Ctrl+K

rics.collections.dicts#

Dict utility functions.

Module Attributes

KT

Key type.

VT

Value type.

HVT

Hashable value type.

OKT

Outer key type.

MakeType

Valid input types for making the InheritedKeysDict.make() function.

Functions

compute_if_absent(d, key[, func])

Compute and store key using func if key is not in the dict.

flatten_dict(d[, join_string, ...])

Flatten a nested dictionary.

reverse_dict(d[, duplicate_key_action])

Swap keys and values.

unflatten_dict(d[, join_string])

Unflatten a flat dictionary.

Classes

InheritedKeysDict([specific, default])

A nested dictionary that returns default-backed child dictionaries.
class KT#

Key type.

alias of TypeVar(‘KT’, bound=Hashable)

class VT#

Value type.

alias of TypeVar(‘VT’)

class HVT#

Hashable value type.

alias of TypeVar(‘HVT’, bound=Hashable)

class OKT#

Outer key type.

alias of TypeVar(‘OKT’, bound=Hashable)

compute_if_absent(d: dict[KT, VT], key: KT, func: Callable[[KT], VT] | None = None) → VT[source]#

Compute and store key using func if key is not in the dict.

Parameters:

  • d – A dict.

  • key – The key to get.

  • func – A function to call for missing keys. Perform regular __getitem__ call if None.

Returns:

The value of k in d.

reverse_dict(d: Mapping[KT, HVT], duplicate_key_action: Literal['ignore', 'warn', 'raise', 'IGNORE', 'WARN', 'RAISE'] | ActionLevel = 'raise') → dict[HVT, KT][source]#

Swap keys and values.

Parameters:

  • d – A dict to reverse.

  • duplicate_key_action – Action to take if the return dict has key collisions in the reversed dict, i.e. there are duplicate values in d. Set to ignore to allow.

Returns:

A reversed copy of d.

Examples

Reversing a dict with two elements.

>>> reverse_dict({"A": 0, "B": 1})
{0: 'A', 1: 'B'}
Raises:

ValueError – If there are duplicate values in d and duplicate_key_action='raise'.

flatten_dict(d: dict[~rics.collections.dicts.KT, ~typing.Any], join_string: str = '.', filter_predicate: ~typing.Callable[[~rics.collections.dicts.KT, ~typing.Any], bool] = <function <lambda>>, string_fn: ~typing.Callable[[~rics.collections.dicts.KT], str] | str | None = <class 'str'>) → dict[str, Any][source]#

Flatten a nested dictionary.

This process is partially (or fully; depends on d and arguments) reversible, see unflatten_dict().

Parameters:

  • d – A dict to flatten. Keys must be strings.

  • join_string – Joiner for nested keys.

  • filter_predicate – A callable (key, value) -> should_keep. Default always returns True.

  • string_fn – A callable which takes a non-string key value and converts into to a string. If None, a type error will be raised for non-string keys. Default is str(key). Pass a string to resolve using get_by_full_name().

Returns:

A flattened version of d.

Examples

Flattening a shallow nested dict.

>>> flatten_dict({"foo": 0, "bar": {"foo": 1, "bar": 2}})
{'foo': 0, 'bar.foo': 1, 'bar.bar': 2}
unflatten_dict(d: dict[str | tuple[str, ...], Any], join_string: str = '.') → dict[str, dict[str, Any] | Any][source]#

Unflatten a flat dictionary.

This process is reversible, see flatten_dict().

Parameters:

  • d – A flat dict to unflatten. Keys must be str or tuple.

  • join_string – Joiner for flattened keys. Ignored if d has tuple-keys.

Returns:

A nested version of d.

Examples

Unflatten a flat dict.

>>> unflatten_dict({"foo": 0, "bar.foo": 1, "bar.bar": 2})
{'foo': 0, 'bar': {'foo': 1, 'bar': 2}}

Tuple keys are also supported, including mixed key types.

>>> unflatten_dict({"foo": 0, ("bar", "foo"): 1, "bar.bar": 2})
{'foo': 0, 'bar': {'foo': 1, 'bar': 2}}
class InheritedKeysDict(specific: dict[OKT, dict[KT, VT]] | None = None, default: dict[KT, VT] | None = None)[source]#

Bases: Mapping[OKT, dict[KT, VT]]

A nested dictionary that returns default-backed child dictionaries.

The length of an InheritedKeysDict is equal to the number of specific outer keys, and is considered True when cast to bool if there are shared and/or specific keys present.

Parameters:

  • default – Shared (fallback) mappings for all contexts.

  • specific – Context-specific mappings, backed by the default fallback mappings.

Examples

A short demonstration.

>>> shared = {0: "fallback-for-0", 1: "fallback-for-1"}
>>> specific = {
...     "ctx0": {0: "c0-v0"},
...     "ctx1": {0: "c1-v0", 1: "c1-v1", 2: "c1-v2"},
... }
>>> ikd = InheritedKeysDict(default=shared, specific=specific)
>>> ikd
InheritedKeysDict(default={0: 'fallback-for-0', 1: 'fallback-for-1'}, specific={'ctx0': {0: 'c0-v0'},
'ctx1': {0: 'c1-v0', 1: 'c1-v1', 2: 'c1-v2'}})

The value of key 0 is inherited for ‘ctx0’. The ‘ctx1’-context defines all shared keys, as well as a unique key.

>>> ikd["ctx0"]
{0: 'c0-v0', 1: 'fallback-for-1'}
>>> ikd["ctx1"]
{0: 'c1-v0', 1: 'c1-v1', 2: 'c1-v2'}

The InheritedKeysDict.__contains__-method is True for all keys. Unknown keys simply return the default values. This will be an empty if no specific keys are specified.

>>> "unseen-key" in ikd
True
>>> ikd["unseen-key"]
{0: 'fallback-for-0', 1: 'fallback-for-1'}

The length of ikd is equal to the number of specific contexts (two in this case).

copy() → InheritedKeysDict[OKT, KT, VT][source]#

Make a copy of this InheritedKeysDict.

classmethod make(arg: InheritedKeysDict[OKT, KT, VT] | _MakeDict[OKT, KT, VT]) → InheritedKeysDict[OKT, KT, VT][source]#

Create instance from a mapping.

The given argument must be on the format:

{
    "default": {key: value},
    "specific": {
        ctx0: {key: value},
        ctx1: {key: value},
        ...
        ctxN: {key: value},
    }
}

No other top-level keys are accepted, but neither default nor context-specific are required.

Parameters:

arg – Input to make an instance from.

Returns:

A new instance.

Raises:

ValueError – If there are any keys other than ‘default’ and ‘context-specific’ present in mapping.

MakeType#

Valid input types for making the InheritedKeysDict.make() function.

alias of Union[InheritedKeysDict[OKT, KT, VT], _MakeDict[OKT, KT, VT]]

On this page