, str], default: {}
|
|
1069
|
-
Packages to use for this flow. The key is the name of the package
|
|
1070
|
-
and the value is the version to use.
|
|
1071
|
-
python : str, optional, default: None
|
|
1072
|
-
Version of Python to use, e.g. '3.7.4'. A default value of None implies
|
|
1073
|
-
that the version used will correspond to the version of the Python interpreter used to start the run.
|
|
1074
|
-
"""
|
|
1075
|
-
...
|
|
1076
|
-
|
|
1077
|
-
@typing.overload
|
|
1078
|
-
def pypi_base(f: typing.Type[FlowSpecDerived]) -> typing.Type[FlowSpecDerived]:
|
|
1079
|
-
...
|
|
1080
|
-
|
|
1081
|
-
def pypi_base(f: typing.Optional[typing.Type[FlowSpecDerived]] = None, *, packages: typing.Dict[str, str] = {}, python: typing.Optional[str] = None):
|
|
1082
|
-
"""
|
|
1083
|
-
Specifies the PyPI packages for all steps of the flow.
|
|
1084
|
-
|
|
1085
|
-
Use `@pypi_base` to set common packages required by all
|
|
1086
|
-
steps and use `@pypi` to specify step-specific overrides.
|
|
1087
|
-
|
|
1088
|
-
Parameters
|
|
1089
|
-
----------
|
|
1090
|
-
packages : Dict[str, str], default: {}
|
|
1091
|
-
Packages to use for this flow. The key is the name of the package
|
|
1092
|
-
and the value is the version to use.
|
|
1093
|
-
python : str, optional, default: None
|
|
1094
|
-
Version of Python to use, e.g. '3.7.4'. A default value of None implies
|
|
1095
|
-
that the version used will correspond to the version of the Python interpreter used to start the run.
|
|
1096
|
-
"""
|
|
1097
|
-
...
|
|
1098
|
-
|
|
1099
|
-
def airflow_external_task_sensor(*, timeout: int, poke_interval: int, mode: str, exponential_backoff: bool, pool: str, soft_fail: bool, name: str, description: str, external_dag_id: str, external_task_ids: typing.List[str], allowed_states: typing.List[str], failed_states: typing.List[str], execution_delta: "datetime.timedelta", check_existence: bool) -> typing.Callable[[typing.Type[FlowSpecDerived]], typing.Type[FlowSpecDerived]]:
|
|
1100
|
-
"""
|
|
1101
|
-
The `@airflow_external_task_sensor` decorator attaches a Airflow [ExternalTaskSensor](https://airflow.apache.org/docs/apache-airflow/stable/_api/airflow/sensors/external_task/index.html#airflow.sensors.external_task.ExternalTaskSensor) before the start step of the flow.
|
|
1102
|
-
This decorator only works when a flow is scheduled on Airflow and is compiled using `airflow create`. More than one `@airflow_external_task_sensor` can be added as a flow decorators. Adding more than one decorator will ensure that `start` step starts only after all sensors finish.
|
|
1103
|
-
|
|
1104
|
-
|
|
1105
|
-
Parameters
|
|
1106
|
-
----------
|
|
1107
|
-
timeout : int
|
|
1108
|
-
Time, in seconds before the task times out and fails. (Default: 3600)
|
|
1109
|
-
poke_interval : int
|
|
1110
|
-
Time in seconds that the job should wait in between each try. (Default: 60)
|
|
1111
|
-
mode : str
|
|
1112
|
-
How the sensor operates. Options are: { poke | reschedule }. (Default: "poke")
|
|
1113
|
-
exponential_backoff : bool
|
|
1114
|
-
allow progressive longer waits between pokes by using exponential backoff algorithm. (Default: True)
|
|
1115
|
-
pool : str
|
|
1116
|
-
the slot pool this task should run in,
|
|
1117
|
-
slot pools are a way to limit concurrency for certain tasks. (Default:None)
|
|
1118
|
-
soft_fail : bool
|
|
1119
|
-
Set to true to mark the task as SKIPPED on failure. (Default: False)
|
|
1120
|
-
name : str
|
|
1121
|
-
Name of the sensor on Airflow
|
|
1122
|
-
description : str
|
|
1123
|
-
Description of sensor in the Airflow UI
|
|
1124
|
-
external_dag_id : str
|
|
1125
|
-
The dag_id that contains the task you want to wait for.
|
|
1126
|
-
external_task_ids : List[str]
|
|
1127
|
-
The list of task_ids that you want to wait for.
|
|
1128
|
-
If None (default value) the sensor waits for the DAG. (Default: None)
|
|
1129
|
-
allowed_states : List[str]
|
|
1130
|
-
Iterable of allowed states, (Default: ['success'])
|
|
1131
|
-
failed_states : List[str]
|
|
1132
|
-
Iterable of failed or dis-allowed states. (Default: None)
|
|
1133
|
-
execution_delta : datetime.timedelta
|
|
1134
|
-
time difference with the previous execution to look at,
|
|
1135
|
-
the default is the same logical date as the current task or DAG. (Default: None)
|
|
1136
|
-
check_existence: bool
|
|
1137
|
-
Set to True to check if the external task exists or check if
|
|
1138
|
-
the DAG to wait for exists. (Default: True)
|
|
878
|
+
var : str, optional, default None
|
|
879
|
+
Name of the artifact in which to store the caught exception.
|
|
880
|
+
If not specified, the exception is not stored.
|
|
881
|
+
print_exception : bool, default True
|
|
882
|
+
Determines whether or not the exception is printed to
|
|
883
|
+
stdout when caught.
|
|
884
|
+
"""
|
|
885
|
+
...
|
|
886
|
+
|
|
887
|
+
def project(*, name: str, branch: typing.Optional[str] = None, production: bool = False) -> typing.Callable[[typing.Type[FlowSpecDerived]], typing.Type[FlowSpecDerived]]:
|
|
888
|
+
"""
|
|
889
|
+
Specifies what flows belong to the same project.
|
|
890
|
+
|
|
891
|
+
A project-specific namespace is created for all flows that
|
|
892
|
+
use the same `@project(name)`.
|
|
893
|
+
|
|
894
|
+
|
|
895
|
+
Parameters
|
|
896
|
+
----------
|
|
897
|
+
name : str
|
|
898
|
+
Project name. Make sure that the name is unique amongst all
|
|
899
|
+
projects that use the same production scheduler. The name may
|
|
900
|
+
contain only lowercase alphanumeric characters and underscores.
|
|
901
|
+
|
|
902
|
+
branch : Optional[str], default None
|
|
903
|
+
The branch to use. If not specified, the branch is set to
|
|
904
|
+
`user.<username>` unless `production` is set to `True`. This can
|
|
905
|
+
also be set on the command line using `--branch` as a top-level option.
|
|
906
|
+
It is an error to specify `branch` in the decorator and on the command line.
|
|
907
|
+
|
|
908
|
+
production : bool, default False
|
|
909
|
+
Whether or not the branch is the production branch. This can also be set on the
|
|
910
|
+
command line using `--production` as a top-level option. It is an error to specify
|
|
911
|
+
`production` in the decorator and on the command line.
|
|
912
|
+
The project branch name will be:
|
|
913
|
+
- if `branch` is specified:
|
|
914
|
+
- if `production` is True: `prod.<branch>`
|
|
915
|
+
- if `production` is False: `test.<branch>`
|
|
916
|
+
- if `branch` is not specified:
|
|
917
|
+
- if `production` is True: `prod`
|
|
918
|
+
- if `production` is False: `user.<username>`
|
|
1139
919
|
"""
|
|
1140
920
|
...
|
|
1141
921
|
|
|
@@ -1342,3 +1122,223 @@ def conda_base(f: typing.Optional[typing.Type[FlowSpecDerived]] = None, *, packa
|
|
|
1342
1122
|
"""
|
|
1343
1123
|
...
|
|
1344
1124
|
|
|
1125
|
+
def airflow_s3_key_sensor(*, timeout: int, poke_interval: int, mode: str, exponential_backoff: bool, pool: str, soft_fail: bool, name: str, description: str, bucket_key: typing.Union[str, typing.List[str]], bucket_name: str, wildcard_match: bool, aws_conn_id: str, verify: bool) -> typing.Callable[[typing.Type[FlowSpecDerived]], typing.Type[FlowSpecDerived]]:
|
|
1126
|
+
"""
|
|
1127
|
+
The `@airflow_s3_key_sensor` decorator attaches a Airflow [S3KeySensor](https://airflow.apache.org/docs/apache-airflow-providers-amazon/stable/_api/airflow/providers/amazon/aws/sensors/s3/index.html#airflow.providers.amazon.aws.sensors.s3.S3KeySensor)
|
|
1128
|
+
before the start step of the flow. This decorator only works when a flow is scheduled on Airflow
|
|
1129
|
+
and is compiled using `airflow create`. More than one `@airflow_s3_key_sensor` can be
|
|
1130
|
+
added as a flow decorators. Adding more than one decorator will ensure that `start` step
|
|
1131
|
+
starts only after all sensors finish.
|
|
1132
|
+
|
|
1133
|
+
|
|
1134
|
+
Parameters
|
|
1135
|
+
----------
|
|
1136
|
+
timeout : int
|
|
1137
|
+
Time, in seconds before the task times out and fails. (Default: 3600)
|
|
1138
|
+
poke_interval : int
|
|
1139
|
+
Time in seconds that the job should wait in between each try. (Default: 60)
|
|
1140
|
+
mode : str
|
|
1141
|
+
How the sensor operates. Options are: { poke | reschedule }. (Default: "poke")
|
|
1142
|
+
exponential_backoff : bool
|
|
1143
|
+
allow progressive longer waits between pokes by using exponential backoff algorithm. (Default: True)
|
|
1144
|
+
pool : str
|
|
1145
|
+
the slot pool this task should run in,
|
|
1146
|
+
slot pools are a way to limit concurrency for certain tasks. (Default:None)
|
|
1147
|
+
soft_fail : bool
|
|
1148
|
+
Set to true to mark the task as SKIPPED on failure. (Default: False)
|
|
1149
|
+
name : str
|
|
1150
|
+
Name of the sensor on Airflow
|
|
1151
|
+
description : str
|
|
1152
|
+
Description of sensor in the Airflow UI
|
|
1153
|
+
bucket_key : Union[str, List[str]]
|
|
1154
|
+
The key(s) being waited on. Supports full s3:// style url or relative path from root level.
|
|
1155
|
+
When it's specified as a full s3:// url, please leave `bucket_name` as None
|
|
1156
|
+
bucket_name : str
|
|
1157
|
+
Name of the S3 bucket. Only needed when bucket_key is not provided as a full s3:// url.
|
|
1158
|
+
When specified, all the keys passed to bucket_key refers to this bucket. (Default:None)
|
|
1159
|
+
wildcard_match : bool
|
|
1160
|
+
whether the bucket_key should be interpreted as a Unix wildcard pattern. (Default: False)
|
|
1161
|
+
aws_conn_id : str
|
|
1162
|
+
a reference to the s3 connection on Airflow. (Default: None)
|
|
1163
|
+
verify : bool
|
|
1164
|
+
Whether or not to verify SSL certificates for S3 connection. (Default: None)
|
|
1165
|
+
"""
|
|
1166
|
+
...
|
|
1167
|
+
|
|
1168
|
+
@typing.overload
|
|
1169
|
+
def pypi_base(*, packages: typing.Dict[str, str] = {}, python: typing.Optional[str] = None) -> typing.Callable[[typing.Type[FlowSpecDerived]], typing.Type[FlowSpecDerived]]:
|
|
1170
|
+
"""
|
|
1171
|
+
Specifies the PyPI packages for all steps of the flow.
|
|
1172
|
+
|
|
1173
|
+
Use `@pypi_base` to set common packages required by all
|
|
1174
|
+
steps and use `@pypi` to specify step-specific overrides.
|
|
1175
|
+
|
|
1176
|
+
Parameters
|
|
1177
|
+
----------
|
|
1178
|
+
packages : Dict[str, str], default: {}
|
|
1179
|
+
Packages to use for this flow. The key is the name of the package
|
|
1180
|
+
and the value is the version to use.
|
|
1181
|
+
python : str, optional, default: None
|
|
1182
|
+
Version of Python to use, e.g. '3.7.4'. A default value of None implies
|
|
1183
|
+
that the version used will correspond to the version of the Python interpreter used to start the run.
|
|
1184
|
+
"""
|
|
1185
|
+
...
|
|
1186
|
+
|
|
1187
|
+
@typing.overload
|
|
1188
|
+
def pypi_base(f: typing.Type[FlowSpecDerived]) -> typing.Type[FlowSpecDerived]:
|
|
1189
|
+
...
|
|
1190
|
+
|
|
1191
|
+
def pypi_base(f: typing.Optional[typing.Type[FlowSpecDerived]] = None, *, packages: typing.Dict[str, str] = {}, python: typing.Optional[str] = None):
|
|
1192
|
+
"""
|
|
1193
|
+
Specifies the PyPI packages for all steps of the flow.
|
|
1194
|
+
|
|
1195
|
+
Use `@pypi_base` to set common packages required by all
|
|
1196
|
+
steps and use `@pypi` to specify step-specific overrides.
|
|
1197
|
+
|
|
1198
|
+
Parameters
|
|
1199
|
+
----------
|
|
1200
|
+
packages : Dict[str, str], default: {}
|
|
1201
|
+
Packages to use for this flow. The key is the name of the package
|
|
1202
|
+
and the value is the version to use.
|
|
1203
|
+
python : str, optional, default: None
|
|
1204
|
+
Version of Python to use, e.g. '3.7.4'. A default value of None implies
|
|
1205
|
+
that the version used will correspond to the version of the Python interpreter used to start the run.
|
|
1206
|
+
"""
|
|
1207
|
+
...
|
|
1208
|
+
|
|
1209
|
+
def airflow_external_task_sensor(*, timeout: int, poke_interval: int, mode: str, exponential_backoff: bool, pool: str, soft_fail: bool, name: str, description: str, external_dag_id: str, external_task_ids: typing.List[str], allowed_states: typing.List[str], failed_states: typing.List[str], execution_delta: "datetime.timedelta", check_existence: bool) -> typing.Callable[[typing.Type[FlowSpecDerived]], typing.Type[FlowSpecDerived]]:
|
|
1210
|
+
"""
|
|
1211
|
+
The `@airflow_external_task_sensor` decorator attaches a Airflow [ExternalTaskSensor](https://airflow.apache.org/docs/apache-airflow/stable/_api/airflow/sensors/external_task/index.html#airflow.sensors.external_task.ExternalTaskSensor) before the start step of the flow.
|
|
1212
|
+
This decorator only works when a flow is scheduled on Airflow and is compiled using `airflow create`. More than one `@airflow_external_task_sensor` can be added as a flow decorators. Adding more than one decorator will ensure that `start` step starts only after all sensors finish.
|
|
1213
|
+
|
|
1214
|
+
|
|
1215
|
+
Parameters
|
|
1216
|
+
----------
|
|
1217
|
+
timeout : int
|
|
1218
|
+
Time, in seconds before the task times out and fails. (Default: 3600)
|
|
1219
|
+
poke_interval : int
|
|
1220
|
+
Time in seconds that the job should wait in between each try. (Default: 60)
|
|
1221
|
+
mode : str
|
|
1222
|
+
How the sensor operates. Options are: { poke | reschedule }. (Default: "poke")
|
|
1223
|
+
exponential_backoff : bool
|
|
1224
|
+
allow progressive longer waits between pokes by using exponential backoff algorithm. (Default: True)
|
|
1225
|
+
pool : str
|
|
1226
|
+
the slot pool this task should run in,
|
|
1227
|
+
slot pools are a way to limit concurrency for certain tasks. (Default:None)
|
|
1228
|
+
soft_fail : bool
|
|
1229
|
+
Set to true to mark the task as SKIPPED on failure. (Default: False)
|
|
1230
|
+
name : str
|
|
1231
|
+
Name of the sensor on Airflow
|
|
1232
|
+
description : str
|
|
1233
|
+
Description of sensor in the Airflow UI
|
|
1234
|
+
external_dag_id : str
|
|
1235
|
+
The dag_id that contains the task you want to wait for.
|
|
1236
|
+
external_task_ids : List[str]
|
|
1237
|
+
The list of task_ids that you want to wait for.
|
|
1238
|
+
If None (default value) the sensor waits for the DAG. (Default: None)
|
|
1239
|
+
allowed_states : List[str]
|
|
1240
|
+
Iterable of allowed states, (Default: ['success'])
|
|
1241
|
+
failed_states : List[str]
|
|
1242
|
+
Iterable of failed or dis-allowed states. (Default: None)
|
|
1243
|
+
execution_delta : datetime.timedelta
|
|
1244
|
+
time difference with the previous execution to look at,
|
|
1245
|
+
the default is the same logical date as the current task or DAG. (Default: None)
|
|
1246
|
+
check_existence: bool
|
|
1247
|
+
Set to True to check if the external task exists or check if
|
|
1248
|
+
the DAG to wait for exists. (Default: True)
|
|
1249
|
+
"""
|
|
1250
|
+
...
|
|
1251
|
+
|
|
1252
|
+
@typing.overload
|
|
1253
|
+
def trigger(*, event: typing.Union[str, typing.Dict[str, typing.Any], None] = None, events: typing.List[typing.Union[str, typing.Dict[str, typing.Any]]] = [], options: typing.Dict[str, typing.Any] = {}) -> typing.Callable[[typing.Type[FlowSpecDerived]], typing.Type[FlowSpecDerived]]:
|
|
1254
|
+
"""
|
|
1255
|
+
Specifies the event(s) that this flow depends on.
|
|
1256
|
+
|
|
1257
|
+
```
|
|
1258
|
+
@trigger(event='foo')
|
|
1259
|
+
```
|
|
1260
|
+
or
|
|
1261
|
+
```
|
|
1262
|
+
@trigger(events=['foo', 'bar'])
|
|
1263
|
+
```
|
|
1264
|
+
|
|
1265
|
+
Additionally, you can specify the parameter mappings
|
|
1266
|
+
to map event payload to Metaflow parameters for the flow.
|
|
1267
|
+
```
|
|
1268
|
+
@trigger(event={'name':'foo', 'parameters':{'flow_param': 'event_field'}})
|
|
1269
|
+
```
|
|
1270
|
+
or
|
|
1271
|
+
```
|
|
1272
|
+
@trigger(events=[{'name':'foo', 'parameters':{'flow_param_1': 'event_field_1'},
|
|
1273
|
+
{'name':'bar', 'parameters':{'flow_param_2': 'event_field_2'}])
|
|
1274
|
+
```
|
|
1275
|
+
|
|
1276
|
+
'parameters' can also be a list of strings and tuples like so:
|
|
1277
|
+
```
|
|
1278
|
+
@trigger(event={'name':'foo', 'parameters':['common_name', ('flow_param', 'event_field')]})
|
|
1279
|
+
```
|
|
1280
|
+
This is equivalent to:
|
|
1281
|
+
```
|
|
1282
|
+
@trigger(event={'name':'foo', 'parameters':{'common_name': 'common_name', 'flow_param': 'event_field'}})
|
|
1283
|
+
```
|
|
1284
|
+
|
|
1285
|
+
|
|
1286
|
+
Parameters
|
|
1287
|
+
----------
|
|
1288
|
+
event : Union[str, Dict[str, Any]], optional, default None
|
|
1289
|
+
Event dependency for this flow.
|
|
1290
|
+
events : List[Union[str, Dict[str, Any]]], default []
|
|
1291
|
+
Events dependency for this flow.
|
|
1292
|
+
options : Dict[str, Any], default {}
|
|
1293
|
+
Backend-specific configuration for tuning eventing behavior.
|
|
1294
|
+
"""
|
|
1295
|
+
...
|
|
1296
|
+
|
|
1297
|
+
@typing.overload
|
|
1298
|
+
def trigger(f: typing.Type[FlowSpecDerived]) -> typing.Type[FlowSpecDerived]:
|
|
1299
|
+
...
|
|
1300
|
+
|
|
1301
|
+
def trigger(f: typing.Optional[typing.Type[FlowSpecDerived]] = None, *, event: typing.Union[str, typing.Dict[str, typing.Any], None] = None, events: typing.List[typing.Union[str, typing.Dict[str, typing.Any]]] = [], options: typing.Dict[str, typing.Any] = {}):
|
|
1302
|
+
"""
|
|
1303
|
+
Specifies the event(s) that this flow depends on.
|
|
1304
|
+
|
|
1305
|
+
```
|
|
1306
|
+
@trigger(event='foo')
|
|
1307
|
+
```
|
|
1308
|
+
or
|
|
1309
|
+
```
|
|
1310
|
+
@trigger(events=['foo', 'bar'])
|
|
1311
|
+
```
|
|
1312
|
+
|
|
1313
|
+
Additionally, you can specify the parameter mappings
|
|
1314
|
+
to map event payload to Metaflow parameters for the flow.
|
|
1315
|
+
```
|
|
1316
|
+
@trigger(event={'name':'foo', 'parameters':{'flow_param': 'event_field'}})
|
|
1317
|
+
```
|
|
1318
|
+
or
|
|
1319
|
+
```
|
|
1320
|
+
@trigger(events=[{'name':'foo', 'parameters':{'flow_param_1': 'event_field_1'},
|
|
1321
|
+
{'name':'bar', 'parameters':{'flow_param_2': 'event_field_2'}])
|
|
1322
|
+
```
|
|
1323
|
+
|
|
1324
|
+
'parameters' can also be a list of strings and tuples like so:
|
|
1325
|
+
```
|
|
1326
|
+
@trigger(event={'name':'foo', 'parameters':['common_name', ('flow_param', 'event_field')]})
|
|
1327
|
+
```
|
|
1328
|
+
This is equivalent to:
|
|
1329
|
+
```
|
|
1330
|
+
@trigger(event={'name':'foo', 'parameters':{'common_name': 'common_name', 'flow_param': 'event_field'}})
|
|
1331
|
+
```
|
|
1332
|
+
|
|
1333
|
+
|
|
1334
|
+
Parameters
|
|
1335
|
+
----------
|
|
1336
|
+
event : Union[str, Dict[str, Any]], optional, default None
|
|
1337
|
+
Event dependency for this flow.
|
|
1338
|
+
events : List[Union[str, Dict[str, Any]]], default []
|
|
1339
|
+
Events dependency for this flow.
|
|
1340
|
+
options : Dict[str, Any], default {}
|
|
1341
|
+
Backend-specific configuration for tuning eventing behavior.
|
|
1342
|
+
"""
|
|
1343
|
+
...
|
|
1344
|
+
|