Tom Hepworth

Add support for pairwise format of clusters

1

Code that allows users to evaluate their clusters in the context of the `linker.predict()` output. Specifically, this allows users the ability to view the match threshold associated with cluster links....

### What happens? When loading a pandas df with array columns into duckdb, it seems to transform them into strings. So an array `[1234, 5678]` will become -> `"[1234, 5678]"`....

It's late, so I'll update this with some further information tomorrow morning as appropriate. This impacts anywhere that either requests nulls be removed/adjusted or that compares strings. This rather tragically...

`random()` in duckdb doesn't register properly in CTE expressions

1

You can see part of the debate [here](https://github.com/moj-analytical-services/splink/pull/545#issuecomment-1167063184) and my temporary fix [here](https://github.com/moj-analytical-services/splink/blob/86ffa1c4b11ae5fcaac7ac2b919ae1cb60510bb9/splink/linker.py#L833). In short, when creating a column using `random()` in a CTE chain and then attempting to filter...

Spell check single files

2

### Type of PR - [ ] BUG - [x] FEAT - [ ] MAINT - [ ] DOC ### Is your Pull Request linked to an existing Issue or...

Add a section on dependency management within Splink

3

### Type of PR - [ ] BUG - [ ] FEAT - [ ] MAINT - [x] DOC ### Is your Pull Request linked to an existing Issue or...

[MAINT] `linker._intermediate_table_cache` logs multiple entries when creating term frequency tables

2

### What happens? See https://github.com/moj-analytical-services/splink/pull/1061#issuecomment-1546708437. In short, when generating our term frequency tables we are currently recording both the physical and templated names within `linker._intermediate_table_cache`. This in turn makes clean-up...

[MAINT] `linker.py` cleanin/refactoring

3

### Is your proposal related to a problem? Tags onto this conversation about cleaning up our settings class - https://github.com/moj-analytical-services/splink/issues/1669. I think we should sit down and work out how...

[FEAT] Cross-Backend / Lazily Evaluated Comparisons and Blocking Rule builders

8

### Is your proposal related to a problem? See this closed issue for some further background, though it's not necessary - https://github.com/moj-analytical-services/splink/issues/1670. ## The Problem Currently, we face the following...

[MAINT] Settings restructuring - bringing everything into a centralised settings object

1

### Is your proposal related to a problem? This proposal builds upon the discussions in issue https://github.com/moj-analytical-services/splink/issues/1055. ### Describe the solution you'd like As Splink has expanded, some of our...