These methods would make it easier to add a subgraph to an existing hyperparameter graph and remove it later if necessary.

I was trying to setup a simple all-continuous PCS (code below).

import ConfigSpace as CS

cs = CS.ConfigurationSpace()

for i in range(5):
    cs.add_hyperparameter(CS.UniformFloatHyperparameter('x%i'%i, i-1, i+1, i ))

default = cs.get_default_configuration()
print(cs.get_hyperparameters())
print(default.get_dictionary())
print(default.get_array())

Running above code, I get the following output (slightly reformatted to improve readability):

[   x0, Type: UniformFloat, Range: [-1.0, 1.0], Default: 0.0,
    x1, Type: UniformFloat, Range: [0.0, 2.0], Default: 1.0,
    x2, Type: UniformFloat, Range: [1.0, 3.0], Default: 2.0,
    x3, Type: UniformFloat, Range: [2.0, 4.0], Default: 3.0,
    x4, Type: UniformFloat, Range: [3.0, 5.0], Default: 4.0]
{'x2': 2.0, 'x3': 3.0, 'x4': 4.0, 'x0': 0.0, 'x1': 1.0}
[ 0.5  0.5  0.5  0.5  0.5]

The last line should be [0,1,2,3,4], but instead consists of 0.5 everywhere.

Hi,

after waiting for more than 30 seconds, I don't get a random_neighbor where the pcs has conditionals, e.g., /data/aad/aclib/target_algorithms/sat/spear-python/spear-params-mixed.pcs

(refactor branch)

Best,
Marius

when there are two conditions which affect the same child:

@1:0:restarts {F,L,D,x,+,no}[x]
@1:S:Luby:aryrestarts {1,2}[1]
@1:2:Luby:restarts [1,65535][1000]il
@1:2:Luby:restarts | @1:0:restarts in {L}
@1:2:Luby:restarts | @1:S:Luby:aryrestarts in {2}

via e-mail from @mlindauer

I want to create configuration from a HPOlib-pkl-file :

hpolib_pkl = cPickle.load(open(path_to_pkl,"r"))
config = configuration_space.Configuration(config_space, hpolib_pkl["trials"][0]["params"])

But it crashes with the error message:

Value 1.0, <type 'str'> for instantiation of hyperparameter 'random_forest:max_features, Type: UniformFloat, Range: [0.5, 5.0], Default: 1' is not a legal value

Did I miss something? Are the dictionaries for a parameter instantiation created by HPOlib in a different format than HPOlibConfigSpace expects?

The HPOlib dictionary looks as follows:

In [350]: hpolib_pkl["trials"][0]["params"]
Out[350]: OrderedDict([('classifier', 'random_forest'), ('imputation:strategy', 'mean'), ('preprocessor', 'None'), ('random_forest:bootstrap', 'True'), ('random_forest:criterion', 'gini'), ('random_forest:max_depth', 'None'), ('random_forest:max_features', '1.0'), ('random_forest:max_leaf_nodes', 'None'), ('random_forest:min_samples_leaf', '1'), ('random_forest:min_samples_split', '2'), ('random_forest:n_estimators', '100'), ('rescaling:strategy', 'min/max')])