AutomationCondition.in_progress() now will be true if an asset partition is part of an in-progress backfill that has not yet executed it. The prior behavior, which only considered runs, is encapsulated in AutomationCondition.execution_in_progress().build_dbt_asset_selection.container_kwargs.stop_timeout can now be set when using the DockerRunLauncher or docker_executor to configure the amount of time that Docker will wait when terminating a run for it to clean up before forcibly stopping it with a SIGKILL signal.AutomationCondition.eager() that could cause it to attempt to launch a second attempt of an asset in cases where it was skipped or failed during a run where one of its parents successfully materialized.AutomationConditionSensorDefinitions to not be evaluated if the use_user_code_server value was toggled after the initial evaluation.AutomationCondition.eager(), AutomationCondition.missing(), and AutomationCondition.on_cron conditions are now compatible with asset checks.AssetSelection.materializable(), which returns only assets that are materializable in an existing selection.AutomationCondition.all_deps_blocking_checks_passed condition, which can be used to prevent materialization when any upstream blocking checks have failed.code_version parameter to the @graph_asset decorator.LaunchPartitionBackfill mutation is submitted to GQL with invalid partition keys, it will now return an early PartitionKeysNotFoundError.AssetSelection.checks_for_assets now accepts AssetKeys and string asset keys, in addition to AssetsDefinitions.get_objects now has a since_last_modified that enables only fetching objects modified after a given timestamp.dagster_aws.pipes.PipesEMRCLient ) for running and monitoring AWS EMR jobs from Dagster.AutomationConditions that operate over dependencies.DatabricksPySparkStepLauncher, so that permissions are now set correctly for non-admin users.fetch_column_metadata did not work properly for models imported through dbt dependencies.DagsterK8sPipesClient.run now shows up in API docs.create_pull_request method on GithubClient that enables creating a pull request.create_ref method on GithubClient that enables creating a new branch.dlt.run(...).fetch_row_count() for both partitioned and non-partitioned assets. Thanks @kristianandre!PartitionedConfig objects can now return a RunConfig without causing a crash.AssetIn.__new__ typing for the dagster_type argument.fetch_row_count() method now works for databases returning uppercase column names. Thanks @kristianandre!GCSFileManager.read operations. Thanks @ollie-bell!JobDefinition, @job, and define_asset_job now take a run_tags parameter. If run_tags are defined, they will be attached to all runs of the job, and tags will not be. If run_tags is not set, then tags are attached to all runs of the job (status quo behavior). This change enables the separation of definition-level and run-level tags on jobs.DAGSTER_COMPUTE_LOG_TAIL_WAIT_AFTER_FINISH can now be used to pause before capturing logs (thanks @HynekBlaha!)kinds parameter is now available on AssetSpec.OutputContext now exposes the AssetSpec of the asset that is being stored as an output (thanks, @marijncv!)AutomationCondition on an asset, the label of this condition will now be shown in the sidebar on the Asset Details page.dagster definitions validate have been added.AssetSpec now has a with_io_manager_key method that returns an AssetSpec with the appropriate metadata entry to dictate the key for the IO manager used to load it. The deprecation warning for SourceAsset now references this method.max_runtime_seconds configuration option to run monitoring, allowing you to specify that any run in your Dagster deployment should terminate if it exceeds a certain runtime. Prevoiusly, jobs had to be individually tagged with a dagster/max_runtime tag in order to take advantage of this feature. Jobs and runs can still be tagged in order to override this value for an individual run.tags and a custom execution_fn on a ScheduleDefinition. Schedule tags are intended to annotate the definition and can be used to search and filter in the UI. They will not be attached to run requests emitted from the schedule if a custom execution_fn is provided. If no custom execution_fn is provided, then for back-compatibility the tags will also be automatically attached to run requests emitted from the schedule.SensorDefinition and all of its variants/decorators now accept a tags parameter. The tags annotate the definition and can be used to search and filter in the UI.dagster definitions validate command to Dagster CLI. This command validates if Dagster definitions are loadable.build_op_context in a unit test would sometimes raise a TypeError: signal handler must be signal.SIG_IGN, signal.SIG_DFL, or a callable object Exception on process shutdown.DefaultScheduleStatus.STOPPED / DefaultSensorStatus.STOPPED after performing a reset.--partition-range option to dagster asset materialize CLI. This option only works for assets with single-run Backfill Policies..without() method to AutomationCondition.eager(), AutomationCondition.on_cron(), and AutomationCondition.on_missing() which allows sub-conditions to be removed, e.g. AutomationCondition.eager().without(AutomationCondition.in_latest_time_window()).AutomationCondition.on_missing(), which materializes an asset partition as soon as all of its parent partitions are filled in.pyproject.toml can now load multiple Python modules as individual Code Locations. Thanks, @bdart!ADLS2PickleIOManager now accepts lease_duration configuration. Thanks, @0xfabioo!sling.replicate(...).fetch_row_count().AutomationCondition.any_downstream_condition() with downstream AutoMaterializePolicy objects.process_config_and_initialize did not properly handle processing nested resource config.DepConditionWrapperCondition instead of the key that they were evaluated against.wait_for_pod check now waits until all pods are available, rather than erroneously returning after the first pod becomes available. Thanks @easontm!AssetSpec constructor now raises an error if an invalid group name is provided, instead of an error being raised when constructing the Definitions object.dagster/relation_identifier metadata is now automatically attached to assets which are stored using a DbIOManager.DbtCliResource, the project_dir argument will now override the DBT_PROJECT_DIR environment variable if it exists in the local environment (thanks, @marijncv!).rows_loaded metadata (thanks, @kristianandre!).asset_selection=[] on RunRequest objects yielded from sensors using asset_selection would select all assets instead of none.assets_dbt_python example.database field set (thanks, @dargmuesli!)dbt.adapters.duckdb module when loading dbt assets without that package installed.DagsterInstance.wipe_asset_partitions.allow_missing_partitions as metadata on a downstream asset.object has no attribute '_base_path' error.statuses filter argument to the sensorsOrError GraphQL field was sometimes ignored when querying GraphQL for multiple sensors at the same time.CapturedLogManager and ComputeLogManager APIs into a single base class.PipesECSClient to allow Dagster to interface with ECS tasks.dbt subprocess to wait 25 seconds for the subprocess to cleanly terminate. Previously, it would only wait 2 seconds.sdf subprocess to wait 25 seconds for the subprocess to cleanly terminate. Previously, it would only wait 2 seconds.AutomationCondition using DagsterDltTranslator.get_automation_condition() (Thanks, aksestok!)dagsterDaemon.runRetries.retryOnAssetOrOpFailure to False in the Dagster Helm chart to prevent op retries and run retries from simultaneously firing on the same failure.recursive parameter (Thanks, chrishiste!)dagster/image label when image is provided from user_defined_k8s_config. (Thanks, @HynekBlaha!)custom_user_agent, was provided by defaultPipesGlueClient.run have been deprecated and will be removed in 1.9.0. The new params argument should be used instead.allow_missing_partitions configuration option.PartitionsDefinitions, there will no longer be an implicit asset job __ASSET_JOB_... for each PartitionsDefinition; there will just be one with all the assets. This reduces the time it takes to load code locations with assets with many different PartitionsDefinitions.OpenAIResource (Thanks, @chasleslr!)MultiPartitionMapping with @dbt_assets (Thanks, @arookieds!)AzureBlobComputeLogManager without a secret_key (Thanks, @ion-elgreco and @HynekBlaha!)AutomationCondition and associated static constructors.PipesCloudWatchMessageReaderDocs
New
OutputDefinition to take is_required rather than is_optional argument. This is to
remain consistent with changes to Field in 0.7.1 and to avoid confusion
with python's typing and dagster's definition of Optional, which indicates None-ability,
rather than existence. is_optional is deprecated and will be removed in a future version.Bugfixes
Dagit
dagster_pandas
dagster_pandas dataframes.dagster_aws
s3_resource no longer uses an unsigned session by default.Bugfixes
Documentation
Breaking Changes
There are a substantial number of breaking changes in the 0.7.0 release.
Please see 070_MIGRATION.md for instructions regarding migrating old code.
Scheduler
The scheduler configuration has been moved from the @schedules decorator to DagsterInstance.
Existing schedules that have been running are no longer compatible with current storage. To
migrate, remove the scheduler argument on all @schedules decorators:
instead of:
@schedules(scheduler=SystemCronScheduler) def define_schedules(): ...
Remove the scheduler argument:
@schedules def define_schedules(): ...
Next, configure the scheduler on your instance by adding the following to
$DAGSTER_HOME/dagster.yaml:
scheduler: module: dagster_cron.cron_scheduler class: SystemCronScheduler
Finally, if you had any existing schedules running, delete the existing $DAGSTER_HOME/schedules
directory and run dagster schedule wipe && dagster schedule up to re-instatiate schedules in a
valid state.
The should_execute and environment_dict_fn argument to ScheduleDefinition now have a
required first argument context, representing the ScheduleExecutionContext
Config System Changes
In the config system, Dict has been renamed to Shape; List to Array; Optional to
Noneable; and PermissiveDict to Permissive. The motivation here is to clearly delineate
config use cases versus cases where you are using types as the inputs and outputs of solids as
well as python typing types (for mypy and friends). We believe this will be clearer to users in
addition to simplifying our own implementation and internal abstractions.
Our recommended fix is not to use Shape and Array, but instead to use our new condensed
config specification API. This allow one to use bare dictionaries instead of Shape, lists with
one member instead of Array, bare types instead of Field with a single argument, and python
primitive types (int, bool etc) instead of the dagster equivalents. These result in
dramatically less verbose config specs in most cases.
So instead of
from dagster import Shape, Field, Int, Array, String # ... code config=Shape({ # Dict prior to change 'some_int' : Field(Int), 'some_list: Field(Array[String]) # List prior to change })
one can instead write:
config={'some_int': int, 'some_list': [str]}
No imports and much simpler, cleaner syntax.
config_field is no longer a valid argument on solid, SolidDefinition, ExecutorDefintion,
executor, LoggerDefinition, logger, ResourceDefinition, resource, system_storage, and
SystemStorageDefinition. Use config instead.
For composite solids, the config_fn no longer takes a ConfigMappingContext, and the context
has been deleted. To upgrade, remove the first argument to config_fn.
So instead of
@composite_solid(config={}, config_fn=lambda context, config: {})
one must instead write:
@composite_solid(config={}, config_fn=lambda config: {})
Field takes a is_required rather than a is_optional argument. This is to avoid confusion
with python's typing and dagster's definition of Optional, which indicates None-ability,
rather than existence. is_optional is deprecated and will be removed in a future version.
Required Resources
All solids, types, and config functions that use a resource must explicitly list that
resource using the argument required_resource_keys. This is to enable efficient
resource management during pipeline execution, especially in a multiprocessing or
remote execution environment.
The @system_storage decorator now requires argument required_resource_keys, which was
previously optional.
Dagster Type System Changes
dagster.Set and dagster.Tuple can no longer be used within the config system.DagsterType, rather than a class than inherits from
RuntimeType. Instead of dynamically generating a class to create a custom runtime type, just
create an instance of a DagsterType. The type checking function is now an argument to the
DagsterType, rather than an abstract method that has to be implemented in
a subclass.RuntimeType has been renamed to DagsterType is now an encouraged API for type creation.bool in addition
to a TypeCheck object.type_check_fn on DagsterType (formerly type_check and RuntimeType, respectively) now
takes a first argument context of type TypeCheckContext in addition to the second argument of
value.define_python_dagster_type has been eliminated in favor of PythonObjectDagsterType .dagster_type has been renamed to usable_as_dagster_type.as_dagster_type has been removed and similar capabilities added as
make_python_type_usable_as_dagster_type.PythonObjectDagsterType and usable_as_dagster_type no longer take a type_check argument. If
a custom type_check is needed, use DagsterType.dagster_pyspark or
dagster_pandas and expecting Pyspark or Pandas types to work as Dagster types, e.g., in type
annotations to functions decorated with @solid to indicate that they are input or output types
for a solid, you will need to call make_python_type_usable_as_dagster_type from your code in
order to map the Python types to the Dagster types, or just use the Dagster types
(dagster_pandas.DataFrame instead of pandas.DataFrame) directly.Other
step_metadata_fn has been removed from SolidDefinition & @solid.SolidDefinition & @solid now takes tags and enforces that values are strings or
are safely encoded as JSON. metadata is deprecated and will be removed in a future version.resource_mapper_fn has been removed from SolidInvocation.New
Dagit now includes a much richer execution view, with a Gantt-style visualization of step execution and a live timeline.
Early support for Python 3.8 is now available, and Dagster/Dagit along with many of our libraries are now tested against 3.8. Note that several of our upstream dependencies have yet to publish wheels for 3.8 on all platforms, so running on Python 3.8 likely still involves building some dependencies from source.
dagster/priority tags can now be used to prioritize the order of execution for the built-in
in-process and multiprocess engines.
dagster-postgres storages can now be configured with separate arguments and environment
variables, such as:
run_storage: module: dagster_postgres.run_storage class: PostgresRunStorage config: postgres_db: username: test password: env: ENV_VAR_FOR_PG_PASSWORD hostname: localhost db_name: test
Support for RunLaunchers on DagsterInstance allows for execution to be "launched" outside of
the Dagit/Dagster process. As one example, this is used by dagster-k8s to submit pipeline
execution as a Kubernetes Job.
Added support for adding tags to runs initiated from the Playground view in Dagit.
Added @monthly_schedule decorator.
Added Enum.from_python_enum helper to wrap Python enums for config. (Thanks @kdungs!)
[dagster-bash] The Dagster bash solid factory now passes along kwargs to the underlying
solid construction, and now has a single Nothing input by default to make it easier to create a
sequencing dependency. Also, logs are now buffered by default to make execution less noisy.
[dagster-aws] We've improved our EMR support substantially in this release. The
dagster_aws.emr library now provides an EmrJobRunner with various utilities for creating EMR
clusters, submitting jobs, and waiting for jobs/logs. We also now provide a
emr_pyspark_resource, which together with the new @pyspark_solid decorator makes moving
pyspark execution from your laptop to EMR as simple as changing modes.
[dagster-pandas] Added create_dagster_pandas_dataframe_type, PandasColumn, and
Constraint API's in order for users to create custom types which perform column validation,
dataframe validation, summary statistics emission, and dataframe serialization/deserialization.
[dagster-gcp] GCS is now supported for system storage, as well as being supported with the Dask executor. (Thanks @habibutsu!) Bigquery solids have also been updated to support the new API.
Bugfix
RunStorage clean up pipeline run tags when a run
is deleted. Requires a storage migration, using dagster instance migrate.@solid and @lambda_solid decorators now correctly wrap their decorated functions, in the
sense of functools.wraps.dagster_k8s has been hardened against various failure modes and is now
compatible with Helm 2.raise_error flag in execute_pipeline now actually raises user exceptions instead
of a wrapper type.Documentation
Thank you Thank you to all of the community contributors to this release!! In alphabetical order: @habibutsu, @kdungs, @vatervonacht, @zzztimbo.
Bugfix
Documentation
New
dagster-celery
Bugfix
@pyspark_solid decorator now handles inputs correctly.Documentation