acceldata-sdk-python 26.8.0__tar.gz → 26.10.0__tar.gz

acceldata_sdk_python-26.10.0/CHANGELOG.txt

@@ -0,0 +1,69 @@
+Version Log
+========================
+
+26.10.0 (10/10/2026)
+-------------------
+- Released **acceldata_sdk_python** version 26.10.0.
+- See [RELEASE_NOTES_26.10.0.md](RELEASE_NOTES_26.10.0.md) for full details.
+
+Generated models (OpenAPI) — catalog:
+- `DataplaneJobStatus` — new optional `outcome`.
+
+Generated models (OpenAPI) — pipeline:
+- `Identifier` — new optional `description`.
+- `Node` — new optional `retry_count`, `scope`.
+- `PipelineListingInfo` — new optional `tags_count`.
+- `PipelineRunInfo` — new response-only `link_confidence` and `link_direction` on `Node.linked_pipeline_run_info`. Callers do not set them.
+- `Span` — new optional `retry_count`.
+
+Compatibility:
+- Additive only. Existing constructors and payloads that omit the new fields still deserialize.
+
+26.9.0 (20/09/2026)
+------------------
+- Released **acceldata_sdk_python** version 26.9.0.
+
+Pipeline Control Plane API (SDK wiring):
+- `GET /torch-pipeline/api/pipelines/summary` — pagination (`page`, `size`, `sortBy`) and filters (name, owner, team, tags, run status/result, etc.).
+- `PUT /torch-pipeline/api/pipelines` — request body is now `PipelineInputRequest` (`pipeline` + `tagWriteMode`).
+- `GET /torch-pipeline/api/pipelines/{identity}` — responses include `tagAssociations` alongside `tags`.
+
+SDK changes:
+- `AdocClient.get_pipelines()` — new optional `pipeline_filter`, `page`, `size`, and `sort_by` parameters.
+- New `PipelineSummaryFilter` SDK type with `to_query_params()` for listing filters.
+- `AdocClient.create_pipeline()` — sends full `PipelineInputRequest` payload including `tagWriteMode` (default `TAGS`).
+- `AdocClient.update_pipeline()` — new optional `tag_write_mode` parameter (default `TAGS`).
+- `CreatePipelineInputRequest` — new `tag_write_mode` and `tag_associations` constructor parameters.
+
+Generated models (OpenAPI) — catalog:
+- `ColumnMapping` — new `lower_bound`, `upper_bound`.
+- `DataAnomalySegmentResult` — new `check_summaries`.
+- `SegmentResultResponse` — new `check_summaries` (list of `CheckSummaryResponse`); new `CheckSummaryResponse` type.
+- `RuleExecution` — new `has_multiple_data_planes`, `left_job_state`, `right_job_state`, `right_request_id`.
+- `YunikornSparkResourceConfig` — new `queue_name`.
+- `NotificationPayload` — `severity` is optional/nullable in OpenAPI (default `"MEDIUM"` on deserialize when omitted or null); applies to catalog and pipeline models.
+- `AnalyticsPipeline` — new `slot_config`, `slots_enabled`; new `SlotConfigInput` type (`provisioned_pushdown`, `provisioned_spark`).
+- `CrawlerRunConfig` — new `asset_batch_size`.
+- `CreateProfile` — new `priority`.
+- `PolicyExecutionRequest` — new `priority`.
+- `ExecutionResult` — new nullable `queued_job_phase` (`WAITING`, `ACQUIRED`, `DISPATCHED`, `IN_PROGRESS`).
+- Policy/rule models — new `tag_associations` on `ColumnMapping`, `DataCadenceRule`, `DataQualityRule`, `Item`, `QualityRule`, `ReconciliationRule`; new `TagDetail` type.
+
+Generated models (OpenAPI) — pipeline:
+- `PipelineInputRequest` — new `tag_write_mode` (default `TAGS`).
+- `Pipeline` — new `tag_associations`, `assembly_id`.
+- `Node` — new `linked_pipeline_run_info`; new `PipelineRunInfo` type.
+- `PipelineRunInfo` — new allowed `source_type`: `AWS_STEP_FUNCTIONS`.
+- `NotificationPayload` — `severity` optional/nullable (default `"MEDIUM"` on deserialize); same contract as catalog.
+
+Tests added:
+- `tests/unit/test_pipeline_summary_filter.py`
+- `tests/unit/test_create_pipeline_input_request.py`
+- `tests/unit/test_pipeline_service_listing_create.py`
+- `tests/unit/test_notification_payload.py`
+- `tests/unit/test_analytics_pipeline_deserialization.py`
+- `tests/unit/test_crawler_run_config_deserialization.py`
+- `tests/unit/test_create_profile_deserialization.py`
+- `tests/unit/test_pipeline_deserialization.py`
+- `tests/unit/test_policy_executions.py` — `TestPolicyExecutionRequestPriority`
+- `tests/unit/test_executor.py` — `ExecutionResult.queued_job_phase` coverage

{acceldata_sdk_python-26.8.0 → acceldata_sdk_python-26.10.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: acceldata_sdk_python
-Version: 26.8.0
+Version: 26.10.0
 Summary: Acceldata SDK Python
 Author-email: acceldata <apisupport@acceldata.io>
 License-Expression: MIT

acceldata_sdk_python-26.10.0/README-features.md

@@ -0,0 +1,288 @@
+# acceldata-sdk-python capabilities
+
+Public entry point: `AdocClient` (`acceldata.client.adoc_client`).
+
+```python
+AdocClient(
+    url,
+    access_key,
+    secret_key,
+    connection_timeout_ms=5000,
+    read_timeout_ms=15000,
+    verify_ssl=True,  # True, False, or a CA bundle path
+)
+```
+
+Calls go to the catalog, pipeline, and tag APIs. Resource objects returned by the client expose the same operations scoped to that object.
+
+## Pipelines
+
+| Capability | Method |
+| --- | --- |
+| Create a pipeline | `AdocClient.create_pipeline(CreatePipelineInputRequest)` |
+| Get a pipeline by identity | `AdocClient.get_pipeline(pipeline_identity)` |
+| Get a pipeline (alias) | `AdocClient.get_pipeline_res(pipeline_identity)` |
+| Wrap an existing pipeline model | `AdocClient.pipeline_resource(pipeline)` |
+| Update a pipeline | `AdocClient.update_pipeline(pipeline, tag_write_mode="TAGS")` |
+| Update a pipeline | `PipelineResource.update(pipeline)` |
+| Replace pipeline tags | `AdocClient.replace_pipeline_tags(pipeline_identity, tags=None, tag_associations=None, tag_write_mode="TAGS")` |
+| Replace pipeline tags | `PipelineResource.replace_tags(tags=None, tag_associations=None, tag_write_mode="TAGS")` |
+| Delete a pipeline | `AdocClient.delete_pipeline(pipeline_id)` |
+| Delete a pipeline | `PipelineResource.delete()` |
+| List pipelines | `AdocClient.get_pipelines(pipeline_filter=None, page=0, size=25, sort_by=None)` |
+
+`PipelineSummaryFilter` fields sent as query params: `name`, `description`, `owner`, `team`, `tag_key_ids`, `tag_key_value_ids`, `tags`, `pipeline_ids`, `pipeline_source`, `instrumented_by`, `namespaces`, `data_source`, `adf_ids`, `run_status`, `run_result`.
+
+`tag_write_mode` is `TAGS` or `ASSOCIATIONS`. `ASSOCIATIONS` uses Tag Service ids (`tag_associations`). `TAGS` uses display-name tags.
+
+`PipelineSourceType` values: `AIRFLOW`, `AZURE_DATA_FACTORY`, `TRINO`, `OTHERS`, `AUTOSYS`, `SPARK`, `DBT`, `FIVETRAN`, `SNAPLOGIC`, `HVR`, `AWS_STEP_FUNCTIONS`.
+
+## Pipeline runs
+
+| Capability | Method |
+| --- | --- |
+| Create a run | `PipelineResource.create_pipeline_run(PipelineRunInput)` |
+| Get a run by id, or by continuation id plus pipeline id | `AdocClient.get_pipeline_run(pipeline_run_id=None, continuation_id=None, pipeline_id=None)` |
+| Get a run on this pipeline | `PipelineResource.get_run(pipeline_run_id=None, continuation_id=None)` |
+| Get the latest run | `PipelineResource.get_latest_pipeline_run()` |
+| List runs | `AdocClient.get_pipeline_runs(pipeline_id)` |
+| List runs | `PipelineResource.get_runs()` |
+| Update a run (status, result, times, context, continuation id) | `PipelineRunResource.update_pipeline_run(request=None, context_data=None, continuation_id=None, started_at=None, finished_at=None, result=None, status=None)` |
+| Graph for a run (nodes and edges) | `PipelineRunResource.get_details()` |
+
+`PipelineRunInput` fields: `context_data`, `continuation_id`, `started_at`, `finished_at`, `result`, `status`. Create uses `context_data`, `continuation_id`, and `started_at`. The created run status is `STARTED`.
+
+`PipelineRunStatus`: `STARTED`, `COMPLETED`, `FAILED`, `ABORTED`.
+
+`PipelineRunResult`: `RUNNING`, `SUCCESS`, `FAILURE`, `CANCELLED`.
+
+A continuation id loads the same run from another call via `get_pipeline_run` or `PipelineResource.get_run`. `started_at` and `finished_at` set historical timestamps.
+
+## Jobs
+
+| Capability | Method |
+| --- | --- |
+| Create a job on a pipeline (requires `pipeline_run_id`) | `PipelineResource.create_job(CreateJobInput)` |
+| Create a job on a run (`pipeline_run_id` defaults to this run) | `PipelineRunResource.create_job(CreateJobInput)` |
+
+`CreateJobInput` fields: `uid`, `name`, `description`, `inputs`, `outputs`, `meta`, `pipeline_run_id`, `bounded_by_span`, `span_uid`, `with_explicit_time`, `span_created_at`, `span_updated_at`, `context`.
+
+`JobInputOutputRef` is either `asset_uid` or `job_uid`. When `bounded_by_span` is true, `span_uid` is required and the SDK also creates a span for the job.
+
+## Spans
+
+| Capability | Method |
+| --- | --- |
+| Create a span | `AdocClient.create_span(pipeline_run_id, span)` |
+| Create the root span | `PipelineRunResource.create_root_span(uid, associated_job_uids=None, with_explicit_time=False, created_at=None, updated_at=None)` |
+| Create a child span | `SpanResource.create_child_span(uid, associated_job_uids=None, with_explicit_time=False, created_at=None, updated_at=None)` |
+| Get the root span | `AdocClient.get_root_span(pipeline_run_id)` |
+| Get the root span | `PipelineRunResource.get_root_span()` |
+| Get the root span for a run or the latest run | `PipelineResource.get_root_span(pipeline_run_id=None, continuation_id=None)` |
+| Get a span by uid | `AdocClient.get_span(pipeline_run_id, span_identity)` |
+| Get a span by uid | `PipelineRunResource.get_span(span_identity)` |
+| List spans for a run | `AdocClient.get_spans(pipeline_run_id)` |
+| List spans for a run | `PipelineRunResource.get_spans()` |
+| List child spans | `AdocClient.get_child_spans(span_id)` |
+| List child spans | `SpanResource.get_child_spans()` |
+| Whether the span has children | `SpanResource.has_children()` |
+| Whether the span is the root | `SpanResource.is_root()` |
+| Whether the span is finished | `SpanResource.is_span_finished()` |
+| Start a span | `SpanResource.start(context_data=None, created_at=None)` |
+| End a span | `SpanResource.end(context_data=None, created_at=None)` |
+| Abort a span | `SpanResource.abort(context_data=None, created_at=None)` |
+| Fail a span | `SpanResource.failed(context_data=None, created_at=None)` |
+| Send a log or custom event | `SpanResource.send_event(LogEvent or GenericEvent)` |
+| Append a span event | `AdocClient.create_span_event(SpanEventRequest)` |
+
+Span event uids: `SPAN.START`, `SPAN.END`, `SPAN.ABORTED`, `SPAN.FAILED`, `LOG`.
+
+`LogEvent(log_data, context_data=None, created_at=None)` sends `LOG`. `GenericEvent(context_data=None, event_uid=..., log_data=None, created_at=None)` sends a custom `event_uid`.
+
+`created_at` and `updated_at` set explicit timestamps for historical loads.
+
+## Datasources and crawlers
+
+| Capability | Method |
+| --- | --- |
+| Get a datasource by name or id | `AdocClient.get_datasource(assembly_identifier, properties=False)` |
+| Get a datasource (alias) | `AdocClient.get_datasource_resource(assembly_identifier, properties=False)` |
+| Get a datasource by id | `AdocClient.get_datasource_by_id(id, properties=False)` |
+| List datasources, optional source-type filter | `AdocClient.get_datasources(source_type=None)` |
+| List all datasources | `AdocClient.get_all_datasources()` |
+| Wrap an assembly | `AdocClient.datasource_resource(assembly)` |
+| Reload a datasource | `DatasourceResource.refresh(properties=False)` |
+| Start a crawler | `AdocClient.start_crawler(datasource_name, transient_retry=None)` |
+| Start a crawler | `DatasourceResource.start_crawler(transient_retry=None)` |
+| Read crawler status | `AdocClient.get_crawler_status(datasource_name, transient_retry=None)` |
+| Read crawler status | `DatasourceResource.get_crawler_status(transient_retry=None)` |
+
+`AssetSourceType` is the catalog source-type filter (Snowflake, BigQuery, Databricks, Kafka, S3, dbt Cloud, and the other catalog source values).
+
+## Assets, metadata, and sample data
+
+| Capability | Method |
+| --- | --- |
+| Get an asset by numeric id or uid | `AdocClient.get_asset(identifier)` |
+| Get an asset by id | `AdocClient.asset_service.get_asset_by_id(id)` |
+| Get an asset by uid | `AdocClient.asset_service.get_asset_by_uid(uid)` |
+| Wrap an asset | `AdocClient.asset_resource(asset)` |
+| List asset type definitions | `AdocClient.get_asset_types()` |
+| List asset type definitions (alias) | `AdocClient.get_all_asset_types()` |
+| Read asset metadata | `AdocClient.get_asset_metadata(asset_id)` |
+| Read asset metadata | `AssetResource.get_metadata()` |
+| Add custom metadata (string map) | `AdocClient.add_custom_metadata(asset_id, custom_metadata)` |
+| Add custom metadata | `AssetResource.add_custom_metadata(list[CustomAssetMetadata])` |
+| Trigger sample data (`sync` polls until finished) | `AdocClient.sample_data(asset_id, sync=True)` |
+| Trigger sample data | `AssetResource.sample_data(sync=True)` |
+| Poll a sample-data request | `AdocClient.get_sample_data_result(request_id)` |
+
+## Profiling
+
+| Capability | Method |
+| --- | --- |
+| Start profiling | `AdocClient.profile_asset(asset_id, profiling_params=None, sync=False, transient_retry=None)` |
+| Start profiling | `AssetResource.start_profile(profiling_params, sync=False, transient_retry=None)` |
+| Latest profile status | `AdocClient.get_latest_profile_status(asset_id)` |
+| Latest profile status | `AssetResource.get_latest_profile_status()` |
+| Profile request by asset and request id | `AdocClient.get_profile_request_details(asset_id, req_id)` |
+| Profile request (alias) | `AdocClient.get_profile_status(asset_id, req_id)` |
+| Refresh a profile request | `ProfileRequestResource.get_status()` |
+| Cancel a profile request | `AdocClient.cancel_profile(profile_req_id)` |
+| Cancel a profile request | `ProfileRequestResource.cancel()` |
+| Wrap a profile request | `AdocClient.profile_request_resource(profile_request)` |
+| Wrap a profile request (alias) | `AdocClient.profile_resource(profile_request)` |
+
+`AssetProfilingParams(profiling_type, marker_config=None)`. `ProfilingType`: `FULL`, `INCREMENTAL`, `SELECTIVE`. Incremental and selective require `marker_config`.
+
+Marker types: `id`, `datetime`, `bound`, `boundDateTime`, `boundFileEvent`, `file`, `offset`, `date_partitioned`, `timestamp`, `pubsub_timestamp`, `pubsub_lookback`.
+
+`sync=False` returns after the request is accepted. `sync=True` polls until the request is no longer running.
+
+## Policies
+
+Executable kinds: data quality, reconciliation, data cadence (freshness).
+
+`RuleType` also includes `EQUALITY`, `DATA_DRIFT`, and `SCHEMA_DRIFT` for listing and execution history. `Executor` and `execute_policy` run data quality, reconciliation, and data cadence.
+
+| Capability | Method |
+| --- | --- |
+| List rules | `AdocClient.list_all_policies(policy_filter, page=0, size=25, with_latest_execution=True, sort_by="updatedAt:DESC")` |
+| Get a rule (untyped, or typed when `type` is set) | `AdocClient.get_policy(type=None, identifier=None)` |
+| Data quality resource | `AdocClient.get_data_quality_rule_resource(identifier)` |
+| Data quality resource (alias) | `AdocClient.get_dq_policy_resource(identifier)` |
+| Reconciliation resource | `AdocClient.get_reconciliation_rule_resource(identifier)` |
+| Reconciliation resource (alias) | `AdocClient.get_recon_policy_resource(identifier)` |
+| Data cadence resource | `AdocClient.get_data_cadence_rule_resource(identifier)` |
+| Data cadence resource (alias) | `AdocClient.get_cadence_policy_resource(identifier)` |
+| Reload a typed rule | `DataQualityRuleResource.refresh()`, `ReconciliationRuleResource.refresh()`, `DataCadenceRuleResource.refresh()` |
+| List executions | `AdocClient.policy_executions(identifier, rule_type, page=0, size=25, sort_by="finishedAt:DESC")` |
+| List executions | `DataQualityRuleResource.policy_executions(...)`, `ReconciliationRuleResource.policy_executions(...)`, `DataCadenceRuleResource.policy_executions(...)` |
+| Start a data quality run | `AdocClient.execute_dq_rule(rule_id, policy_execution_request)` |
+| Start a data quality run | `DataQualityRuleResource.execute(policy_execution_request)` |
+| Start a reconciliation run | `AdocClient.execute_reconciliation_rule(rule_id, policy_execution_request)` |
+| Start a reconciliation run | `ReconciliationRuleResource.execute(policy_execution_request)` |
+| Start a cadence / freshness run | `AdocClient.execute_freshness_rule(rule_id, policy_execution_request)` |
+| Start a cadence / freshness run | `DataCadenceRuleResource.execute(policy_execution_request)` |
+| Start a run and optionally wait | `AdocClient.execute_policy(rule_type, rule_id, policy_execution_request, sync=True, failure_strategy=FailureStrategy.DoNotFail, sleep_interval=5.0, total_retries=-1, transient_retry=None, pipeline_run_id=None)` |
+| Start a run and optionally wait | `DataQualityRuleResource.execute_policy(...)`, `ReconciliationRuleResource.execute_policy(...)`, `DataCadenceRuleResource.execute_policy(...)` |
+| Data quality execution details | `AdocClient.get_dq_rule_execution_details(execution_id)` |
+| Data quality execution details | `DataQualityRuleResource.get_execution_details(execution_id)` |
+| Reconciliation execution details | `AdocClient.get_reconciliation_rule_execution_details(execution_id)` |
+| Reconciliation execution details | `ReconciliationRuleResource.get_execution_details(execution_id)` |
+| Data cadence execution details | `AdocClient.get_data_cadence_rule_execution_details(execution_id)` |
+| Data cadence execution details | `DataCadenceRuleResource.get_execution_details(execution_id)` |
+| Data quality result | `AdocClient.get_dq_rule_result(execution_id)` |
+| Data quality result | `DataQualityRuleResource.get_result(execution_id)` |
+| Reconciliation result | `AdocClient.get_reconciliation_rule_result(execution_id)` |
+| Reconciliation result | `ReconciliationRuleResource.get_result(execution_id)` |
+| Data cadence result | `AdocClient.get_freshness_rule_result(execution_id)` |
+| Data cadence result | `DataCadenceRuleResource.get_result(execution_id)` |
+| Current execution status | `AdocClient.get_policy_execution_status(rule_type, execution_id, transient_retry=None)` |
+| Current execution status | typed resource `.get_policy_execution_status(execution_id, transient_retry=None)` |
+| Wait for a terminal result | `AdocClient.get_policy_execution_result(rule_type, execution_id, failure_strategy=FailureStrategy.DoNotFail, transient_retry=None)` |
+| Wait for a terminal result | typed resource `.get_policy_execution_result(...)` |
+| Cancel an execution | `AdocClient.cancel_rule_execution(execution_id)` |
+| Cancel an execution | typed resource `.cancel_execution(execution_id)` |
+| Enable a rule | `AdocClient.enable_rule(rule_id)` |
+| Enable a rule | typed resource `.enable()` |
+| Disable a rule | `AdocClient.disable_rule(rule_id)` |
+| Disable a rule | typed resource `.disable()` |
+
+`PolicyFilter` fields: `period`, `tags`, `last_execution_result`, `assets`, `policy_type`, `enable`, `active`, `data_sources`.
+
+`ExecutionPeriod`: `Last15minutes`, `Last30minutes`, `Last1hour`, `Last3hours`, `Last6hours`, `Last12hours`, `Last24hours`, `Today`, `Yesterday`, `Last7days`, `Thismonth`, `Last1month`, `Last3month`.
+
+`PolicyType`: `DATA_QUALITY`, `RECONCILIATION`, `DATA_CADENCE`.
+
+`PolicyExecutionType` on the request: `FULL`, `INCREMENTAL`, `SELECTIVE`.
+
+`PolicyExecutionInput` fields: `executionType`, `markerConfigs`, `ruleItemSelections`, `includeInQualityScore`, `pipelineRunId`, `columnVariables`, `sparkSQLDynamicFilterVariableMapping`, `sparkFilterSelectedColumns`, `sparkResourceConfig`.
+
+Selective marker configs used by policy runs include ID bounds, date-time bounds, file-event bounds, and pubsub / Kafka timestamp.
+
+`FailureStrategy`: `DoNotFail`, `FailOnWarning`, `FailOnError`.
+
+`Executor` is the handle returned by `execute_policy`:
+
+| Capability | Method |
+| --- | --- |
+| Start the run held by this executor | `Executor.execute(rule_id, policy_execution_request, sleep_interval=5.0, total_retries=-1, failure_strategy=..., transient_retry=None)` |
+| Status of this run | `Executor.get_status()` |
+| Status of an execution id | `Executor.get_execution_status(execution_id, transient_retry=None)` |
+| Wait for this run | `Executor.get_result(sleep_interval=5.0, total_retries=-1, failure_strategy=..., transient_retry=None)` |
+| Wait for an execution id | `Executor.get_execution_result(execution_id, ...)` |
+| Cancel this run | `Executor.cancel()` |
+| Cancel an execution | `Executor.cancel_rule_execution(execution_id)` |
+| Execution details | `Executor.get_dq_rule_execution_details(execution_id)`, `Executor.get_reconciliation_rule_execution_details(execution_id)` |
+| Enable or disable a rule | `Executor.enable_rule(rule_id)`, `Executor.disable_rule(rule_id)` |
+
+`sync=True` on `execute_policy` blocks until a terminal result. `sync=False` returns after the run is accepted. `total_retries=-1` means no cap on `RUNNING` / `WAITING` poll rounds.
+
+## Tags
+
+| Capability | Method |
+| --- | --- |
+| Create a tag key | `AdocClient.create_tag(CreateTagRequest)` |
+| Wrap a tag response | `AdocClient.tag_resource(tag)` |
+| Add values to a key | `AdocClient.add_values_to_tag_key(key_id, values)` |
+| Add values to a key | `TagResource.add_values(values)` |
+| Search keys and values | `AdocClient.search_tag_key_values(SearchKeyValuesRequest)` |
+| Associate tag values with an entity | `AdocClient.associate_tags(AssociateTagsRequest)` |
+| Tags linked to an entity | `AdocClient.get_tags_for_entity(entity_id, entity_type)` |
+
+`CreateTagType`: `UNSPECIFIED_TYPE`, `GOVERNED`, `USER`.
+
+`SearchKeyValuesRequest` query fields: `page`, `size`, `sort_by`, `name` (key name), `key_type`, `key_status`, `key_ids`, `key_created_by`, `value`, `value_status`, `data_protection`, `key_value_ids`, `value_created_by`.
+
+`TagEntityType`: `ASSET`, `POLICY`, `BULK_POLICY_RULE`, `BULK_POLICY_RULESET`, `POLICY_RULE_ITEM`, `RECONCILIATION_COLUMN_MAPPING`, `USER_DEFINED_FUNCTION`, `LABELER`.
+
+## Asset tags and labels
+
+These methods are on the client and on `AssetResource`. They are marked deprecated in the SDK. Tag association calls above cover the same entities.
+
+| Capability | Method |
+| --- | --- |
+| List asset tags | `AdocClient.get_asset_tags(asset_id)`, `AssetResource.get_asset_tags()` |
+| Add an asset tag | `AdocClient.add_asset_tag(asset_id, AssetTagRequest)`, `AssetResource.add_asset_tag(payload)` |
+| List asset labels | `AdocClient.get_asset_labels(asset_id)`, `AssetResource.get_asset_labels()` |
+| Add asset labels | `AdocClient.add_asset_labels(asset_id, AssetLabelRequest)`, `AssetResource.add_asset_labels(payload)` |
+| List catalog tags | `AdocClient.asset_service.get_tags()` |
+
+## Catalog lookups on the client services
+
+These are called on the service objects held by `AdocClient`, not as methods of `AdocClient` itself.
+
+| Capability | Method |
+| --- | --- |
+| Connection types | `AdocClient.generic_service.get_connection_types()` |
+| Source types | `AdocClient.generic_service.get_source_types()` |
+| Analytics pipeline by id or name | `AdocClient.generic_service.get_analysis_pipeline(pipeline_identifier)` |
+
+## Retries
+
+`RetryConfig(enabled=True, max_attempts=10, initial_interval_seconds=30, max_interval_seconds=600, retry_status_codes=...)`.
+
+Default retry HTTP statuses: 408, 409, 422, 429, 500, 502, 503, 504. Connection failures are retried the same way.
+
+`transient_retry` is a parameter on crawler start, crawler status, profiling, policy status, policy result, and `execute_policy` / `Executor`.

{acceldata_sdk_python-26.8.0 → acceldata_sdk_python-26.10.0}/README-pipeline.md

@@ -24,14 +24,13 @@ The rest of this guide is organized by topic (definition, runs, spans, jobs, eve
 
 ## Creating a pipeline
 
-`CreatePipelineInputRequest` accepts an optional `tags` argument: a list of `Tag` objects attached to the pipeline definition for discovery and display in ADOC. Each tag has a `name` (stable identifier, often a `key:value` style string) and a `displayName` (human-readable label in the UI). When `tags` is omitted, the pipeline is created with no tags.
-
-Updating labels after the fact is a separate call: `PipelineResource.replace_tags(...)` **replaces the entire tag list** with the supplied list. To add or remove a single tag, send a new full list that reflects the desired state.
+Create and update use a `PipelineInputRequest`-shaped body: a `pipeline` object plus
+`tagWriteMode` (default `TAGS`). Prefer `CreatePipelineInputRequest`, which builds that
+payload for you.
 
 ```python
 from acceldata.client.adoc_client import AdocClient
 from acceldata.models.api.pipeline.meta import Meta
-from acceldata.models.api.pipeline.tag import Tag
 from acceldata.models.sdk.pipeline.create_pipeline_input_request import (
     CreatePipelineInputRequest,
 )
@@ -48,15 +47,135 @@ pipeline = client.create_pipeline(
         name="Monthly reporting pipeline",
         description="Builds monthly reporting datasets",
         meta=Meta(owner="data-team", team="analytics", code_location="..."),
+    )
+)
+print(pipeline.to_dict())
+```
+
+For tagging pipelines, see [Pipeline tags](#pipeline-tags) below.
+
+## Pipeline tags
+
+> **Since 26.9.0:** `tagAssociations`, `tag_write_mode` on create/update, and the
+> deprecation of legacy `pipeline.tags` apply from **acceldata-sdk-python 26.9.0**
+> onwards.
+>
+> **Deprecation:** The legacy `pipeline.tags` field is **deprecated from 26.9.0** and
+> will be removed **after 3 releases**. Affected APIs include
+> `CreatePipelineInputRequest.tags`, and display-name writes via
+> `PipelineResource.replace_tags` / `AdocClient.replace_pipeline_tags` with the default
+> `tag_write_mode="TAGS"`. Tags are migrating to **`tagAssociations`**, which reference
+> existing Tag Service key/value ids (`keyId`, `keyValueId`). **New code should use
+> `tag_associations` with `tag_write_mode="ASSOCIATIONS"`.** The `tags` field still
+> appears in API responses for backward compatibility. To create keys/values and resolve
+> ids, see the [Tags guide](README-unified-tags.md).
+
+### `tag_write_mode`
+
+Available from **26.9.0** on `CreatePipelineInputRequest`, `AdocClient.create_pipeline()`,
+and `AdocClient.update_pipeline()`.
+
+Create and update pipeline calls send `tagWriteMode` (SDK: `tag_write_mode`) to select
+which tag payload is applied. On **update**, the supplied tag set **replaces** the
+existing associations — an empty list clears all tags.
+
+| `tag_write_mode` | Default | Payload applied | `pipeline.tags` on write | `pipeline.tagAssociations` on write | Response includes |
+| --- | --- | --- | --- | --- | --- |
+| `TAGS` | Yes | Legacy display-name tags | Each `Tag.displayName` is find-or-created in the Tag Service (as both key name and value name), then associated with the pipeline. `Tag.name` is stored but not used for find-or-create. | Ignored | Both `tags` and `tagAssociations` (with resolved key/value names) |
+| `ASSOCIATIONS` | No | Tag Service id references | Ignored entirely | Each `TagDetail` with `keyId` and `keyValueId` is associated with the pipeline. Optional `keyName` / `valueName` are read-only in responses. | Both `tags` and `tagAssociations` (with resolved key/value names) |
+
+| SDK parameter | Wire field | Allowed values |
+| --- | --- | --- |
+| `tag_write_mode` | `tagWriteMode` | `TAGS`, `ASSOCIATIONS` |
+
+### Create with `tagAssociations` (recommended)
+
+Use this when tag keys and values already exist in the Tag Service (or after creating
+them via the [Tags guide](README-unified-tags.md)).
+
+```python
+from acceldata.models.api.pipeline.tag_detail import TagDetail
+from acceldata.models.sdk.pipeline.create_pipeline_input_request import (
+    CreatePipelineInputRequest,
+)
+
+pipeline = client.create_pipeline(
+    CreatePipelineInputRequest(
+        uid="monthly_reporting_pipeline_tagged",
+        name="Monthly reporting pipeline (associations)",
+        tag_associations=[
+            TagDetail(keyId=12, keyValueId=34),
+            TagDetail(keyId=12, keyValueId=56),
+        ],
+        tag_write_mode="ASSOCIATIONS",
+    )
+)
+```
+
+### Create with `tags` (deprecated)
+
+Each tag has a `name` (stable identifier, often a `key:value` style string) and a
+`displayName` (human-readable label in the UI). Prefer `tag_associations` instead.
+
+```python
+from acceldata.models.api.pipeline.tag import Tag
+from acceldata.models.sdk.pipeline.create_pipeline_input_request import (
+    CreatePipelineInputRequest,
+)
+
+# tag_write_mode="TAGS" is the default; shown here for clarity
+pipeline = client.create_pipeline(
+    CreatePipelineInputRequest(
+        uid="monthly_reporting_pipeline",
+        name="Monthly reporting pipeline",
         tags=[
             Tag(name="env:prod", displayName="Environment: Production"),
             Tag(name="domain:finance", displayName="Domain: Finance"),
         ],
+        tag_write_mode="TAGS",
     )
 )
-print(pipeline.to_dict())
 ```
 
+### Replace tags on an existing pipeline
+
+Both approaches below **replace the entire tag set** with the supplied list. To add or
+remove a single tag, send the full list that reflects the desired state.
+
+#### Using `replace_tags` with `tagAssociations` (recommended)
+
+```python
+from acceldata.models.api.pipeline.tag_detail import TagDetail
+
+pipeline.replace_tags(
+    tag_associations=[
+        TagDetail(keyId=12, keyValueId=34),
+        TagDetail(keyId=15, keyValueId=78),
+    ],
+    tag_write_mode="ASSOCIATIONS",
+)
+```
+
+Or via the client:
+
+```python
+client.replace_pipeline_tags(
+    pipeline_uid,
+    tag_associations=[TagDetail(keyId=12, keyValueId=34)],
+    tag_write_mode="ASSOCIATIONS",
+)
+```
+
+#### Using `replace_tags` with display-name tags (deprecated)
+
+> **Deprecated from 26.9.0.** Display-name tag writes via `replace_tags` /
+> `replace_pipeline_tags` with the default `tag_write_mode="TAGS"` will be **removed
+> after 3 releases**. Migrate to `tag_associations` with
+> `tag_write_mode="ASSOCIATIONS"` (see above).
+
+Use this when you want to set display-name tags without pre-resolving Tag Service ids.
+It performs a GET + PUT under the default `tag_write_mode="TAGS"`.
+
 ```python
 from acceldata.models.api.pipeline.tag import Tag
 
@@ -68,6 +187,24 @@ pipeline.replace_tags(
 )
 ```
 
+Alternatively, call `AdocClient.replace_pipeline_tags(pipeline_uid, tags)` for the same
+behavior without a `PipelineResource` instance.
+
+You can also use `update_pipeline` directly if you already have the full pipeline object:
+
+```python
+from acceldata.models.api.pipeline.tag_detail import TagDetail
+
+updated = pipeline.pipeline.model_copy(
+    update={
+        "tagAssociations": [
+            TagDetail(keyId=12, keyValueId=34),
+        ],
+    }
+)
+client.update_pipeline(updated, tag_write_mode="ASSOCIATIONS")
+```
+
 ## Pipeline runs
 
 ```python
@@ -406,14 +543,89 @@ runs = client.get_pipeline_runs(pipeline.id)
 runs = pipeline.get_runs()
 ```
 
-### Get all pipelines
+### Get all pipelines (summary listing)
+
+`get_pipelines()` calls `GET /pipelines/summary` and returns a paginated listing of
+pipeline summaries (metadata plus latest run info when available).
 
-`get_pipelines()` returns a listing object; iterate `pipelines` for each row (including `pipeline_summary` with `uid` and `name`).
+| Parameter | Default | Description |
+| --- | --- | --- |
+| `pipeline_filter` | `None` | `PipelineSummaryFilter` instance |
+| `page` | `0` | 0-based page index |
+| `size` | `25` | Results per page |
+| `sort_by` | `None` | Sort as `field:direction`, e.g. `name:asc`, `createdAt:desc`, `latestRunResult:desc` |
+
+Pass filters via `PipelineSummaryFilter`. All filter values are sent as query-string
+parameters (camelCase on the wire). The server parses comma-separated values where
+noted below — matching `PipelineQueryEnhancer` on the torch-pipeline service.
+
+#### Filter categories
+
+| Category | SDK fields | Wire format | Notes |
+| --- | --- | --- | --- |
+| Single value | `name`, `description`, `pipeline_source`, `instrumented_by`, `run_status` | One string | Partial match for `name` / `description`; single enum for `run_status` (`runStatus`) and `pipeline_source` (`pipelineSource`, or `ALL`) |
+| Comma-separated strings | `owner`, `team`, `run_result`, `data_source`, `namespaces` | `"a,b"` string | Pass multiple values as one comma-separated string, e.g. `owner="alice,bob"` |
+| Comma-separated ids | `tag_key_ids`, `tag_key_value_ids`, `tags`, `pipeline_ids`, `adf_ids` | `"1,2"` string | Tag Service / pipeline / ADF ids. SDK also accepts a single `int` or `list[int]` and joins them for you |
+
+#### Filter field reference
+
+| SDK field | Query param | Type on wire |
+| --- | --- | --- |
+| `name` | `name` | string (partial match) |
+| `description` | `description` | string (partial match) |
+| `owner` | `owner` | comma-separated string |
+| `team` | `team` | comma-separated string |
+| `tag_key_ids` | `tagKeyIds` | comma-separated long ids |
+| `tag_key_value_ids` | `tagKeyValueIds` | comma-separated long ids |
+| `tags` | `tags` | comma-separated long ids (alias for `tagKeyValueIds`; merged server-side) |
+| `pipeline_ids` | `pipelineIds` | comma-separated long ids |
+| `pipeline_source` | `pipelineSource` | string (`AIRFLOW`, `ALL`, …) |
+| `instrumented_by` | `instrumentedBy` | string (`SDK`, `OPENLINEAGE`; requires `pipeline_source=AIRFLOW`) |
+| `namespaces` | `namespaces` | comma-separated string (OpenLineage; requires `instrumented_by=OPENLINEAGE` on Airflow) |
+| `data_source` | `dataSource` | comma-separated string |
+| `adf_ids` | `adfIds` | comma-separated long ids (Azure Data Factory integration ids) |
+| `run_status` | `runStatus` | string (single latest-run status) |
+| `run_result` | `runResult` | comma-separated string (e.g. `SUCCESS,FAILURE`) |
 
 ```python
+from acceldata.models.sdk.pipeline.pipeline_summary_filter import PipelineSummaryFilter
+
+# First page (defaults: page=0, size=25)
 listing = client.get_pipelines()
 for row in listing.pipelines:
     print(row.pipeline_summary.uid, row.pipeline_summary.name)
+
+# Paginated + filtered (strings and comma-separated values)
+listing = client.get_pipelines(
+    pipeline_filter=PipelineSummaryFilter(
+        name="reporting",
+        owner="data-team",
+        tag_key_value_ids="34,56",
+        run_result="SUCCESS",
+    ),
+    page=0,
+    size=10,
+    sort_by="name:asc",
+)
+
+# Filter by pipeline ids — pass a list of ints or a comma-separated string
+listing = client.get_pipelines(
+    pipeline_filter=PipelineSummaryFilter(
+        pipeline_ids=[487648, 489112],
+    ),
+)
+
+# Tag Service filters — list[int] is supported for id fields
+listing = client.get_pipelines(
+    pipeline_filter=PipelineSummaryFilter(
+        tag_key_ids=[12],
+        tags=[34, 56],  # alias for tagKeyValueIds
+    ),
+)
+
+for row in listing.pipelines:
+    summary = row.pipeline_summary
+    print(summary.uid, summary.name, summary.owner, row.latest_run_result)
 ```
 
 ### Delete a pipeline

{acceldata_sdk_python-26.8.0 → acceldata_sdk_python-26.10.0}/README-unified-tags.md

@@ -82,6 +82,24 @@ client.associate_tags(
 )
 ```
 
+Pipelines can also take Tag Service associations at create/update time via
+`tagWriteMode=ASSOCIATIONS` on `CreatePipelineInputRequest` (see
+[Pipelines guide](README-pipeline.md)).
+
+```python
+from acceldata.models.api.pipeline.tag_detail import TagDetail
+from acceldata.models.sdk.pipeline import CreatePipelineInputRequest
+
+pipeline = client.create_pipeline(
+    CreatePipelineInputRequest(
+        uid="etl_pipeline",
+        name="ETL pipeline",
+        tag_associations=[TagDetail(keyId=12, keyValueId=key_value_ids[0])],
+        tag_write_mode="ASSOCIATIONS",
+    )
+)
+```
+
 ## Verify associations
 
 ```python