Migration Reference

This guide helps you move from the legacy Python package acceldata-sdk (import namespace acceldata_sdk, client TorchClient) to acceldata-sdk-python (import namespace acceldata, client AdocClient).

Legacy package notice
acceldata-sdk-python is the supported path for catalog, pipeline, policy, and tagging workflows against ADOC. acceldata-sdk is now in maintenance mode and is supported for up to three additional releases. For an overview of why to migrate, see Migration Guide and New Features Overview.

acceldata-sdk-python uses OpenAPI-generated models (Pydantic v2), typed resource wrappers, and a clearer separation between API transport errors and SDK workflow errors.

At a Glance

Topic	Legacy (acceldata-sdk)	New (acceldata-sdk-python)
PyPI package	`acceldata-sdk`	`acceldata-sdk-python`
Import root	`acceldata_sdk`	`acceldata`
Client class	`TorchClient`	`AdocClient`
Python version	3.7+	3.10+
Models	Hand-written classes / dataclasses	OpenAPI-generated Pydantic models + SDK resource wrappers
Pipeline return types	`Pipeline`, `PipelineRun` with embedded client	`PipelineResource`, `PipelineRunResource`, `SpanResource`
Policy execution type param	`policy_type: PolicyType`	`rule_type: RuleType`
Incremental policy runs	`incremental=True` on `execute_policy`	`PolicyExecutionInput(executionType=PolicyExecutionType.INCREMENTAL)`
Job I/O references	`Node`	`JobInputOutputRef`
Job / pipeline metadata	`JobMetadata`, `PipelineMetadata`	`Meta`
Policy status	`get_policy_status`	`get_policy_execution_status`
SDK usage errors	`TorchSdkException`	`AcceldataSdkException` (+ specific subclasses)
Connection timeouts	`torch_connection_timeout_ms`, `torch_read_timeout_ms`	`connection_timeout_ms`, `read_timeout_ms`

Prerequisites

Upgrade Python to 3.10 or newer. The new SDK does not support Python 3.7–3.9.
Align SDK and ADOC versions. Install an acceldata-sdk-python release that matches your ADOC deployment.
Plan a dependency swap, not a side-by-side install. Both packages target overlapping functionality but use different import paths. Migrate imports and remove acceldata-sdk from requirements.txt or pyproject.toml when done.

Step 1: Change the Package

Before:

pip install acceldata-sdk

After:

pip uninstall acceldata-sdk # when you are ready to cut over pip install acceldata-sdk-python

Update dependency files:

- acceldata-sdk>=26.4.0 + acceldata-sdk-python>=<target-version>

Step 2: Update Imports and Client Construction

Client

Legacy	New
`from acceldata_sdk.torch_client import TorchClient`	`from acceldata.client.adoc_client import AdocClient`
`TorchClient(url=..., access_key=..., secret_key=...)`	`AdocClient(url=..., access_key=..., secret_key=...)`
`torch_connection_timeout_ms`	`connection_timeout_ms`
`torch_read_timeout_ms`	`read_timeout_ms`
`do_version_check` (deprecated, unused)	Removed — not supported

Before:

from acceldata_sdk.torch_client import TorchClient client = TorchClient( url="https://<your-adoc-url>", access_key="<access-key>", secret_key="<secret-key>", torch_connection_timeout_ms=10_000, torch_read_timeout_ms=20_000, )

After:

from acceldata.client.adoc_client import AdocClient client = AdocClient( url="https://<your-adoc-url>", access_key="<access-key>", secret_key="<secret-key>", connection_timeout_ms=10_000, read_timeout_ms=20_000, # Optional: verify_ssl=False, or a path to a CA bundle for private TLS )

Constants

Legacy	New
`acceldata_sdk.constants.TORCH_CONNECTION_TIMEOUT_MS`	`acceldata.constants.CONNECTION_TIMEOUT_MS`
`acceldata_sdk.constants.TORCH_READ_TIMEOUT_MS`	`acceldata.constants.READ_TIMEOUT_MS`

Environment Variables (Scripts / Operators)

If you configure the client from the environment (common in Airflow and CI), rename the timeout variables:

Legacy	New
`TORCH_CONNECTION_TIMEOUT_MS`	`CONNECTION_TIMEOUT_MS`
`TORCH_READ_TIMEOUT_MS`	`READ_TIMEOUT_MS`

URL, ACCESS_KEY, and SECRET_KEY are unchanged.

Step 3: Error Handling

Before:

from acceldata_sdk.errors import APIError, TorchSdkException

After:

from acceldata.exceptions import APIError, ApiException, AcceldataSdkException

Legacy	New	When Raised
`APIError`	`APIError` (and status-specific subclasses such as `NotFoundError`)	ADOC returned a non-2xx HTTP response.
—	`ApiException`	Network or transport failure before a response.
`TorchSdkException`	`AcceldataSdkException`	Invalid SDK usage or workflow failure (for example, a policy result with `FailOnError`).

Step 4: Pipelines

The largest behavioral change is the resource wrapper pattern. Legacy TorchClient methods returned Pipeline / PipelineRun objects that carried a hidden client reference and exposed methods like create_job and create_span directly on the run.

The new SDK returns PipelineResource and PipelineRunResource wrappers around generated API models. Chaining is similar, but types and some method names differ.

Create a Pipeline

Legacy	New
`CreatePipeline`	`CreatePipelineInputRequest`
`PipelineMetadata`	`Meta` (`acceldata.models.api.pipeline.meta.Meta`)
`client.create_pipeline(CreatePipeline(...))` → `Pipeline`	`client.create_pipeline(CreatePipelineInputRequest(...))` → `PipelineResource`

Before:

from acceldata_sdk.models.pipeline import CreatePipeline, PipelineMetadata pipeline = client.create_pipeline( CreatePipeline( uid="my_pipeline", name="My pipeline", meta=PipelineMetadata(owner="team-a", team="data", codeLocation="..."), ) )

After:

from acceldata.models.api.pipeline.meta import Meta from acceldata.models.sdk.pipeline.create_pipeline_input_request import CreatePipelineInputRequest pipeline = client.create_pipeline( CreatePipelineInputRequest( uid="my_pipeline", name="My pipeline", meta=Meta(owner="team-a", team="data", code_location="..."), ) ) # Inspect API-shaped data: print(pipeline.to_dict())

CreatePipelineInputRequest accepts both Meta with code_location and legacy-style objects that expose codeLocation; the SDK normalizes either form.

Load and List Pipelines

Legacy	New
`client.get_pipeline(uid)` → `Pipeline`	`client.get_pipeline(uid)` → `PipelineResource`
`client.get_pipelines()`	`client.get_pipelines()` → `PipelinesListingResponse`
—	`client.get_pipeline_res(uid)`
—	`client.delete_pipeline(pipeline_id)`
—	`client.update_pipeline(pipeline)`
—	`client.replace_pipeline_tags(identity, tags)`

Pipeline Runs

Legacy	New
`pipeline.create_pipeline_run(...)` on `Pipeline`	`pipeline.create_pipeline_run(...)` on `PipelineResource` → `PipelineRunResource`
`client.get_pipeline_run(pipeline_run_id=..., continuation_id=..., pipeline_id=...)`	Same keyword-only lookup on `AdocClient`; returns `PipelineRunResource`
`run.create_job(CreateJob(...))`	`run.create_job(CreateJobInput(...))` on `PipelineRunResource`
`run.create_span(...)`, `run.get_root_span()`	`run.create_root_span()`, `run.get_root_span()`, `SpanResource` helpers
`client.get_spans(pipeline_run_id)`	`client.get_spans(pipeline_run_id)` → `SpansResponse`
—	`client.get_span(pipeline_run_id, span_identity)`
—	`client.create_span(...)`, `client.get_child_spans(...)`

Continuation IDs for multi-stage workflows are unchanged in semantics.

Jobs and Lineage References

Legacy	New
`CreateJob`	`CreateJobInput`
`Node(asset_uid="DS.table")` or `Node(job_uid="...")`	`JobInputOutputRef(asset_uid="DS.table")` or `JobInputOutputRef(job_uid="...")`
`JobMetadata`	`Meta`

Before:

from acceldata_sdk.models.job import CreateJob, Node, JobMetadata run.create_job( CreateJob( uid="extract_job", name="Extract", inputs=[Node(asset_uid="WAREHOUSE.db.schema.table")], outputs=[Node(job_uid="transform_job")], meta=JobMetadata(owner="etl", team="data", codeLocation="..."), bounded_by_span=True, span_uid="extract_span", ) )

After:

from acceldata.models.api.pipeline.meta import Meta from acceldata.models.sdk.pipeline import CreateJobInput, JobInputOutputRef run.create_job( CreateJobInput( uid="extract_job", name="Extract", inputs=[JobInputOutputRef(asset_uid="WAREHOUSE.db.schema.table")], outputs=[JobInputOutputRef(job_uid="transform_job")], meta=Meta(owner="etl", team="data", code_location="..."), bounded_by_span=True, span_uid="extract_span", ) )

Spans and Events

Legacy	New
`acceldata_sdk.models.span_context.SpanContext`	`acceldata.models.sdk.pipeline.span_resource.SpanResource`
`acceldata_sdk.events.generic_event.GenericEvent`	`acceldata.models.sdk.pipeline.events.GenericEvent`
`acceldata_sdk.events.log_events.LogEvent`	`acceldata.models.sdk.pipeline.events.LogEvent`

Step 5: Catalog — Datasources, Assets, Profiling

Datasources

Legacy	New
`client.get_datasource(name, properties=False)`	`client.get_datasource(name, properties=False)` → `DatasourceResource`
`client.get_datasource_by_id(id, properties=False)`	`client.get_datasource_by_id(id, properties=False)` → `DatasourceResource`
`client.get_datasources(type=...)`	`client.get_datasources(type=...)`
`client.get_all_datasources()`	`client.get_all_datasources()` → `list[DatasourceResource]`
`client.start_crawler(name)`	`client.start_crawler(name)`
`client.get_crawler_status(name)`	`client.get_crawler_status(name)`

AssetSourceType moved to acceldata.models.sdk.catalog.asset_source_type.

Assets

Legacy	New
`client.get_asset(identifier)` → asset object with methods	`client.get_asset(identifier)` → `AssetResource`
`asset.profile_asset(...)`, `asset.sample_data()` on model	`client.profile_asset(...)`, `client.sample_data(...)`, or `AssetResource` methods
`client.get_asset_types()`	`client.get_asset_types()` / `client.get_all_asset_types()`
`client.get_profile_status(asset_id, req_id)`	`client.get_profile_status(asset_id, req_id)` → `ProfileRequestResource`
—	`client.get_asset_tags`, `add_asset_tag`, `get_asset_labels`, `add_asset_labels`, `get_asset_metadata`, `add_custom_metadata`

Step 6: Policies and Rule Execution

Most policy flows map one-to-one, but parameter names and execution scoping changed.

Fetch and List Policies

Legacy	New
`client.get_policy(type=PolicyType.DATA_QUALITY, identifier="...")`	`client.get_policy(PolicyType.DATA_QUALITY, "...")`, or typed `get_dq_policy_resource` / `get_recon_policy_resource` / `get_cadence_policy_resource`
`client.list_all_policies(filter, page=..., size=...)`	`client.list_all_policies(policy_filter=..., page=..., size=...)`
`client.policy_executions(identifier, RuleType.DATA_QUALITY, ...)`	Same; `RuleType` enum values differ (see below)

Import paths:

# Legacy from acceldata_sdk.constants import PolicyType, FailureStrategy, RuleExecutionStatus from acceldata_sdk.models.ruleExecutionResult import PolicyFilter, RuleType # New from acceldata.models.sdk.catalog import PolicyType, PolicyFilter, RuleType, RuleExecutionStatus from acceldata.models.sdk.catalog.executor import FailureStrategy

RuleType in the new SDK uses enum member names (for example, RuleType.DATA_QUALITY) for query parameters; legacy used wire strings such as 'DATA-QUALITY'. PolicyType wire values (DATA-QUALITY, RECONCILIATION, DATA_CADENCE) are unchanged for get_policy.

Execute a Policy

Legacy	New
`execute_policy(policy_type, policy_id, sync=..., incremental=..., failure_strategy=..., pipeline_run_id=..., policy_execution_request=...)`	`execute_policy(rule_type, rule_id, policy_execution_request, *, sync=..., failure_strategy=..., pipeline_run_id=..., sleep_interval=..., total_retries=..., transient_retry=...)`
`incremental=True`	`PolicyExecutionInput(executionType=PolicyExecutionType.INCREMENTAL)`
`incremental=False` (full run)	`PolicyExecutionInput(executionType=PolicyExecutionType.FULL)`
Optional `policy_execution_request`	Required `policy_execution_request` (use `PolicyExecutionInput` for typical cases)
Returns execution handle / result inline when `sync=True`	Returns `Executor`; call `executor.get_result()` / `get_status()` when `sync=False`

Before (full run, synchronous):

from acceldata_sdk.constants import PolicyType, FailureStrategy result = client.execute_policy( PolicyType.DATA_QUALITY, policy_id=123, sync=True, incremental=False, failure_strategy=FailureStrategy.FailOnError, pipeline_run_id=run_id, )

After (equivalent):

from acceldata.models.sdk.catalog import PolicyExecutionType, RuleType from acceldata.models.sdk.catalog.policy_execution_request import PolicyExecutionInput from acceldata.models.sdk.catalog.executor import FailureStrategy executor = client.execute_policy( RuleType.DATA_QUALITY, 123, PolicyExecutionInput(executionType=PolicyExecutionType.FULL), sync=True, failure_strategy=FailureStrategy.FailOnError, pipeline_run_id=run_id, ) # When sync=True, executor already holds the terminal result; you can also call: # result = executor.get_result()

Asynchronous execution:

executor = client.execute_policy( RuleType.DATA_QUALITY, 123, PolicyExecutionInput(executionType=PolicyExecutionType.FULL), sync=False, ) status = executor.get_status() result = executor.get_result(failure_strategy=FailureStrategy.DoNotFail)

Status, Results, and Per-Type Helpers

Legacy	New
`get_policy_status(policy_type, execution_id)`	`get_policy_execution_status(rule_type, execution_id)`
`get_policy_execution_result(policy_type, execution_id, failure_strategy=...)`	`get_policy_execution_result(rule_type, execution_id, failure_strategy=...)`
`get_dq_rule_result` / `get_reconciliation_rule_result` / `get_freshness_rule_result` → `RuleResult`	Same method names; return `ExecutionResult` (typed API models)
`execute_dq_rule`, `execute_reconciliation_rule`, `execute_freshness_rule`	Still available on `AdocClient`; prefer unified `execute_policy`
`cancel_rule_execution`, `enable_rule`, `disable_rule`	Unchanged names

Step 7: Pipeline Tags

Pipeline tag types moved to generated models:

from acceldata.models.api.pipeline.tag import Tag Tag(name="env:prod", displayName="Environment: Production")

Removed or Not Exposed on AdocClient

The following legacy TorchClient methods are not on AdocClient. Migrate only if you still depend on them.

Legacy Method	Notes
`get_supported_sdk_versions()`	Removed.
`get_torch_version()`	Removed.
`do_version_check` constructor flag	Removed.
`get_property_templates()`	Was unimplemented in the legacy SDK, and remains unimplemented here.
`get_all_source_types()`	Not on the client; use asset or datasource type listing APIs.
`get_tags()` (global tag list)	Use asset tag APIs or pipeline tag replacement.
`get_analysis_pipeline(id)`	Available on `client.generic_service.get_analysis_pipeline(pipeline_identifier)` — accepts a numeric ID or name; returns `AnalyticsPipeline`.
`get_connection_types()`	Available on `client.generic_service.get_connection_types()`.
`create_connection()`, `check_connection()`	Deprecated; moved to the management service.

Transient HTTP retries on policy executions are new in acceldata-sdk-python.

What's Next

After you complete this section, explore:

Migration Guide and New Features Overview – Review the support policy and the full migration checklist.
Acceldata SDK for Python (acceldata-sdk-python) – Review installation, client setup, and error handling in the new SDK.

Last updated on

Was this page helpful?