latqcdtools.statistics.bootstr

_autoSeed(seed) -> int:
'''
We use seed=None to flag the seed should be automatically chosen. The problem is that we need
seed to be an integer when enforcing that different bootstrap samples use different seeds. 
'''

bootstr(func, data, numb_samples, sample_size=None, same_rand_for_obs=False, conf_axis=1, return_sample=False, seed=None, err_by_dist=False, args=(), nproc=6):
'''
Bootstrap for arbitrary functions. This routine resamples the data and passes them to the function in the same
format as in the input. So the idea is to write a function that computes an observable from a given data set. This
function can be put into this bootstrap routine and will get bootstrap samples as input. Based on the output of the
function, the bootstrap mean and error are computed. The function may return multiple observables that are either
scalars or numpy objects. You can pass a multidimensional object as data, but the bootstrap function has to know
which axis should be resampled which is controlled by conf_axis (default = 0 for one dimensional arrays and
default = 1 for higher order arrays.)

    Parameters
    ----------
    func : callable
        The function that calculates the observable

    data : array_like
        Input data

    numb_samples : integer
        Number of bootstrap samples

    sample_size : integer, optional, default = 0
        Size of sample

    same_rand_for_obs : boolean, optional, default = False
        Use the same random numbers for each observable accessed by index conf_axis - 1. Please note:
            - Objects that are accessed by an axis >= conf_axis + 1 do always have the same random numbers.
            - Objects that are accessed by axis conf_axis < conf_axis - 1 never share the same random numbers.

    conf_axis : integer, optional, default = 0 for dim(data) = 1 or default = 1 for dim(data) >= 2
        Axis that should be resampled

    return_sample : boolean, optional, default = False
        Along with the mean and the error also return the results from the individual samples

    seed: integer, optional, default = None
        seed for the random generator. If None, the default seed from numpy is used (probably from time)

    same_rand_for_obs : boolean, optional, default = False
        same random numbers per observable

    err_by_dist : boolean, optional, default = False
        Compute the error from the distribution using the median and the 68% quantile

    args : array_like or dict, default = ()
        optional arguments to be passed to func. If a dictionary the are passed as **args.

    nproc : integer
        Number of threads to use if you choose to parallelize. nproc=1 turns off parallelization.
'''

bootstr_from_gauss(func, data, data_std_dev, numb_samples, sample_size=1, same_rand_for_obs=False, return_sample=False, seed=None, err_by_dist=True, useCovariance=False, Covariance=None, args=(), nproc=6, asym_err=False):
'''
Same as standard bootstrap routine, but the data are generated by gaussian noise around the mean values in data.
The width of the distribution is controlled by data_std_dev. Note, that the function has to average over samples.
This means that data_std_dev should always be the standard deviation of a single measurement and not the standard
deviation of a mean.
'''

estimateCovariance(data_std_dev, numb_samples, seed=None):
'''
Estimate a covariance matrix using data and their corresponding std_dev. The idea is to estimate this with a
bootstrap under the assumption that each datum X_i is maximally correlated with every other datum X_j. Here
data should be 1-dimensional.
'''

recurs_append(data, sample_data, axis, conf_axis, sample_size, same_rand_for_obs, i, my_seed):
'''
Recursive function to fill the sample. 
'''

class nimbleBoot(func, data, numb_samples, sample_size, same_rand_for_obs, conf_axis, return_sample, seed, err_by_dist, args, nproc):
'''
'''

class nimbleGaussianBoot(func, data, data_std_dev, numb_samples, sample_size, same_rand_for_obs, return_sample, seed, err_by_dist, useCovariance, Covariance, args, nproc, asym_err):
'''
'''