Comparator
Install
python setup.py install
If you're using poetry, you can add comparator to your git sources:
# pyproject.toml
[tool.poetry.dependencies]
comparator = {git = "git@github.com:dean-shaff/comparator.git"}Now calling poetry update will install comparator
Usage
comparator offers a simple interface for comparing single dimensional
arrays, either real or complex. Each comparator object has associated
with it "operators", "products", and a "domain". "Operators" are operations
that apply to pairs of arrays, producing a single array as output. Examples
might be cross correlation and difference squared. "Products"
are operation that apply to single arrays, producing a single number as output.
Examples include mean and max. The "domain" is the subset of input arrays on
which to operate.
import matplotlib.pyplot as plt
import comparator.util
from comparator import TimeFreqDomainComparator
comp = TimeFreqDomainComparator()
comp.freq.domain = [0, 2**15]
# we can also set the domain with a slice object:
comp.freq.domain = slice(0, None)
# we can also set the domain to a fraction of the size of the input arrays:
comp.freq.domain = [0.1, 0.8]
# Alternatively, do the following:
comp.freq.set_fft_size(2**15)
# operators can map from any number of inputs to a single output.
# Its convenient to operate on the input arrays themselves, hence 'this'
comp.operators["this"] = lambda a: a
comp.operators["diff"] = lambda a, b: a - b
comp.products["mean"] = np.mean
# only the time domain will have the cross correlation ("xcorr") operator
comp.time.operators["xcorr"] = lambda a, b: scipy.signal.fftconvolve(a, b.conj)
# only the time domain will have the "max" product
comp.time.products["max"] = np.amax
# generate some random data to compare
a, b, c = [np.random.rand(100) + 1j*np.random.rand(100) for i in range(3)]
res_prod, res_op = comp(a, b, c)
# print out the result product matrix or array in scientific notation
print(f"{res_prod:.4e}")
# util has a corner plot utility
comparator.util.corner_plot(res_op)
plt.show()
# Instead of calling the Comparator object itself, we could get something
# specific from each of the subdomains:
comp.freq(a, b, c)
comp.freq(a, b, c)
comp.time(a, b)