Design spaces

A DesignSpace describes the parameters an agent searches over. It is built from a list of spec dicts, one per parameter:

from banditry import DesignSpace

space = DesignSpace.parse([
    {"name": "learning_rate", "type": "num",  "lb": 1e-4, "ub": 1e-1},
    {"name": "num_layers",    "type": "int",  "lb": 1,    "ub": 8},
    {"name": "use_dropout",   "type": "bool"},
    {"name": "optimiser",     "type": "cat",  "categories": ["adam", "sgd", "rmsprop"]},
])

Parameter types

type	Spec keys	Domain
"num"	lb, ub	continuous float in [lb, ub]
"int"	lb, ub	integer in [lb, ub] (inclusive)
"bool"	—	True / False
"cat"	categories	one of the listed values

Suggestions come back as a pandas DataFrame with one column per parameter in the original (untransformed) domain — optimiser is "adam", not an index; use_dropout is a bool.

Column ordering

space.para_names lists numeric-ish parameters first (num, int, bool), then categorical ones. Suggestion DataFrames follow this order.

Transforms

Internally, models see a transformed representation: numeric-ish parameters as a float tensor, categorical parameters as integer category indices (fed to embeddings or one-hot encodings by the surrogates).

x_num, x_cat = space.transform(df)      # DataFrame -> (FloatTensor, LongTensor)
df_back = space.inverse_transform(x_num, x_cat)

Custom parameter types

Register a subclass of Parameter under a new type name to extend the spec format:

from banditry import DesignSpace
from banditry.variable_domains import NumericParameter

class LogUniformParameter(NumericParameter):
    ...  # override transform / inverse_transform / sample

DesignSpace.register_parameter_type("lognum", LogUniformParameter)
space = DesignSpace.parse([{"name": "lr", "type": "lognum", "lb": 1e-5, "ub": 1e-1}])

Full API: Design spaces reference.