python_utils

Back to API tree index

Purpose

Numeric and Python utility helpers shared across modules.

Module summary

No module-level description available.

Top-level functions

Function	Summary
abbar	return c1 \* np.conjugate(c2) with numba vectorization
abs2	x.real^2 + x.imag^2
abs2sum	sum x.real^2 + x.imag^2 along last axis
archive_paths	No docstring summary available.
argmax_lastax	get axis=-1 (innermost \`column’) index of max in arr
argmaxnd	get unraveled (tuple) index of maximal element in arr
check_equal	recursive x == y that handles ragged mixed-type containers
flat_in_log	get n samples from log-uniform distribution between vmin and vmax
fmt	formatting number as string
import_matplotlib	Avoid annoying errors due to loading matplotlib on the server
invert_dict	return dictionary with inverted key-value pairs
is_numpy_float	check if type(x) is one of numpy’s float types
is_numpy_int	check if type(x) is one of numpy’s integer types
load_module	load python module by name (string)
load_symmetrix_matrix	load n x n symmetric matrix from 1d array formatted as in store_symmetric_matrix()
merge_dicts_safely	merge multiple dictionaries into one, accepting repeated keys if values are consistent, otherwise raise ValueError
next_power	get next power of 2 following n
nfft_of_rfft	ASSUMES nfft even <=> len(rfft_arr) odd
nfft_of_rfftlen	ASSUMES nfft even <=> rfftlen odd
npy_append_cols	append columns to array from .npy file at path infile if outfile_new given, write the result there; otherwise rewrite infile_npy with appended columns
npy_append_rows	append rows to array from .npy file at path infile if outfile_new given, write the result there; otherwise rewrite infile_npy with appended rows
printarr	wrapper for np.array2string printing
rand_azim	get nangles (int or tuple) samples ~ U(0, 2\*np.pi)
rand_cos	get nangles (int or tuple) samples ~ U(-1, 1)
rand_polar	get nangles (int or tuple) samples ~ np.arccos(U(-1, 1))
rfftlen_of_nfft	ASSUMES nfft even <=> len(rfft_arr) odd
store_symmetrix_matrix	store n x n symmetric matrix as 1d array of length n\*(n+1)/2
tukwin_back	back taper of tukey window (last half of hann)
tukwin_bandpass	taper_width is frequency interval length in Hz to be tapered at each end f_nyq & rfft_len are the nyquist frequency and length of the rfftfreq array
tukwin_front	front taper of tukey window (first half of hann)
tukwin_npts	get tukey window based on numbers of total and windowed points
zeropad_end	pad arr axis=-1 with zeros at (right/back) end to length pad_to_N if arr.shape[-1] < pad_to_N, raise ValueError
zip_to_array	O(10^3)x faster np.array([row for row in np.broadcast(\*args)]) usable when args are all at most 1d.

No public classes were found.