0.12.15

Features

• feat: content-aware column widths, width estimation tests, and screenshot comparison infra (#606)
• feat: add opt-in compact_number displayer (#597) (#601)
• Add CLAUDE.md for Claude Code development guidance (#598)
• feat: add color_static color rule; use it to highlight join-key columns in compare (#594)
• feat: add /load_compare endpoint for col_join_dfs diff styling (#593)

Bug Fixes

• fix: set defaultMinWidth to prevent narrow columns (#599) (#600)

Install

pip install buckaroo==0.12.15

0.12.14

Bug Fixes

• fix: Playwright row count off-by-one with pinned columns (#590)
• Fix BuckarooCompare for arbitrary join keys (#589)
• Fix left-pinned index column for summary stats (#587)
• fix: harden release workflow — tag-based versioning, no push to main (#581)

Install

pip install buckaroo==0.12.14

0.12.13

Bug Fixes

• Fix left-pinned index column for summary stats (#587)
• fix: harden release workflow — tag-based versioning, no push to main (#581)
• fix: handle zero PRs in release notes; extract CHANGELOG updater to script (#580)
• fix: plain release notes fallback + sync version to 0.12.11 (#579)
• fix: MCP + server reliability, observability, and hardening (#578)

Other

• for the PR (#577)

Install

pip install buckaroo==0.12.13

0.12.4

Fixes #422 which makes buckaroo pandera function again.

Lazy Infinite Polars Buckaroo

This release adds Lazy Infinite Buckaroo

• when reading Polars LazyFrames for LazyInfinitePolarsBuckaroo, the entire dataframe is no longer loaded into memory
• Adds the execution framework that reliably executes code in separate processes with timeouts. This gives much better reliability when running on large dataframes in an interactive environment
• Uses the Execution Framework for summary stats for the LazyInfinitePolarsBuckaroo Widget.
• adds read() that works on csvs, parquet, and many files to give a LazyInfinitePolarsBuckaroo widget that won't blow out your memory.

Polars Bug Fixes for 0.11.2

• Release 0.11.2
• Sort broken for polars https://github.com/paddymul/buckaroo/issues/410
• Polars doesn't display row indexes properly for PolarsBuckarooInfiniteWidget https://github.com/paddymul/buckaroo/issues/408
• Buckaroo isn't the default display method for Polars in Marimo https://github.com/paddymul/buckaroo/issues/409

0.11.1 Bug fixes

• Release 0.11.1
• Fixes sort for pandas with new column naming
• Fixes lowcode UI with new column naming
• More reliable default analysis for pandas
• Start of the ddd gallery for showing tricky dataframes
• Start of Patrick big file blog post

0.11.0 Multi Index support

0.11.0 release notes

Multi-Index support

This release now supports pandas multi indexes in the main index and columns. Multi index Columns appear nested like they do in the pandas repr. Named indexes for columns and main index are also supported and displayed.

A lot more typechecking

Buckaroo has added a lot of typing for the pluggable_analysis_framework. There will be follow on releases with more typing and mypy (or beartype) enabled for CI. These types also extend to styling

More disciplined serialization of dataframes

Up until now, Buckaroo depended on a lot of implicit handling of column names and index names. This interacted poorly with multi-index dataframes in particular. Now, before serialization, every column is renamed to a letter based named between a-zz. The index is treated as a column explicitly renamed to "index". Internally, the pluggable_analysis_framework already dealt mostly with series, now it only deals withs series with explicit orig_col_name, and rewritten_col_name keys.

Extensive tests

There are even more tests now, especially around styling and the pluggable_analysis_framework

Starker colors in table display

Zebra striping is more pronounced. There are thicker borders around index columns

Buckaroo-js-core improvements to handle nested columns

Supporting nested columns required modifications to Buckaroo js core for support.

Introducing the DDD

This release adds the Dastardly Datafarame Dataset. A collection of dataframes that are tricky to display, serialize, or otherwise deal with. Through hard faught experience, Buckaroo has dealt with bugs or edge cases related to each dataframe in the library. At a minimum, all dataframes display in some way unless otherwise noted. Most display with full features, there are a couple of errors, but having a comprehensive test set is a very helpful start.

What's coming next

Nested column support will be used to better display polars structs. There need to be new styling primitives to handle structs (polars see structs as a single entity, but the table needs to treat each struct member as an independent column).

Styling tweaks, there are improvements to be made with some of the combinations of nested and named multi-indexes, and the thick border delineating the index columns.

0.10.5 Pandera Tool

• Release 0.10.5
• Adds Pandera tool which allows easy visualization of validation errors of DataFrames. Live marimo demo https://marimo.io/p/@paddy-mullen/buckaroo-pandera

* Fixes #393 Columns with a period don't display properly

0.10.4 Rendering fixes and improvements

• released 0.10.4
• Much better rendering performance
• No longer run react in Strict Mode, better performance
• Much smarter way of choosing highlighting the active column which took the re-render time from 360ms to 60ms.
• 0.10.3 fixed bug where multiple buckaroo widgets in marimo-wasm would be unstyled