spines¶

Spines

Skeletons for parameterized models.

class spines.Model(*args, **kwargs)[source]¶

Bases: spines.transforms.base.Transform

Spines primary Model class

fit(*args, **kwargs) → None[source]¶

Fits the model

Generally this method is used for a single iteration of model fitting (and for simple models this may be the only call which is required). The train method can call this function multiple times and update the model iteratively (where that approach is appropriate).

Parameters

args (optional) – Arguments to use in fit call.
kwargs (optional) – Any additional keyword arguments to use in fit call.

Returns

Either returns None if adjustments to the model’s parameters happen internally, otherwise returns the dictionary of updated parameters to apply.

Return type

None or dict

See also

train(), transform()

get_hyper_params() → Dict[str, Any][source]¶

Gets the current hyper-parameter values

Returns: Copy of the currently set hyper-parameter values.
Return type: dict

hyper_parameters¶

Hyper-parameters which are currently set.

Type: ParameterStore

predict(*args, **kwargs) → Any[source]¶

Predict outputs for the given inputs

Parameters

args (optional) – Additional arguments to pass to predict call.
kwargs (optional) – Additional keyword arguments to pass to predict call.

Returns

Predictions from the given data.

Return type

object

set_hyper_parameter(name: str, value: Any) → None[source]¶

Sets a hyper-parameter value

Sets a hyper-parameter’s value if the given hyper_param and value are valid.

Parameters

name (str) – Hyper-parameter to set value for.
value – Value to set.

Raises

MissingParameterException – If the given name hyper-parameter does not exist.
InvalidParameterException – If the given value is not valid for the specified hyper-parameter.

set_hyper_params(**hyper_params) → None[source]¶

Sets the values of this model’s hyper-parameters

Parameters: hyper_params – Hyper-parameter values to set.
Raises: InvalidParameterException – If one of the given hyper-parameter values is not valid.

transform(*args, **kwargs) → Any[source]¶

Transforms the given input data

Parameters

args (optional) – Additional arguments to pass to predict call.
kwargs (optional) – Additional keyword arguments to pass to predict call.

Returns

Transformed inputs.

Return type

object

unset_hyper_parameter(name: str) → Any[source]¶

Un-sets a hyper-parameter

Un-sets the specified hyper-parameter’s value from the set of hyper-parameters and returns the previously set value.

Parameters: name (str) – Name of the hyper-parameter to clear the value for.
Returns: Previously set value of the hyper-parameter.
Return type: object
Raises: MissingParameterException – If the given name hyper-parameter does not exist.

class spines.Parameter(*value_type, default: Optional[Any] = None, desc: Optional[str] = None)[source]¶

Bases: object

Parameter class

Parameters

value_type (type or Iterable of type) – The type(s) of values allowed for this parameter.
default (object, optional) – Default value for this parameter, if any.
desc (str, optional) – Description for this parameter, if any.

check(value: Any) → bool[source]¶

Checks the given value for validity

Based on this particular Parameter’s settings this method checks the given value to see if it’s in-line with the specifications.

Parameters: value – Parameter value to check validity of.
Returns: Whether or not the value is valid for the parameter.
Return type: bool

See also

preprocess()

default¶

Default value to use for this parameter.

Type: object

desc¶

Description of this parameter.

Type: str

name¶

Name of this parameter.

Type: str

preprocess(value: Any) → Any[source]¶

Pre-process/massage parameter values into correct type

This is used to try and convert parameter values to the allowed types (if possible). For instance the allowed type may be a float but the user provides 1, we would likely not want to break in that situation, but simply convert the int given into a float. However, this can also be used to perform more complex initialization or customization on given parameter values, if needed.

The default method is to try and cast the given value as the types given in this Parameter’s _value_types attribute and returning the first successful cast, otherwise returns the given input value.

This method can be easily extended or overridden for custom Parameter classes to handle more complex value types.

Parameters: value (object) – The value to pre-process for this parameter.
Returns: The pre-processed value.
Return type: object

Note

This method is, by default, only called if the given value fails this Parameter’s check call.

See also

check()

required¶

Whether or not this parameter is required to be set.

Type: bool

value_type¶

The types of values allowed for this option.

Type: tuple

class spines.HyperParameter(*value_type, default: Optional[Any] = None, desc: Optional[str] = None)[source]¶

Bases: spines.parameters.base.Parameter

Hyper-parameter

spines.get_config(setting: str = _MISSING, default: Any = _MISSING) → spines.config.core.Config[source]¶

Get the global spines configuration

Parameters

setting (str, optional) – Setting(s) to get from the global configuration.
default (optional) – Value to return if setting is not set (if not given an error will occur if accessing a non-existant setting).

Returns

The global configuration settings for the current spines project or the values for the settings given.

Return type

Config or dict

Raises

ValueError – If the given setting(s) doesn’t exist and no default value is provided.

spines.load_config(*path, update: bool = False) → None[source]¶

Loads the global spines configuration

Parameters

path (str (or multiple), optional) – File(s) to load the configuration from (defaults to the correct spines hierarchy for configuration files).
update (bool, optional) – Update the current configuration as opposed to replacing it with the newly loaded on (default is False, replace it).

spines.set_config(section: Optional[str] = None, **settings) → None[source]¶

Sets global configuration setting(s)

Parameters

section (str, optional) – Configuration section to set settings in.
settings – The setting(s) in the global configuration to update.