Skip to content

Latest commit

 

History

History
174 lines (105 loc) · 9.4 KB

CHANGELOG.md

File metadata and controls

174 lines (105 loc) · 9.4 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

[4.0.5] - 2024-11-06

  • Dataframes to be registered when using compare_two_records, to avoid problems with data typing (because the input data can have an explicit schema) (#2493)

[4.0.4] - 2024-10-13

Added

  • cluster_pairwise_predictions_at_multiple_thresholds to more efficiently cluster at multiple thresholds (#2437)

Fixed

  • Fixed issue with profile_columns using latest Altair version (#2466)

[4.0.3] - 2024-09-19

Added

  • Cluster without linker by @RobinL in #2412
  • Better autocomplete for dataframes by @RobinL in #2434

[4.0.2] - 2024-09-19

Added

  • Match weight and m and u probabilities charts now have improved tooltips (#2392)
  • Added new AbsoluteDifferenceLevel comparison level for numerical columns (#2398)
  • Added new CosineSimilarityLevel and CosineSimilarityAtThresholds for comparing array columns using cosine similarity (#2405)
  • Added new ArraySubsetLevel for comparing array columns (#2416)

Fixed

  • Fixed issue where ColumnsReversedLevel required equality on both columns (#2395)

[4.0.1] - 2024-09-06

Added

  • When using DuckDB, you can now pass duckdb.DuckDBPyRelations as input tables to the Linker (#2375)
  • It's now possible to fix values for m and u probabilities in the settings such that they are not updated/changed during training. (#2379)
  • All charts can now be returned as vega lite spec dictionaries (#2361)

Fixed

  • Completeness chart now works correctly with indexed columns in spark (#2309)
  • Completeness chart works even if you have a source_dataset column (#2323)
  • SQLiteAPI can now be instantiated without error when opting not to register custom UDFs (#2342)
  • Splink now runs properly when working in read-only filesystems (#2357)
  • Infinite Bayes factor no longer causes SQL error in Spark (#2372)
  • splink_datasets is now functional in read-only filesystems (#2378)

4.0.0 - 2024-07-24

Major release - see our blog for what's changed

3.9.15 - 2024-06-18

Fixed

  • Activates higher_is_more_similar kwarg in cl.distance_function_at_thresholds, see here
  • linker.save_model_to_json() now correctly serialises tf_minimum_u_value and reloads. See here.
  • Performance improvements on code geenration, see here

3.9.14 - 2024-03-25

Fixed

  • IndexError: List index out of range error due to API change SQLGlot>=23.0.0, see here

Added

  • Ability to override detection of exact match level for tf adjustments. See here for example.
  • Added method for computing graph metrics (#2027)

3.9.13 - 2024-03-04

  • Support for Databricks Runtime 13.x+

Fixed

  • Bug that prevented sqlglot <= 17.0.0 from working properly (#1996)
  • Fixed issues relating to duckdb 0.10.1 (#1999)
  • Update sqlglot compatibility to support latest version (#1998)

3.9.12 - 2024-01-30

Fixed

  • Support sqlalchemy >= 2.0.0 (#1908)

3.9.11 - 2024-01-17

Added

  • Ability to block on array columns by specifying arrays_to_explode in your blocking rule. (#1692)
  • Added ability to sample by density in cluster studio by @zslade in (#1754)

Changed

  • Splink now fully parallelises data linkage when using DuckDB (#1796)

Fixed

  • Allow salting in EM training (#1832)

3.9.10 - 2023-12-07

Changed

  • Remove unused code from Athena linker (#1775)
  • Add argument for register_udfs_automatically (#1774)

Fixed

  • Fixed issue with _source_dataset_col and _source_dataset_input_column (#1731)
  • Delete cached tables before resetting the cache (#1752

3.9.9 - 2023-11-14

Changed

  • Upgraded sqlglot to versions >= 13.0.0 (#1642)
  • Improved logging output from settings validation (#1636) and corresponding documentation (#1674)
  • Emit a warning when using a default (i.e. non-trained) value for probability_two_random_records_match (#1653)

Fixed

  • Fixed issue causing occasional SQL errors with certain database and catalog combinations (#1558)
  • Fixed issue where comparison vector grid not synced with corresponding histogram values in comparison viewer dashboard (#1652)
  • Fixed issue where composing null levels would mistakenly sometimes result in a non-null level (#1672)
  • Labelling tool correctly works even when offline (#1646)
  • Explicitly cast values when using the postgres linker (#1693)
  • Fixed issue where parameters to completeness_chart were not being applied (#1662)
  • Fixed issue passing boto3_session into the Athena linker (#1733)

3.9.8 - 2023-10-05

Added

  • Added ability to delete tables with Spark when working in Databricks (#1526)

Changed

  • Re-added support for python 3.7 (specifically >= 3.7.1) and adjusted dependencies in this case (#1622)

Fixed

  • Fix behaviour where using to_csv with Spark backend wouldn't overwrite files even when instructed to (#1635)
  • Corrected path for Spark .jar file containing UDFs to work correctly for Spark < 3.0 (#1622)
  • Spark UDF damerau_levensthein is now only registered for Spark >= 3.0, as it is not compatible with earlier versions (#1622)