Operator Protocol

Core operator protocol for data transformations.

See Also

• Core Overview - All core protocols
• Operators - Operator implementations
• Element Operator - Element transforms
• Operators Tutorial

---

datarax.core.operator

OperatorModule - base class for parametric transformation modules.

This module provides OperatorModule, the base class for all parametric, differentiable data transformations in Datarax.

Key Features:

• Config-based initialization with OperatorConfig
• Stochastic mode (with random parameter generation)
• Deterministic mode (no randomness)
• Batch processing with vmap
• JIT compatibility with static branching
• Statistics system (inherited from DataraxModule)

logger module-attribute

logger = getLogger(__name__)

OperatorModule

OperatorModule(config: OperatorConfig, *, rngs: Rngs | None = None, name: str | None = None)

Bases: DataraxModule

Base class for parametric, differentiable operators.

Operators work on Batch[Element] data and can have learnable parameters. They support both stochastic (random) and deterministic modes.

The operator pattern separates RNG generation from transformation: 1. generate_random_params() - Generates batch-level random parameters (impure) 2. apply() - Applies transformation to single element (pure function) 3. apply_batch() - Orchestrates batch processing with vmap (concrete implementation)

Parameters:

Name	Type	Description	Default
config	OperatorConfig	OperatorConfig (already validated via post_init)	required
rngs	Rngs | None	Random number generators (required if stochastic=True)	None
name	str | None	Optional name for the operator	None

Attributes:

Name	Type	Description
config	OperatorConfig	Operator configuration
stochastic		Whether this operator uses randomness (from config)
stream_name		RNG stream name (from config, required if stochastic=True)

Parameters:

Name	Type	Description	Default
config	OperatorConfig	Operator configuration (already validated)	required
rngs	Rngs | None	Random number generators (required if stochastic=True)	None
name	str | None	Optional operator name	None

Raises:

Type	Description
ValueError	If stochastic=True but rngs is None

config instance-attribute

config: OperatorConfig

stochastic instance-attribute

stochastic = static(stochastic)

stream_name instance-attribute

stream_name = static(stream_name)

rngs instance-attribute

rngs = rngs

name instance-attribute

name = static(name)

generate_random_params

generate_random_params(rng: Array, data_shapes: PyTree) -> PyTree

Generate random parameters for batch transformation.

This method generates batch-level random parameters (one per batch element). It is impure (uses RNG) and should be called once per batch.

Required for stochastic operators. Deterministic operators can leave default.

Parameters:

Name	Type	Description	Default
rng	Array	JAX random key	required
data_shapes	PyTree	PyTree with same structure as batch.data, containing shapes Examples: {"image": (batch_size, H, W, C)}	required

Returns:

Type	Description
PyTree	PyTree of random parameters matching batch structure.
PyTree	Can be any structure (scalars, arrays, dicts, etc.)

Raises:

Type	Description
NotImplementedError	If stochastic=True but not implemented

apply

apply(data: PyTree, state: PyTree, metadata: dict[str, Any] | None, random_params: Any = None, stats: dict[str, Any] | None = None) -> tuple[PyTree, PyTree, dict[str, Any] | None]

Apply operator to single element (no batch dimension).

This is a PURE FUNCTION that transforms a single data element. It should not access self.rngs or generate random numbers. All randomness comes through random_params argument.

Subclasses MUST implement this method.

Parameters:

Name	Type	Description	Default
data	PyTree	Element data PyTree (typically dict[str, Array], no batch dim)	required
state	PyTree	Element state PyTree (typically dict[str, Any])	required
metadata	dict[str, Any] | None	Element metadata as structured dict	required
random_params	Any	Random parameters for this element (from generate_random_params)	None
stats	dict[str, Any] | None	Optional statistics (from get_statistics() or passed explicitly)	None

Returns:

Type	Description
PyTree	Tuple of (transformed_data, new_state, new_metadata)
PyTree	All return values are PyTrees matching input structure

Examples:

Example implementation:

def apply(self, data, state, metadata, random_params=None, stats=None):
    # Apply random brightness
    factor = random_params if random_params is not None else 1.0
    transformed = {"image": data["image"] * factor}
    return transformed, state, metadata

get_output_structure

get_output_structure(sample_data: PyTree, sample_state: PyTree) -> tuple[PyTree, PyTree]

Declare output PyTree structure for vmap axis specification.

Default uses jax.eval_shape to discover structure automatically. Override for efficiency or when eval_shape doesn't work (e.g., data-dependent shapes).

Parameters:

Name	Type	Description	Default
sample_data	PyTree	Single element data (not batched)	required
sample_state	PyTree	Single element state (not batched)	required

Returns:

Type	Description
PyTree	Tuple of (output_data_structure, output_state_structure) with None leaves.
PyTree	The structure (keys/nesting) matters, leaf values are ignored.

Example override for operator that adds keys

def get_output_structure(self, sample_data, sample_state): out_data = { **jax.tree.map(lambda _: None, sample_data), "score": None, "alignment": None, } return out_data, sample_state

apply_batch

apply_batch(batch: Batch, stats: dict[str, Any] | None = None) -> Batch

Process entire batch with vmap and optional RNG generation.

This method implements the batch processing logic for both stochastic and deterministic modes. It uses static branching on self.stochastic for JIT compilation efficiency.

The implementation delegates to _vmap_apply() for the shared computational core, then wraps the result in a Batch object.

Parameters:

Name	Type	Description	Default
batch	Batch	Input batch (Batch[Element] structure)	required
stats	dict[str, Any] | None	Optional statistics (if None, uses get_statistics())	None

Returns:

Type	Description
Batch	Transformed batch with same structure

Note

This method is concrete (not abstract). Subclasses typically don't override it, but can if they need custom batch processing logic.

output_spec

output_spec(input_spec: PyTree) -> PyTree

Return the operator's output spec given an input spec.

Most operators (normalization, additive noise, simple element-wise transforms) do not change shape; the default returns input_spec unchanged. Shape-changing operators (Resize, Crop, Reshape) MUST override this method.

Parameters:

Name	Type	Description	Default
input_spec	PyTree	PyTree of jax.ShapeDtypeStruct describing the input element (matching the upstream DataSourceModule.element_spec() or another operator's output_spec).	required

Returns:

Type	Description
PyTree	PyTree of jax.ShapeDtypeStruct describing the operator's output.
PyTree	By default, equal to input_spec.

get_operation_stats

get_operation_stats() -> dict[str, int]

Get operation statistics.

Note: This method converts JAX arrays to Python ints for introspection. It is intended for use outside of JIT-compiled functions.

Returns:

Type	Description
dict[str, int]	Dictionary with 'applied_count' and 'skipped_count'

reset_operation_stats

reset_operation_stats() -> None

Reset operation statistics to zero.

Note: Creates new JAX arrays to reset the counters.

compute_statistics

compute_statistics(data: Any) -> dict[str, Any] | None

Compute statistics from data using batch_stats_fn.

If batch_stats_fn is not configured, returns None. Computed statistics are cached in _computed_stats.

Parameters:

Name	Type	Description	Default
data	Any	Input data to compute statistics from	required

Returns:

Type	Description
dict[str, Any] | None	Dictionary of statistics, or None if no batch_stats_fn configured

get_statistics

get_statistics() -> dict[str, Any] | None

Get current statistics.

Returns precomputed_stats if configured (unless reset was called), otherwise returns cached computed statistics, or None if no statistics available.

Returns:

Type	Description
dict[str, Any] | None	Dictionary of statistics, or None if no statistics available

set_statistics

set_statistics(stats: dict[str, Any]) -> None

Manually set statistics.

This overwrites any previously computed statistics and clears reset flag.

Parameters:

Name	Type	Description	Default
stats	dict[str, Any]	Dictionary of statistics to set	required

reset_statistics

reset_statistics() -> None

Reset all statistics to None.

This clears both computed statistics and marks that precomputed_stats should be ignored (via internal flag). After reset, get_statistics() will return None until new statistics are set or computed.

reset_cache

reset_cache() -> None

Clear the cache.

Only has effect if cacheable=True in config.

copy

copy(*, config: DataraxModuleConfig | None = None, rngs: Rngs | None = None, name: str | None = None) -> DataraxModule

Create a copy of this module with optional config/parameter changes.

This allows creating a new module instance with modified configuration while preserving other attributes. Useful for hyperparameter tuning.

Parameters:

Name	Type	Description	Default
config	DataraxModuleConfig | None	New config (if None, uses current config)	None
rngs	Rngs | None	New RNG state (if None, uses current rngs)	None
name	str | None	New name (if None, uses current name)	None

Returns:

Type	Description
DataraxModule	New module instance with updated parameters

Examples:

Change configuration

new_config = DataraxModuleConfig(cacheable=True) new_module = module.copy(config=new_config)

Change name only

renamed = module.copy(name="new_name")

Note

Subclasses can override this method to provide more fine-grained control over copying, such as allowing individual config field updates without requiring dataclass replace().

get_state

get_state() -> dict[str, Any]

Get module state for checkpointing.

This method implements the Checkpointable protocol using NNX state management. It extracts all state variables from the module and converts them to a serializable format.

Returns:

Type	Description
dict[str, Any]	A dictionary containing the internal state of the component.

set_state

set_state(state: dict[str, Any]) -> None

Restore module state from a checkpoint.

This method implements the Checkpointable protocol using NNX state management. It restores the module state from a serialized format. Restoration is strict: checkpoint structure must match module state.

Parameters:

Name	Type	Description	Default
state	dict[str, Any]	A dictionary containing the internal state to restore.	required

Raises:

Type	Description
TypeError	If state is not a dictionary.
ValueError	If checkpoint structure does not match module state.

clone

clone() -> DataraxModule

Create a new instance with the same state as this module.

Uses NNX's clone function for proper deep cloning of all state.

Returns:

Type	Description
DataraxModule	A new module instance with the same state.

requires_rng_streams

requires_rng_streams() -> list[str] | None

Get the list of RNG streams required by this module.

Returns:

Type	Description
list[str] | None	A list of required RNG stream names, or None if no RNG streams
list[str] | None	are required.

ensure_rng_streams

ensure_rng_streams(stream_names: list[str]) -> None

Ensure that the required RNG streams are available.

Parameters:

Name	Type	Description	Default
stream_names	list[str]	A list of available RNG stream names.	required

Raises:

Type	Description
ValueError	If a required RNG stream is not available.

extract_batch_size

extract_batch_size(data_shapes: PyTree) -> int

Extract batch size from a PyTree of shape tuples.

Traverses the PyTree treating tuples as atomic leaves (since JAX normally unfolds tuples as nodes) and returns the first axis of the first leaf shape.

Parameters:

Name	Type	Description	Default
data_shapes	PyTree	PyTree with same structure as batch data, where each leaf is a shape tuple (e.g. (batch_size, H, W, C)).	required

Returns:

Type	Description
int	The batch size (first element of the first shape found).

Raises:

Type	Description
ValueError	If the shape tree has no leaves.