Low-level analysis functions

Analysis routines.

class py_alf.ana.Parameters(directory, obs_name=None)

Object representing the “parameters” file.

Parameters:

directorypath-like object

Directory of “parameters” file.

obs_namestr, optional

Observable name. If this is set, the object tries to get a parameters not from the namelist ‘var_errors’, but from a namelist called obs_name, while ‘var_errors’ is the fallback options. Parameters will be written to namelist obs_name.

N_min()

Get minimal number of bins, given the parameters in this object.

N_rebin()

Get N_rebin.

N_skip()

Get N_skip.

set_N_rebin(parameter)

Update N_rebin.

set_N_skip(parameter)

Update N_skip.

write_nml()

Write namelist to file. Preseves comments.

class py_alf.ana.ReadObs(directory, obs_name, bare_bins=False, substract_back=True)

Read, skip, rebin and jackknife scalar-type bins.

Bins get skipped and rebinned according to N_skip an N_rebin retrieved through Parameters, then jackknife resampling is applied. Saves jackknife bins.

Cf. read_scal(), read_latt(), read_hist().

Parameters:

directorypath-like object

Directory containing the observable.

obs_namestr

Name of observable.

bare_binsbool, default=False

Do not perform skipping, rebinning, or jackknife resampling.

substract_backbool, default=True

Substract background. Applies to correlation functions.

all()

Return all bins.

jack(N_rebin)

Return jackknife bins. Object has to be created with bare_bins=True.

Parameters:

N_rebinint

Overwrite N_rebin from parameters.

slice(n)

Return n-th bin.

py_alf.ana.ana_eq(directory, obs_name, sym=None)

Analyze given equal-time correlators.

If sym is given, it symmetrizes the bins prior to calculating the error. Cf. symmetrize().

py_alf.ana.ana_hist(directory, obs_name)

Analyze given histogram observables.

py_alf.ana.ana_scal(directory, obs_name)

Analyze given scalar observables.

Parameters:

directorypath-like object

Directory containing the observable.

obs_namestr

Name of the observable.

py_alf.ana.ana_tau(directory, obs_name, sym=None)

Analyze given time-displaced correlators.

If sym is given, it symmetrizes the bins prior to calculating the error. Cf. symmetrize().

py_alf.ana.error(jacks, imag=False)

Calculate expectation values and errors of given jackknife bins.

Parameters:

jacksarray-like object

Jackknife bins.

imagbool, default=False

Output with imaginary part.

Returns:

tuple of numpy arrays

(expectation values, errors).

py_alf.ana.jack(X, par, N_skip=None, N_rebin=None)

Create jackknife bins out of input bins after skipping and rebinning.

Parameters:

Xarray-like object

Input bins. Bins run over first index.

parParameters

Parameters object.

N_skipint, default=par.N_skip()

Number of bins to skip.

N_rebinint, default=par.N_rebin()

Number of bins to recombine into one.

Returns:

numpy array

Jackknife bins after skipping and rebinning.

py_alf.ana.load_res(directories)

Read analysis results from multiple simulations.

Read from pickled dictionaries ‘res.pkl’ and return everything in a single pandas DataFrame with one row per simulation.

Parameters:

directorieslist of path-like objects

Directories containing analyzed simulation results.

Returns:

dfpandas.DataFrame

Contains analysis results and Hamiltonian parameters. One row per simulation.

py_alf.ana.read_hist(directory, obs_name, bare_bins=False)

Read, skip, rebin and jackknife histogram-type bins.

Bins get skipped and rebinned according to N_skip an N_rebin retrieved through Parameters, then jackknife resampling is applied.

Parameters:

directorypath-like object

Directory containing the observable.

obs_namestr

Name of the observable.

bare_binsbool, default=False

Do not perform skipping, rebinning, or jackknife resampling.

Returns:

array

Observables. shape: (N_bins, N_classes).

array

Sign. shape: (N_bins,).

array

Proportion of observations above upper bound. shape: (N_bins,).

array

Proportion of observations below lower bound. shape: (N_bins,).

N_classesint

Number of classes between upper and lower bound.

upperfloat

Upper bound.

lowerfloat

Lower bound.

py_alf.ana.read_latt(directory, obs_name, bare_bins=False, substract_back=True)

Read, skip, rebin and jackknife lattice-type bins (_eq and _tau).

Bins get skipped and rebinned according to N_skip an N_rebin retrieved through Parameters, then jackknife resampling is applied.

Parameters:

directorypath-like object

Directory containing the observable.

obs_namestr

Name of the observable.

bare_binsbool, default=False

Do not perform skipping, rebinning, or jackknife resampling.

substract_backbool, default=True

Substract background from correlation functions.

Returns:

array

Observables. shape: (N_bins, N_orb, N_orb, N_tau, latt.N).

array

Background. shape: (N_bins, N_orb)

array

Sign. shape: (N_bins,).

N_orbint

Number of orbitals.

N_tauint

Number of imaginary time steps.

dtaufloat

Imaginary time step length.

lattLattice

See py_alf.Lattice.

py_alf.ana.read_scal(directory, obs_name, bare_bins=False)

Read, skip, rebin and jackknife scalar-type bins.

Bins get skipped and rebinned according to N_skip an N_rebin retrieved through Parameters, then jackknife resampling is applied.

Parameters:

directorypath-like object

Directory containing the observable.

obs_namestr

Name of the observable.

bare_binsbool, default=False

Do not perform skipping, rebinning, or jackknife resampling.

Returns:

array

Observables. shape: (N_bins, N_obs).

array

Sign. shape: (N_bins,).

N_obsint

Number of observables.

py_alf.ana.rebin(X, N_rebin)

Combine each N_rebin bins into one bin.

If the number of bins (=N0) is not an integer multiple of N_rebin, the last N0 modulo N_rebin bins are discarded.

py_alf.ana.symmetrize(latt, syms, dat)

Symmetrize a dataset.

Parameters:

lattLattice

See py_alf.Lattice.

symslist

List of symmetry operations, including the identity of the form sym(latt, i) -> i_tranformed

datarray-like object

Data to symmetrize. The symmetrization is with respect to the last index of dat.

Returns:

dat_symnumpy array

Symmetrized data.

\(\phantom{\xi}\)