Sampler#

class Sampler(sampling_plan, **kwargs)[source]#

Bases: object

Generate samples based on a sampling plan. Initiate the class by passing a do_mpc.sampling.SamplingPlanner (sampling_plan) object. The class can be configured to create samples based on the defined cases in the sampling_plan.

The class can be created with optional keyword arguments which are passed to set_param().

Configuration and sampling:

  1. (Optional) use set_param() to configure the class. Use data_dir to choose the save location for the samples.

  2. Set the sample generating function with set_sample_function(). This function is executed for each of the samples in the sampling_plan.

  3. Use sample_data() to generate all samples defined in the sampling_plan. A new file is written for each sample.

  4. Or: Create an individual sample result with sample_idx(), where an index (int) referring to the sampling_plan determines the sampled case.

Note

By default, the Sampler will only create samples that do not already exist in the chosen data_dir.

Example:

sp = do_mpc.sampling.SamplingPlanner()

# Plan with two variables alpha and beta:
sp.set_sampling_var('alpha', np.random.randn)
sp.set_sampling_var('beta', lambda: np.random.randint(0,5))

plan = sp.gen_sampling_plan(n_samples=10)

sampler = do_mpc.sampling.Sampler(plan)

# Sampler computes the product of two variables alpha and beta
# that were created in the SamplingPlanner:

def sample_function(alpha, beta):
    return alpha*beta

sampler.set_sample_function(sample_function)

sampler.sample_data()
Parameters

sampling_plan (list) –

Methods#

sample_data#

sample_data(self)#

Sample data after having configured the Sampler. No user input is required and the method will iterate through all the items defined in the sampling_plan (obtained with do_mpc.sampling.SamplingPlanner). :rtype: None

Note

Depending on your sample_function (set with set_sample_function()) and the total number of samples, executing this method may take some time.

Note

If sampler.set_param(overwrite = False) (default) data will only be sampled for instances that do not yet exist.

sample_idx#

sample_idx(self, idx)#

Sample case based on the index of the sample.

Parameters

idx (int) – Index of the sampling_plan for which the sample should be created.

Raises

  • assertion – Index must be between 0 and n_samples.

  • assertion – sample_function must be set prior to sampling data.

Return type

None

set_param#

set_param(self, **kwargs)#

Configure the do_mpc.sampling.Sampler class.

Parameters must be passed as pairs of valid keywords and respective argument. For example:

sampler.set_param(overwrite = True)
Parameters

  • overwrite (bool) – Should previously created results be overwritten. Default is False

  • sample_name (str) – Naming scheme for samples.

  • save_format (str) – Choose either pickle or mat.

  • print_progress (bool) – Print progress-bar to terminal. Default is True.

Return type

None

set_sample_function#

set_sample_function(self, sample_function)#

Set sample generating function. The sampling function produces a sample result for each sample definition in the sampling_plan and is called in the method sample_data().

It is important that the sample function only uses keyword arguments with the same name as previously defined in the sampling_plan.

Example:

sp = do_mpc.sampling.SamplingPlanner()

sp.set_sampling_var('alpha', np.random.randn)
sp.set_sampling_var('beta', lambda: np.random.randint(0,5))

sampler = do_mpc.sampling.Sampler(plan)

def sample_function(alpha, beta):
    return alpha*beta

sampler.set_sample_function(sample_function)
Parameters

sample_function (Callable[[Union[FunctionType, BuiltinMethodType], Union[FunctionType, BuiltinMethodType]], Union[FunctionType, BuiltinMethodType]]) – Function to create each sample of the sampling plan.

Return type

None

Attributes#

data_dir#

Sampler.data_dir#

Set the save directory for the results. If the directory does not exist yet, it is created. If the directory is nested all (non-existing) parent folders are also created.

Example:

sampler = do_mpc.sampling.Sampler()
sampler.data_dir = './samples/experiment_1/'

This will set the directory to the indicated path. If the path does not exist, all folders are created.