v3.9.15

What's Changed

• Document first-time developer setup, add conda option by @zmbc in #2083
• fix links by @RobinL in #2097
• Add dirty reload for much faster updates by @RobinL in #2096
• Add documentation for spellchecker and spellcheck docs by @zslade in #2025
• Add graph definition to docs by @zslade in #1979
• Minor fixes to spellchecker by @zslade in #2113
• Changing args as kwargs by @jlb52 in #2116
• Update threshold_selection_tool.json by @aalexandersson in #2120
• Fix broken link by @samnlindsay in #2098
• added tf_minimum_u_value to as_dict method by @aymonwuolanne in #2122
• Fix a bug in conda script and make minor improvements to quickstart by @zmbc in #2125
• Fix documentation Github Action for forks by @zmbc in #2126
• Add better check for whether conda is already installed by @zmbc in #2130
• Update PULL_REQUEST_TEMPLATE.md with spellchecker tick box by @zslade in #2128
• Clusters topic guide by @zslade in #1883
• Splink blog March 2024: Splink 3 update and Splink 4 development announcement by @RobinL in #2081
• Fix link to linter by @RobinL in #2121
• add probabilistic section to graphs definitions by @RossKen in #2137
• Update PULL_REQUEST_TEMPLATE.md by @zslade in #2138
• Minor bug in filtering predict table by @samnlindsay in #2152
• Update documentation on settings validation in response to code changes by @ThomasHepworth in #2149
• Remove reference to github action that will not come to be by @zslade in #2163
• Fixing spurious error messages with Databricks enable_splink by @aymonwuolanne in #2159
• Fix Splink 4 blog post link by @probjects in #2172
• Make spellcheck work cross-platform by @zmbc in #2131
• add marie curie by @RobinL in #2201
• Fix bug giving warning messages in term_frequencies.py by @DavidFrenchSG in #2204
• Fix lint by @RobinL in #2205
• Improve performance of SQL generation by using deepcopy less by @RobinL in #2212
• 3.9.15 release by @RobinL in #2213

New Contributors

• @zmbc made their first contribution in #2083
• @jlb52 made their first contribution in #2116
• @aalexandersson made their first contribution in #2120
• @probjects made their first contribution in #2172
• @DavidFrenchSG made their first contribution in #2204

Full Changelog: v3.9.14...v3.9.15

v4.0.0.dev6

What's Changed

• Group linker functions thematically by @RobinL in #2192

Full Changelog: v4.0.0.dev5...v4.0.0.dev6

v4.0.0.dev5

What's Changed

• Further tidying of blocking analysis - better typing by @RobinL in #2186
• Misc tidying by @ADBond in #2182
• Consolidate accuracy 2 by @RobinL in #2187
• Move all backend functions to internals/ directory by @RobinL in #2189

Full Changelog: v4.0.0.dev4...v4.0.0.dev5

v4.0.0.dev4

What's Changed

• Simple extension to term frequency adjustments for inexact matches by @samkodes in #2020
• Update bug report template by @ADBond in #2073
• update colab links by @RobinL in #2080
• Fix mkdocs rendering symbols in notebook code by @ADBond in #2033
• Enqueue and compute methods by @RobinL in #2086
• rm deprecated action and bash scripts by @ThomasHepworth in #2094
• Fix sqlglot>=23.0.0 issue by @RobinL in #2079
• 3.9.14 release by @RobinL in #2095
• Document first-time developer setup, add conda option by @zmbc in #2083
• fix links by @RobinL in #2097
• Add dirty reload for much faster updates by @RobinL in #2096
• Remove _pipeline from linker and refactor CTE pipeline by @RobinL in #2069
• Splink 4 blocking rule/blocking rule creator fixes by @RobinL in #2103
• remove deprecated and outdated code by @RobinL in #2107
• Further br fixes by @RobinL in #2106
• Fix find matches input column by @RobinL in #2109
• tf_logic_simplify by @RobinL in #2110
• Add documentation for spellchecker and spellcheck docs by @zslade in #2025
• Add graph definition to docs by @zslade in #1979
• Minor fixes to spellchecker by @zslade in #2113
• Changing args as kwargs by @jlb52 in #2116
• Update threshold_selection_tool.json by @aalexandersson in #2120
• Fix broken link by @samnlindsay in #2098
• added tf_minimum_u_value to as_dict method by @aymonwuolanne in #2122
• Stricter mypy checks by @ADBond in #2108
• Merge 3 4 2123 by @RobinL in #2124
• Fix a bug in conda script and make minor improvements to quickstart by @zmbc in #2125
• Refactor and simplify how TF adjustments are made in _find_new_matches_mode and _compare_two_records_mode by @RobinL in #2111
• Faster tests: Split out tests into separate backends and use altair 5.3.0 by @RobinL in #2117
• Fix documentation Github Action for forks by @zmbc in #2126
• Add better check for whether conda is already installed by @zmbc in #2130
• Restore Settings Validation (Splink 4) by @ADBond in #2127
• Update PULL_REQUEST_TEMPLATE.md with spellchecker tick box by @zslade in #2128
• Clusters topic guide by @zslade in #1883
• Splink blog March 2024: Splink 3 update and Splink 4 development announcement by @RobinL in #2081
• Merge/splink 3 to 4 by @RobinL in #2134
• Fix link to linter by @RobinL in #2121
• add probabilistic section to graphs definitions by @RossKen in #2137
• Update PULL_REQUEST_TEMPLATE.md by @zslade in #2138
• Remove flags from block_using_rules_sqls logic ( _find_new_matches_mode and _compare_two_records_mode etc.) by @RobinL in #2129
• Merge/splink 3 to 4 by @RobinL in #2148
• Process input tables simplification by @RobinL in #2143
• Type decorator by @ADBond in #2151
• Allow df_concat to be created without a linker by @RobinL in #2144
• Specify generic types by @ADBond in #2153
• switch to ruff by @RobinL in #2156
• Mark spark tests by @ADBond in #2161
• Fix bugs in calculations for true negatives when using accuracy _from_column functions by @RobinL in #2150
• Move missingness chart out of linker and move profile_columns to splink.exploratory by @RobinL in #2157
• Test pythons > 3.9 in CI by @ADBond in #2164
• Adding type-hints, part 1 by @ADBond in #2169
• More type hints - remaining incomplete definitions by @ADBond in #2171
• Estimate u - default value warning by @ADBond in #2181
• Refactor blocking to not need linker by @RobinL in #2180

New Contributors

• @samkodes made their first contribution in #2020
• @jlb52 made their first contribution in #2116
• @aalexandersson made their first contribution in #2120

Full Changelog: v4.0.0.dev3...v4.0.0.dev4

v3.9.14

What's Changed

• Update u probability formula and example in fellegi_sunter.md by @jacuna88 in #2036
• Splink 3: Increment minimum python version from 3.7 to 3.8 by @RobinL in #2031
• Make graph metrics public by @zslade in #2027
• Add PUDL to list of use cases by @zaneselvans in #2044
• Threshold selection tool by @samnlindsay in #2003
• Simple extension to term frequency adjustments for inexact matches by @samkodes in #2020
• Update bug report template by @ADBond in #2073
• Fix mkdocs rendering symbols in notebook code by @ADBond in #2033
• rm deprecated action and bash scripts by @ThomasHepworth in #2094
• Fix sqlglot>=23.0.0 issue by @RobinL in #2079
• 3.9.14 release by @RobinL in #2095

New Contributors

• @jacuna88 made their first contribution in #2036
• @zaneselvans made their first contribution in #2044
• @samkodes made their first contribution in #2020

Full Changelog: v3.9.13...v3.9.14

v4.0.0.dev3

update release workflow

v3.9.13

What's Changed

• Mkdocs preprocess hooks by @ADBond in #1913
• Docs workflow - build and check links on PRs by @ADBond in #1915
• minor homepage tweaks by @RossKen in #1919
• Model evaluation guide by @RossKen in #1916
• convert accuracy metrics to float by @ThomasHepworth in #1893
• Use CASE instead of bool to float casting in truth_space_table by @cinnq346 in #1928
• add NICD x Gateshead use case by @RossKen in #1931
• Update venv to use a custom name and edit errors by @ThomasHepworth in #1918
• Add comparison level validation check by @ThomasHepworth in #1926
• Update load settings and make it the defacto load logic by @ThomasHepworth in #1921
• Cast row_count as float8 in truth_table by @cinnq346 in #1936
• Trim documentation dependencies by @ADBond in #1917
• Fix docs build by @ADBond in #1953
• Implement is_bridge edge metric by @ADBond in #1894
• add parameter to anonymise waterfall chart by @RossKen in #1938
• Clarify naming of hide_details on waterfall chart by @RobinL in #1963
• (Try to) fix css styling for the summary/details tags in .vega-embed by @RobinL in #1966
• Accuracy chart - altair bug by @RossKen in #1965
• use .sql not .execute by @RobinL in #1952
• CI - update splink4 'to-merge' branch by @ADBond in #1984
• sqlglot.parse_one - use read keyword argument by @ADBond in #1996
• Edge evaluation guide by @RossKen in #1927
• Adding support for DBR 13.x and 14.x by @boobay in #1973
• SplinkDataFrame metadata in clustering + metrics by @ADBond in #1981
• Refine additional installs in the readme by @ThomasHepworth in #2007
• compute_graph_metrics - compute what we can without igraph by @ADBond in #1982
• Add a section on dependency management within Splink by @ThomasHepworth in #1985
• Spell check single files by @ThomasHepworth in #2000
• Change file name to reflect graph naming conventions by @zslade in #2015
• Relax Splink 3 Dependency Requirements - demonstrate all tests pass with latest sqlglot by @RobinL in #1998
• Fix test failures in duckdb 0.10.0 by @RobinL in #1999
• v3.9.13 release by @RobinL in #2024

New Contributors

• @cinnq346 made their first contribution in #1928
• @boobay made their first contribution in #1973

Full Changelog: v3.9.12...v3.9.13

v3.9.12

What's Changed

• Update mkdocs.yml by @RossKen in #1858
• Add support for SaltedBlockingRule for EM training (again) by @RobinL in #1853
• Update performance.md by @DanielOX in #1865
• add initial usecases to homepage by @RossKen in #1864
• fix edit link by @RossKen in #1866
• Minor correction to docstring by @zslade in #1867
• Fixes #1872 Update deduplicate_1k_synthetic.ipynb to fix spark error by @w2o-hbrashear in #1873
• Document duckdb parallelism by @RobinL in #1877
• Ethics Blog & blog docs by @RossKen in #1849
• Initial evaluation topic guide by @RossKen in #1876
• Update 2024-01-25-ethics.md by @RossKen in #1879
• add datafirst datasets to use cases by @RossKen in #1880
• Minor tweaks to sampling by cluster size by @zslade in #1829
• fix broken link by @RossKen in #1900
• Update sampling logic for density by @zslade in #1831
• return data class instead of dictionary by @zslade in #1887
• CI link-checking + fixed links by @ADBond in #1902
• SQLAlchemy 1.x and 2.x compatibility: Use explicit transactions, remove sqlalchemy version constraint by @RobinL in #1908
• Type hinting and variable renaming (mypy conformance stage 1) by @ADBond in #1780
• 3.9.12 Release by @RobinL in #1911

New Contributors

• @DanielOX made their first contribution in #1865
• @w2o-hbrashear made their first contribution in #1873

Full Changelog: v3.9.11...v3.9.12

v3.9.11

What's Changed

• Faster duckdb train u by @RobinL in #1800
• remove reference to deleted token by @RossKen in #1802
• material by mkdocs upgrade by @RossKen in #1803
• authors yml format fix by @RossKen in #1804
• Fix broken links by @RossKen in #1805
• Cluster studio sample by density by @zslade in #1754
• Cluster metrics - node degree + cluster centralisation by @ADBond in #1806
• Parallelise duckdb resulting in e.g. 2-4x speedup on 6 core machine by @RobinL in #1796
• Remove brittleness of convergence test by @RobinL in #1798
• Enable salting for EM training by @RobinL in #1832
• Fix linting workflows by @ADBond in #1836
• Refactor of 1664: add ability to do efficient blocking based on list/array intersections by @RobinL in #1692
• changelog: note add ability to block on array columns to by @RobinL in #1847
• 3.9.11 release by @RobinL in #1848

Full Changelog: v3.9.10...v3.9.11

v3.9.10

What's Changed

• Bump aiohttp from 3.8.5 to 3.8.6 by @dependabot in #1741
• Bump urllib3 from 1.26.16 to 1.26.18 by @dependabot in #1656
• BlockingRule: Refactor to enable better iteration by @RobinL in #1701
• Fix issue with _source_dataset_col and _source_dataset_input_column by @RobinL in #1731
• [MAINT] Improve speed of tests by @RobinL in #1736
• When you go to open an issue, add a link to discussions below our issue templates by @RobinL in #1746
• Finds blocking rules which return a comparison count below a given threshold by @RobinL in #1665
• Compute the cost of combinations of blocking rules by @RobinL in #1667
• Fix docs build by @ADBond in #1748
• [BUG] Delete cached tables before resetting the cache by @ThomasHepworth in #1752
• Automatically detect blocking rules for prediction and blocking rules for EM training by @RobinL in #1668
• added argument for register_udfs_automatically by @JonathanLaidler in #1774
• Make notebook tests run faster by @RobinL in #1772
• Improve speed of link only sample test by @RobinL in #1773
• Remove unused code and improve the Athena Linker by @ThomasHepworth in #1775
• Fixes to _compute_cluster_metrics by @zslade in #1763
• Add Mypy setup to pyproject.toml by @ADBond in #1779
• Introduce a ColumnTreeBuilder to aid in the construction of our column ASTs by @ThomasHepworth in #1757
• [MAINT] Revamp the settings validation steps by @ThomasHepworth in #1764
• Fix brl comp test by @ADBond in #1784
• add cs awards to readme by @RossKen in #1792
• v3.9.10 by @RossKen in #1790
• Blog - Dec 2023 by @RossKen in #1791

New Contributors

• @dependabot made their first contribution in #1741
• @JonathanLaidler made their first contribution in #1774

Full Changelog: v3.9.9...v3.9.10