Pyapacheatlas

Release Notes

New Features
* Supporting Azure-Identity package
* You may now, optionally, install the azure-identity package and use, for example, DefaultAzureCredential or AzureCLICredential to authenticate.

from pyapacheatlas.core.client import PurviewClient
from azure.identity import DefaultAzureCredential, AzureCliCredential

cred = DefaultAzureCredential() For Managed Identity or Service Principal auth
cred = AzureCLICredential() For my personal credentials

client = PurviewClient(
account_name=os.environ.get("NEW_PURVIEW_NAME", ""),
authentication=cred
)

* Adding Purview Column Mapping Lineage to Excel Reader
* Users may now specify how columns map to each other for a given process.
* This requires that the process has a columnMapping attribute (and is likely a custom process type).

Breaking Changes
* ColumnLineages tab has been renamed to FineGrainColumnLineage, this was done to help disambiguate when this tab should be used.

Release Notes

New Features
* None

Bug Fixes
* amiket23 added the `includeTermHierarchy=true` query parameter to `PurviewClient.import_terms` which is currently undocumented 🙄
* Updated the `pyapacheatlas.core.util.GuidTracker` to return a `string` rather than an `int`. This caused an issue in working with relationship attributes in the bulk excel upload.

Release Notes

New Features
* **SDK:**
* Added support for assigning glossary terms programmatically with the `AtlasClient.assignTerm`. This allows for assigning a single term to an existing entity without affecting any other term attached to the entity.
* Added a `batch_size` parameter to the `AtlasClient.upload_entities` to support batching large uploads (anecdotally > 4,000 entities) into groups with related dependencies in the same group being uploaded.
* **Excel:**
* Added support for Classification Defs as a new tab (not supporting Purview Classification rules).
* Added support for including Entity Def Super types in the EntityDefs tab.
* Supporting adding glossary terms to a bulk upload through the `[Relationship] meanings` column.


Bug Fixes
* None

Release Notes

New Features
* Added support for partial entity updates from AtlasClient. You can provide either a guid and a single attribute (in a dict) or a qualified name and type along with a dict of multiple attributes.
* Added support in the Readers for Experts and Owners in the bulk entities update (for Purview). You can now add an experts or owners column and will create the appropriate data structure for upload. Only supports using Azure Active Directory object ids currently.
* Made AtlasEntity (and AtlasProcess) `to_json` smarter by recognizing when you provide a guid.
* If you provide a guid (should be a negative number) it expects that you are going to use it as part of a bulk upload.
* If you do not provide a guid it expects that the json produced is being used as a reference rather than as part of a bulk upload.

Release Notes

Bug Fixes
* bsherwin fixed 89 (with PR 91) causing running PyApacheAtlas on PowerShell to fail when doing an authentication.

Release Notes

New Features
* Supporting bulk import and export of glossary terms through CSVs.
* Bulk Import terms with `import_terms` from a CSV.
* Check bulk import status with `import_terms_status`.
* Bulk Export terms with `export_terms` and write to a CSV.
* This should not be confused with creating terms programmatically with `upload_terms`.
* This should not be conused with getting the glossary terms for the entire glossary with `get_glossary` or `get_glossary_term`.
* Supporting the LineageREST APIs for Purview (AnalyticJeremy)
* Get the json lineage data for a given guid.
* Page through the lineage for complex lineage data.

Improved Features
* Improved Azure Purview advanced search by abstracting away pagination (slyons)
* When using the ExcelReader/Reader classes to parse the bulk entities for upload, it will not set propagate = False on classifications to avoid cascading classifications.
* Advanced users can still change the propagation after parsing and before upload.

Bug Fixes
* Fixed a bug where the PurviewClient.is_purivew was not being set properly due to incorrect regex pattern.
* When handling a Json Decode error, the error was being raised again but incorrectly. Changed it to a ValueError which returns the body of the message that was trying to be parsed.