Expand description
Vortex string array backed by the OnPair short-string
compression library, with cast and filter pushdown.
The default training preset is dict-12 (12 bits per token, dictionary
capped at 4 096 entries). See onpair_compress for the entry point and
OnPairArray for the resulting array type.
Structs§
- Bits
- Code width: the maximum dictionary size is
2^bits. Validated to9..=16at construction, so aBitsalways holds an in-range value. - Config
- Training configuration. See
DEFAULT_CONFIGfor a reasonable starting point. - OnPair
- Zero-sized VTable marker for the OnPair encoding.
- OnPair
Data - Inner data for an OnPair-encoded array.
- OnPair
Metadata - Wire-format metadata persisted alongside the OnPair buffer + slot children.
- OnPair
Slots - OnPair
Slots View - Borrowed view of
OnPairSlots. - Threshold
- Dynamic-threshold sample fraction. Validated to
(0.0, 1.0]at construction, so aThresholdalways holds an in-range value.
Enums§
- OnPair
Error - Error returned by the public training and encoding API.
Constants§
- DEFAULT_
DICT12_ CONFIG - Default OnPair training configuration: 12-bit codes (“dict-12”).
Traits§
- OnPair
Array Ext - Convenience methods on top of the macro-generated
OnPairArraySlotsExt. - OnPair
Array Slots Ext - Typed array accessors for
OnPair.
Functions§
- onpair_
compress - Compress a byte-string accessor (typically a
VarBinArrayorVarBinViewArray). - onpair_
compress_ array - Compress any
ArrayRefwhose canonical form is a string array, by first canonicalising toVarBinViewArray. - onpair_
compress_ array_ default - Convenience: build a default
ExecutionCtxfromLEGACY_SESSION. - onpair_
compress_ iter - Compress an iterable of optional byte strings via the OnPair encoder.
Type Aliases§
- OnPair
Array - An
OnPair-encoded Vortex array.