coexist.access.AccessSetup#
- class coexist.access.AccessSetup(script_path)[source]#
Bases:
object
Structure storing constant attributes for an ACCES optimisation run.
Code validation and generation are handled too.
- Attributes:
- parameters
pd.DataFrame
The free parameters extracted from the user script.
- parameters_scaled
pd.DataFrame
The free parameters scaled to the phenotype space, such that the initial variance (sigma) is unity.
- scaling
np.ndarray
A vector of values that the free parameters are scaled by; it is the initial variance (sigma) given by the user.
- script
str
The modified user script that will be executed.
- population
int
The number of simulations to be run in parallel in each epoch.
- target
float
The target scaled variance - i.e. decrease the uncertainty from the initial 1 down to target.
- seed: int
The random seed defining a single ACCES run.
- rng: np.random.Generator
The random number generator used, seeded with seed.
- parameters
- __init__(script_path)[source]#
Given a path to a user-defined simulation script, extract the free parameters and generate the ACCES script.
Methods
__init__
(script_path)Given a path to a user-defined simulation script, extract the free parameters and generate the ACCES script.
setup_complete
(population, target, seed)Set up the final attributes before starting the ACCES run - i.e. the ones set in the
Access.learn
method.Return the initial parameter combinations to start CMA-ES with.
validate_parameters
(parameters)Validate the free parameters extracted from a user script (a
pandas.DataFrame
).- static validate_parameters(parameters)[source]#
Validate the free parameters extracted from a user script (a
pandas.DataFrame
).