dictionary#

Transformations and utilities for manipulating dictionaries.

Note that while most transformations in this module (CreateDictFromValue, Subset, and GetValue) support the metric AddRemoveKeys, AugmentDictTransformation does not. Because of this, none of the included derived transformations (such as create_copy_and_transform_value()) support AddRemoveKeys. Instead, use transformations in add_remove_keys.

Functions#

`create_copy_and_transform_value()`	Returns a transformation that transforms and re-adds a value in the input dict.
`create_rename()`	Returns a transformation that renames a single key.
`create_apply_dict_of_transformations()`	Returns a transformation that applies all given transformations to the input.
`create_transform_value()`	Returns a transformation that transforms a single value in the input dict.
`create_transform_all_values()`	Returns a transformation that transforms every value in the input dict.

create_copy_and_transform_value(input_domain, input_metric, key, new_key, transformation, hint)#

Returns a transformation that transforms and re-adds a value in the input dict.

The returned transformation has roughly the same behavior as

def copy_and_transform_value(data):
    data[new_key] = transformation(data[key])
    return data

The input is a dictionary, a single value is transformed and added to the dictionary at a new key. Note that the original value is left unchanged in the dictionary.

Parameters

input_domain (tmlt.core.domains.collections.DictDomain) – The domain for the input data.
input_metric (tmlt.core.metrics.DictMetric) – The metric for the input data.
key (Any) – The key containing the data to transform.
new_key (Any) – The key to store the transformed data.
transformation (Any) – The transformation to apply.
hint (Callable[[Any, Any], Any]) – A hint for the transformation.

Return type

tmlt.core.transformations.base.Transformation

create_rename(input_domain, input_metric, key, new_key)#

Returns a transformation that renames a single key.

The returned transformation has roughly the same behavior as

def rename(data):
    data[new_key] = data.pop(key)
    return data

Parameters

input_domain (tmlt.core.domains.collections.DictDomain) – The domain for the input data.
input_metric (tmlt.core.metrics.DictMetric) – The metric for the input data.
key (Any) – The original key.
new_key (Any) – The new key.

Return type

tmlt.core.transformations.base.Transformation

create_apply_dict_of_transformations(transformation_dict, hint_dict)#

Returns a transformation that applies all given transformations to the input.

The returned transformation has roughly the same behavior as

def apply_dict_of_transformations(data):
    return {
        key: transformation(data)
        for key, transformation in transformation_dict.items()
    }

The input is a single element, and the output is a dictionary where each value is the input transformed by the corresponding transformation.

Parameters

transformation_dict (Mapping[Any, tmlt.core.transformations.base.Transformation]) – A dictionary of transformations with matching input domains and input metrics.
hint_dict (Mapping[Any, Callable[[Any, Any], Any]]) – A dictionary of hints for the corresponding transformations in transformation_dict.

Return type

tmlt.core.transformations.base.Transformation

create_transform_value(input_domain, input_metric, key, transformation, hint)#

Returns a transformation that transforms a single value in the input dict.

The returned transformation has roughly the same behavior as

def transform_value(data):
    data[key] = transformation(data[key])
    return data

Notice that the input is a dictionary, a single value is transformed, while the other values are left unchanged.

Parameters

input_domain (tmlt.core.domains.collections.DictDomain) – The domain for the input data.
input_metric (tmlt.core.metrics.DictMetric) – The metric for the input data.
key (Any) – The key to transform.
transformation (Any) – The transformation to apply.
hint (Callable[[Any, Any], Any]) – A hint for the transformation.

Return type

tmlt.core.transformations.base.Transformation

create_transform_all_values(transformation_dict, hint_dict)#

Returns a transformation that transforms every value in the input dict.

The returned transformation has roughly the same behavior as

def transform_all_values(data):
    return {
        key: transformation(data[key])
        for key, transformation in transformation_dict.items()
    }

Notice that the input is a dictionary, and each value in the dictionary is transformed by the corresponding transformation.

Parameters

transformation_dict (Mapping[Any, tmlt.core.transformations.base.Transformation]) – A dictionary of transformations with matching input domains and input metrics.
hint_dict (Mapping[Any, Callable[[Any, Any], Any]]) – A dictionary of hints for the corresponding transformations in transformation_dict.

Return type

tmlt.core.transformations.base.Transformation

Classes#

`CreateDictFromValue`	Create a dictionary from an object.
`AugmentDictTransformation`	Applies transformation to a dictionary and appends the output to the input.
`Subset`	Retrieve a subset of a dictionary by keys.
`GetValue`	Retrieve an object from a dictionary.

class CreateDictFromValue(input_domain, input_metric, key, use_add_remove_keys=False)#

Bases: tmlt.core.transformations.base.Transformation

Create a dictionary from an object.

Parameters

input_domain (tmlt.core.domains.base.Domain) –
input_metric (tmlt.core.metrics.Metric) –
key (Any) –
use_add_remove_keys (bool) –

__init__(input_domain, input_metric, key, use_add_remove_keys=False)#

Constructor.

Parameters

input_domain (DomainDomain) – Domain of input objects.
input_metric (MetricMetric) – Distance metric on input objects.
key (AnyAny) – Key for constructing dictionary with given object.
use_add_remove_keys (boolbool (default: False)) – Whether to use AddRemoveKeys as the output metric instead of DictMetric.

property key#

Returns the key for the created dictionary.

Return type: Any

stability_function(d_in)#

Returns the smallest d_out satisfied by the transformation.

The returned d_out is {self.key: d_in}.

Parameters: d_in (Any) – Distance between inputs under input_metric.
Return type: Any

__call__(val)#

Returns dictionary with value associated with specified key.

Parameters: val (Any) –
Return type: Dict[Any, Any]

property input_domain#

Return input domain for the measurement.

Return type: tmlt.core.domains.base.Domain

property input_metric#

Distance metric on input domain.

Return type: tmlt.core.metrics.Metric

property output_domain#

Return input domain for the measurement.

Return type: tmlt.core.domains.base.Domain

property output_metric#

Distance metric on input domain.

Return type: tmlt.core.metrics.Metric

stability_relation(d_in, d_out)#

Returns True only if close inputs produce close outputs.

See the privacy and stability tutorial (add link?) for more information.

Parameters

d_in (Any) – Distance between inputs under input_metric.
d_out (Any) – Distance between outputs under output_metric.

Return type

bool

__or__(other: Transformation) → Transformation#
__or__(other: tmlt.core.measurements.base.Measurement) → tmlt.core.measurements.base.Measurement: Return this transformation chained with another component.

class AugmentDictTransformation(transformation)#

Bases: tmlt.core.transformations.base.Transformation

Applies transformation to a dictionary and appends the output to the input.

Parameters: transformation (tmlt.core.transformations.base.Transformation) –

__init__(transformation)#

Constructor.

Parameters: transformation (TransformationTransformation) – Transformation to be applied to input dictionary.

property inner_transformation#

Returns the inner transformation.

Return type: tmlt.core.transformations.base.Transformation

stability_function(d_in)#

Returns the smallest d_out satisfied.

Returns {**d_in, **self.transformation.stability_function(d_in)}.

Parameters: d_in (Dict[Any, Any]) – Distance between inputs under input_metric.
Raises: NotImplementedError – If self.inner_transformation.stability_function(d_in) raises NotImplementedError.
Return type: Dict[Any, Any]

stability_relation(d_in, d_out)#

Returns True if close inputs produce close outputs.

Returns True if both of the following are true:

d_in[key] <= d_out[key] for all augmented keys.
self.inner_transformation.stability_relation(d_in, original_d_out)

where original_d_out is the subset of d_out excluding the augmented keys.

Parameters

d_in (Any) – Distance between inputs under input_metric.
d_out (Any) – Distance between outputs under output_metric.

Return type

bool

__call__(input_dict)#

Applies transformation on given key to produce augmented dictionary.

Parameters: input_dict (Dict[Any, Any]) –
Return type: Dict[Any, Any]

property input_domain#

Return input domain for the measurement.

Return type: tmlt.core.domains.base.Domain

property input_metric#

Distance metric on input domain.

Return type: tmlt.core.metrics.Metric

property output_domain#

Return input domain for the measurement.

Return type: tmlt.core.domains.base.Domain

property output_metric#

Distance metric on input domain.

Return type: tmlt.core.metrics.Metric

__or__(other: Transformation) → Transformation#
__or__(other: tmlt.core.measurements.base.Measurement) → tmlt.core.measurements.base.Measurement: Return this transformation chained with another component.

class Subset(input_domain, input_metric, keys)#

Bases: tmlt.core.transformations.base.Transformation

Retrieve a subset of a dictionary by keys.

Parameters

input_domain (tmlt.core.domains.collections.DictDomain) –
input_metric (Union[tmlt.core.metrics.DictMetric, tmlt.core.metrics.AddRemoveKeys]) –
keys (List[Any]) –

__init__(input_domain, input_metric, keys)#

Constructor.

Parameters

input_domain (DictDomainDictDomain) – Domain of input dictionaries.
input_metric (DictMetric | AddRemoveKeysUnion[DictMetric, AddRemoveKeys]) – Distance metric over input dictionaries.
keys (List[Any]List[Any]) – Keys to be used for extracting subset.

property keys#

Returns the keys to keep.

Return type: List[Any]

stability_function(d_in)#

Returns the smallest d_out satisfied by the transformation.

The returned d_out is {key: d_in[key] for key in self.keys}.

Parameters: d_in (Any) – Distance between inputs under input_metric.
Return type: Any

__call__(input_dict)#

Returns subset of dictionary specified by keys.

Parameters: input_dict (Any) –
Return type: Any

property input_domain#

Return input domain for the measurement.

Return type: tmlt.core.domains.base.Domain

property input_metric#

Distance metric on input domain.

Return type: tmlt.core.metrics.Metric

property output_domain#

Return input domain for the measurement.

Return type: tmlt.core.domains.base.Domain

property output_metric#

Distance metric on input domain.

Return type: tmlt.core.metrics.Metric

stability_relation(d_in, d_out)#

Returns True only if close inputs produce close outputs.

See the privacy and stability tutorial (add link?) for more information.

Parameters

d_in (Any) – Distance between inputs under input_metric.
d_out (Any) – Distance between outputs under output_metric.

Return type

bool

__or__(other: Transformation) → Transformation#
__or__(other: tmlt.core.measurements.base.Measurement) → tmlt.core.measurements.base.Measurement: Return this transformation chained with another component.

class GetValue(input_domain, input_metric, key)#

Bases: tmlt.core.transformations.base.Transformation

Retrieve an object from a dictionary.

Parameters

input_domain (tmlt.core.domains.collections.DictDomain) –
input_metric (Union[tmlt.core.metrics.DictMetric, tmlt.core.metrics.AddRemoveKeys]) –
key (Any) –

__init__(input_domain, input_metric, key)#

Constructor.

Parameters

input_domain (DictDomainDictDomain) – Domain of input dictionaries.
input_metric (DictMetric | AddRemoveKeysUnion[DictMetric, AddRemoveKeys]) – Distance metric for input dictionaries.
key (AnyAny) – Key for retrieval.