xtructure.hashtable package

Submodules

xtructure.hashtable.table module

HashTable data container and public API.

class xtructure.hashtable.table.HashTable(seed: int, capacity: int, _capacity: int, bucket_size: int, size: int, table: Xtructurable, bucket_fill_levels: Array | ndarray | bool | number, fingerprints: Array | ndarray | bool | number, max_probes: int)

Bases: object

Bucketed Double Hash Table Implementation

Uses double hashing with buckets to resolve collisions.

bucket_fill_levels: Array | ndarray | bool | number

bucket_size: int

static build(dataclass: Xtructurable, seed: int, capacity: int, bucket_size: int = 8, hash_size_multiplier: int = 2, max_probes: int | None = None) → HashTable

Initialize a new hash table backed by JAX-friendly storage.

capacity: int

fingerprints: Array | ndarray | bool | number

from_tuple()

insert(input: Xtructurable) → tuple[HashTable, bool, HashIdx]

lookup(input: Xtructurable) → tuple[HashIdx, bool]

lookup_bucket(input: Xtructurable) → tuple[BucketIdx, bool, Array | ndarray | bool | number]

lookup_parallel(inputs: Xtructurable, filled: Array | ndarray | bool | number | bool = True) → tuple[HashIdx, Array | ndarray | bool | number]

max_probes: int

parallel_insert(inputs: Xtructurable, filled: Array | ndarray | bool | number | bool = None, unique_key: Array | ndarray | bool | number = None)

replace(**kwargs)

seed: int

size: int

table: Xtructurable

to_tuple()

Module contents

class xtructure.hashtable.HashIdx(index: FieldDescriptor(dtype = <class 'jax.numpy.uint32'>, fill_value=4294967295, intrinsic_shape=(), bits=None, packed_bits=None, unpacked_dtype=None, unpacked_intrinsic_shape=None, fill_value_factory=None, validator=None))

Bases: object

property at

property batch_shape

property bytes

Convert entire state tree to flattened byte array.

check_invariants()

classmethod default(shape: Tuple[int, ...] = ()) → T

default_dtype = (<class 'jax.numpy.uint32'>,)

default_shape = ((),)

property dtype

flatten()

from_tuple()

hash(seed=0)

Main hash function that converts state to uint32 lanes and hashes them.

hash_with_uint32ed(seed=0)

Main hash function that converts state to uint32 lanes and hashes them. Returns both hash value and its uint32 representation.

index: FieldDescriptor(dtype=<class 'jax.numpy.uint32'>, fill_value=4294967295, intrinsic_shape=(), bits=None, packed_bits=None, unpacked_dtype=None, unpacked_intrinsic_shape=None, fill_value_factory=None, validator=None)

is_xtructed = True

classmethod load(path: str) → T

Loads an instance from a .npz file.

property ndim: int

Return number of batch dimensions for structured instances.

padding_as_batch(batch_shape: tuple[int, ...])

classmethod random(shape=(), key=None)

replace(**kwargs)

reshape(*new_shape: int | tuple[int, ...]) → T

save(path: str, *, packed: bool = True)

Saves the instance to a .npz file.

property shape

str(**kwargs)

property structured_type: StructuredType

to_tuple()

transpose(axes: tuple[int, ...] | None = None) → T

property uint32ed

Convert pytree to uint32 array.

class xtructure.hashtable.HashTable(seed: int, capacity: int, _capacity: int, bucket_size: int, size: int, table: Xtructurable, bucket_fill_levels: Array | ndarray | bool | number, fingerprints: Array | ndarray | bool | number, max_probes: int)

Bases: object

Bucketed Double Hash Table Implementation

Uses double hashing with buckets to resolve collisions.

bucket_fill_levels: Array | ndarray | bool | number

bucket_size: int

static build(dataclass: Xtructurable, seed: int, capacity: int, bucket_size: int = 8, hash_size_multiplier: int = 2, max_probes: int | None = None) → HashTable

Initialize a new hash table backed by JAX-friendly storage.

capacity: int

fingerprints: Array | ndarray | bool | number

from_tuple()

insert(input: Xtructurable) → tuple[HashTable, bool, HashIdx]

lookup(input: Xtructurable) → tuple[HashIdx, bool]

lookup_bucket(input: Xtructurable) → tuple[BucketIdx, bool, Array | ndarray | bool | number]

lookup_parallel(inputs: Xtructurable, filled: Array | ndarray | bool | number | bool = True) → tuple[HashIdx, Array | ndarray | bool | number]

max_probes: int

parallel_insert(inputs: Xtructurable, filled: Array | ndarray | bool | number | bool = None, unique_key: Array | ndarray | bool | number = None)

replace(**kwargs)

seed: int

size: int

table: Xtructurable

to_tuple()