0.15.4

Features

• feat(stats): expose summary-stat cache hit/miss + timing as a structured signal (#944)

Bug Fixes

• fix(stats): surface summary-stat cache write side via telemetry (#951) (#952)

CI/CD

• ci: remove standalone release-npm.yml workflow (#947)

Install

pip install buckaroo==0.15.4

0.15.3

Features

• feat(stats): expose summary-stat cache hit/miss + timing as a structured signal (#944)
• feat(duckdb-node): search — filtered stats/rows + highlight_phrase (#940)

Bug Fixes

• fix(row-cache): clamp negative start in View.rowidsInRange (#942)

Install

pip install buckaroo==0.15.3

0.15.2

Features

• feat(duckdb-node): search — filtered stats/rows + highlight_phrase (#940)
• feat(duckdb-node): histogram summary stat (numeric .01–.99 + categorical) (#939)
• feat(duckdb-node): DuckDB backend for buckaroo-js-core (#930) (#936)

Bug Fixes

• fix(transport): json_columns so native parquet string cells survive decode (#938)
• fix(ci): exclude slow MCP tests from dependency-check workflow (#932)

Refactoring

• refactor(stats): remove buckaroo-managed source materialization (#929)

Other

• docs: plan unified dataframe transport layer (#933)
• build(deps): bump actions/checkout from 6 to 7 (#931)

Install

pip install buckaroo==0.15.2

0.15.1

Features

• feat(perf): BUCKAROO_PERF toggle — instrument stats pipeline (all backends) + first data pull (#926)

Refactoring

• refactor(stats): remove buckaroo-managed source materialization (#929)

Install

pip install buckaroo==0.15.1

0.15.0

Features

• feat(perf): BUCKAROO_PERF toggle — instrument stats pipeline (all backends) + first data pull (#926)

Bug Fixes

• fix(ci): isolate pytest-check-links so Max Versions matrix can collect (#927)
• fix(stats): stream to temp parquet in _create_base_table instead of materializing InMemoryTable (#922)

Install

pip install buckaroo==0.15.0

0.14.20

Features

• feat(stats): perf/memory smoke tests for summary stats (perf_smoke_test) (#917)

Bug Fixes

• fix(stats): stream to temp parquet in _create_base_table instead of materializing InMemoryTable (#922)
• fix(stats): classify distinct_count by approx support across all arrow dtypes (#918) (#919)
• fix(stats): materialize the stat source only for chained sources (#915) (#916)

Other

• Remove all vestiges of dfstats_v1 (full @stat port) (#912)

Install

pip install buckaroo==0.14.20

0.14.19

Features

• feat(stats): perf/memory smoke tests for summary stats (perf_smoke_test) (#917)

Bug Fixes

• fix(stats): classify distinct_count by approx support across all arrow dtypes (#918) (#919)
• fix(stats): materialize the stat source only for chained sources (#915) (#916)
• fix(xorq): materialize source on cold cached load; log cache stats (#910) (#913)

Install

pip install buckaroo==0.14.19

0.14.18

Features

• feat(server): optional dataframe backends — add [server] extra, make polars/xorq additive (#905)

Bug Fixes

• fix(xorq): materialize source on cold cached load; log cache stats (#910) (#913)
• fix(stats): sample the categorical histogram input above 100k rows (#909)
• fix(stats): use approx_nunique for distinct_count in the xorq batch aggregate (#908)
• fix(server): skip /load_expr pipeline for warm sessions; share xorq backend (#896, #899) (#901)

Install

pip install buckaroo==0.14.18

0.14.17

Bug Fixes

• fix(server): skip /load_expr pipeline for warm sessions; share xorq backend (#896, #899) (#901)

Performance

• perf(serialization): project all_stats wire payload to displayed stats only (#880) (#883)

Install

pip install buckaroo==0.14.17

0.14.16

Features

• feat(xorq): load display klasses from <project_root>/display/*.py (#894)

Performance

• perf(serialization): project all_stats wire payload to displayed stats only (#880) (#883)

Chores

• chore: remove erroneous Bloomberg attribution (#898)

Install

pip install buckaroo==0.14.16