Ayush Chaurasia cb57b7655e feat(python): add Permutation.from_table() with chainable HF/torch-style API ...

Hides the two-step `permutation_builder(t).shuffle().execute()` /
`Permutation.from_tables(t, perm_tbl)` dance behind a single chainable
entry point that ML engineers already expect from HuggingFace and
PyTorch:

  perm = Permutation.from_table(table).shuffle(seed=42)

Builder operations (shuffle, filter, split_*) are accumulated lazily and
the underlying permutation table is built only once on first read.
Subsequent read / transform / format calls are forwarded transparently
to the materialized Permutation.

Closes lancedb/lancedb#3244

2026-04-29 22:21:59 +05:30

..

lancedb

feat(python): add Permutation.from_table() with chainable HF/torch-style API

2026-04-29 22:21:59 +05:30

tests

feat(python): add Permutation.from_table() with chainable HF/torch-style API

2026-04-29 22:21:59 +05:30