gym_gz.randomizers.model

gym_gz.randomizers.model#

gym_gz.randomizers.model.sdf#

class gym_gz.randomizers.model.sdf.Distribution(value)#

Bases: Enum

An enumeration.

Gaussian = 2#
Uniform = 1#
class gym_gz.randomizers.model.sdf.GaussianParams(variance, mean)#

Bases: NamedTuple

mean: float#

Alias for field number 1

variance: float#

Alias for field number 0

class gym_gz.randomizers.model.sdf.Method(value)#

Bases: Enum

An enumeration.

Absolute = 1#
Additive = 2#
Coefficient = 3#
class gym_gz.randomizers.model.sdf.RandomizationData(xpath, distribution, parameters, method, ignore_zeros, force_positive, element)#

Bases: NamedTuple

distribution: str#

Alias for field number 1

element: Element#

Alias for field number 6

force_positive: bool#

Alias for field number 5

ignore_zeros: bool#

Alias for field number 4

method: Method#

Alias for field number 3

parameters: Union[UniformParams, GaussianParams]#

Alias for field number 2

xpath: str#

Alias for field number 0

class gym_gz.randomizers.model.sdf.RandomizationDataBuilder(randomizer)#

Bases: object

Builder class of a RandomizationData object.

Parameters:

randomizer (SDFRandomizer) – The SDFRandomizer object to which the created randomization will be inserted.

add()#

Close the chaining of methods and return to the SDF randomizer the configuration.

Raises:

RuntimeError – If the XPath pattern does not find any match in the SDF.

Return type:

None

at_xpath(xpath)#

Set the XPath pattern associated to the randomization.

Parameters:

xpath (str) – The XPath pattern.

Return type:

RandomizationDataBuilder

Returns:

The randomization builder to allow chaining methods.

force_positive(force_positive=True)#

Force the randomized value to be greater than zero.

This option is helpful to enforce that values e.g. the mass will stay positive regardless of the applied distribution parameters.

Parameters:

force_positive (bool) – True to force positive parameters, false otherwise.

Return type:

RandomizationDataBuilder

Returns:

The randomization builder to allow chaining methods.

ignore_zeros(ignore_zeros)#

Ignore the randomization of values that are zero.

If the value to randomize has a default value of 0 in the SDF, when this method is chained the randomization is skipped. In the case of a multi-match XPath pattern, the values that are not zero are not skipped.

Parameters:

ignore_zeros (bool) – True if zeros should be ignored, false otherwise.

Return type:

RandomizationDataBuilder

Returns:

The randomization builder to allow chaining methods.

method(method)#

Set the randomization method.

Parameters:

method (Method) – The desired randomization method.

Return type:

RandomizationDataBuilder

Returns:

The randomization builder to allow chaining methods.

sampled_from(distribution, parameters)#

Set the distribution associated to the randomization.

Parameters:
Return type:

RandomizationDataBuilder

Returns:

The randomization builder to allow chaining methods.

class gym_gz.randomizers.model.sdf.SDFRandomizer(sdf_model)#

Bases: object

Randomized SDF files generator.

Parameters:

sdf_model (str) – The absolute path to the SDF file.

Raises:

ValueError – If the SDF file does not exist.

clean()#

Clean the SDF randomizer.

Return type:

None

find_xpath(xpath)#

Find the elements that match an XPath pattern.

This method could be helpful to test the matches of a XPath pattern before using it in at_xpath().

Parameters:

xpath (str) – The XPath pattern.

Return type:

List[Element]

Returns:

A list of elements matching the XPath pattern.

get_active_randomizations()#

Return the active randomizations.

This method could be helpful also in the case of multi-match XPath patterns to validate that the inserted randomizations have been processed correctly.

Return type:

List[RandomizationData]

Returns:

The list of the active randomizations.

insert(randomization_data)#

Insert a randomization.

Parameters:

randomization_data – A new randomization.

Return type:

None

new_randomization()#

Start the chaining to build a new randomization.

Return type:

RandomizationDataBuilder

Returns:

A randomization builder.

process_data()#

Process all the inserted randomizations.

Raises:

RuntimeError – If the XPath of a randomization has no matches.

Return type:

None

sample(pretty_print=False)#

Sample a randomized SDF string.

Parameters:

pretty_print – True to pretty print the output.

Raises:

  • ValueError – If the distribution of a randomization is not recognized.

  • ValueError – If the method of a randomization is not recognized.

Return type:

str

Returns:

The randomized model as SDF string.

seed(seed)#

Seed the SDF randomizer.

Parameters:

seed (int) – The seed number.

Return type:

None

class gym_gz.randomizers.model.sdf.UniformParams(low, high)#

Bases: NamedTuple

high: float#

Alias for field number 1

low: float#

Alias for field number 0