def test___init__():
""" Test initialization of LRSpectrum """
good_test = 'lrspectrum/test/data/example_methane.log'
# Error checking
with pytest.raises(TypeError):
LRSpectrum(0)
# Check initialized variables
lr = LRSpectrum(good_test, program='dummy')
assert lr.logfile == [good_test]
assert lr.name is None
assert lr.roots == {}
assert lr.freq is None
assert lr.spect is None
assert lr.broad is None
assert lr.wlim is None
assert lr.res is None
# Check initialization for passing a list
lr = LRSpectrum([good_test] * 2, program='dummy')
assert lr.logfile == [good_test] * 2
def test__lorentz():
""" Test generation of lorentzian distribution """
# Because this is a hidden method, it is not meant to be called directly,
# and type checking is not performed
# Test values
single_root = 'lrspectrum/test/data/single_root.log'
lr = LRSpectrum(single_root, program='Gaussian')
lr.freq = np.array([0])
# Un-normalized (0,1) lorentz should be 1 at x=0
result = lr._lorentz(1, 0, 1) * np.pi
assert np.allclose(np.ones((1, )), result)
# Test non-zero, normalized and scaled
lr.freq = np.array([1.5])
expected = np.array([0.01565458])
result = lr._lorentz(1, 0.3, 0.12)
assert np.allclose(expected, result)
This may be convienent when trying to solve for a large number states in
a dense region of the spectrum (like XAS, for example)
lrspectrum must be installed (or a symbolic link to ../lrspectrum/lrspectrum.py
must be included in the calling directory)
"""
import os
import re
import matplotlib.pyplot as plt
from lrspectrum import LRSpectrum
# LRSpectrum supports logfiles listed as multiple params
lr0 = LRSpectrum('example_0.log',
'example_1.log',
'example_2.log',
name='LogParams')
# Or as a list of logfiles.
logs = ['example_0.log', 'example_1.log', 'example_2.log']
lr1 = LRSpectrum(logs, name='LogList')
# The second is particularly convenient when your log files are systematically
# named. One option to take advantage of this is as follows.
lglst = []
for fil in os.listdir('.'):
rexp = re.compile('example_\d.log')
if rexp.match(fil) is not None:
lglst.append(fil)
"""
This is an example of generating one plot with multiple spectra
lrspectrum must be installed (or a symbolic link to ../lrspectrum
must be included in the calling directory)
"""
import os
import re
from lrspectrum import LRSpectrum
# Multiple instances of LRSpectrum can be created with different logfiles
lr0 = LRSpectrum('example_ammonia.log', name='ammonia')
lr1 = LRSpectrum('example_formaldehyde.log', name='formaldehyde')
lr2 = LRSpectrum('example_methane.log', name='methane')
# Generating the spectra is the same as before
lr0.gen_spect()
lr1.gen_spect()
lr2.gen_spect()
# Turn off roots with sticks=False
lr0.plot(sticks=False)
lr1.plot(sticks=False)
lr2.plot(sticks=False, show=True)
# While the above works alright for small sets of logfiles, a better method can
# iterate through a list of LRSpectrum to minimize repeated code.
lrlst = []
rexp = re.compile('example_[^\W\d_]+.log')
def test_parse_log():
""" Test LRSpectrum.parse_log """
good_test = 'lrspectrum/test/data/example_methane.log'
# Error checking
lr = LRSpectrum(good_test, program='dummy')
with pytest.raises(TypeError):
lr.parse_log(program=0)
with pytest.raises(ValueError):
lr.parse_log(program='Fake program')
# Check parser independent values
lr.logfile = [good_test]
lr.parse_log(program='testing')
assert lr.roots == {'1': 1, '2': 1, '3': 2, '4': 3, '5': 5}
# Finally, check that the update call correctly overlaps dicts
lr.logfile = [good_test, good_test]
lr.parse_log(program='testing')
assert lr.roots == {'1': 1, '2': 1, '3': 2, '4': 3, '5': 5}
def test_gen_spect():
""" Test LRSpectrum.gen_spect """
single_root = 'lrspectrum/test/data/single_root.log'
lr = LRSpectrum(single_root, program='Gaussian')
lr.gen_spect(broad=0.5)
# Test error raising
# Test broad
with pytest.raises(TypeError):
lr.gen_spect(broad={'a': 0.5})
# Test wlim
with pytest.raises(TypeError):
lr.gen_spect(wlim=1)
with pytest.raises(TypeError):
lr.gen_spect(wlim='a')
with pytest.raises(IndexError):
lr.gen_spect(wlim=[1])
# Test res
with pytest.raises(TypeError):
lr.gen_spect(res='a')
# Test meth
with pytest.raises(TypeError):
lr.gen_spect(meth=5)
with pytest.raises(ValueError):
lr.gen_spect(meth='ThIs iS nOt A sUpPoRtEd MeThOd')
# Test automatic spectrum generation
expected = np.array([2.9451285853942224, 7.418871414605778], dtype='float')
actual = np.array([i for i in lr.wlim], dtype='float')
assert np.allclose(expected, actual)
# Test values
wlim = (2.9451285853942224, 7.418871414605778)
# Lorentzian
expected = np.loadtxt('lrspectrum/test/data/single_root_spect.txt')
lr.gen_spect(broad=0.5, wlim=wlim, res=100, meth='lorentz')
assert np.allclose(lr.freq, expected[:, 0])
assert np.allclose(lr.spect, expected[:, 1])
# Gaussian
expected = np.loadtxt('lrspectrum/test/data/single_root_gaussian.txt')
lr.gen_spect(broad=0.5, wlim=wlim, res=100, meth='gaussian')
assert np.allclose(lr.freq, expected[:, 0])
assert np.allclose(lr.spect, expected[:, 1])
def test_plot():
""" Test LRSpectrum.plot internals """
# Because most of the arguments are passed to pyplot, the error checking is
# handled there. We only check the arguments we use to shift and scale
# Test exceptions
single_root = 'lrspectrum/test/data/single_root.log'
lr = LRSpectrum(single_root, program='Gaussian')
# x options
with pytest.raises(TypeError):
lr.plot(do_spect=False, show=False, xscale='a')
with pytest.raises(TypeError):
lr.plot(do_spect=False, show=False, xshift='a')
with pytest.raises(TypeError):
lr.plot(do_spect=False, show=False, xlim=123)
with pytest.raises(IndexError):
lr.plot(do_spect=False, show=False, xlim=[1])
with pytest.raises(TypeError):
lr.plot(do_spect=False, show=False, xlim=(1, 'b'))
# y options
with pytest.raises(TypeError):
lr.plot(do_spect=False, show=False, yscale='a')
with pytest.raises(TypeError):
lr.plot(do_spect=False, show=False, yshift='a')
with pytest.raises(TypeError):
lr.plot(do_spect=False, show=False, ylim=123)
with pytest.raises(IndexError):
lr.plot(do_spect=False, show=False, ylim=[1])
with pytest.raises(TypeError):
lr.plot(do_spect=False, show=False, ylim=(1, 'b'))
# Returns None when calling with no spectrum
result = lr.plot(do_spect=True, show=False)
assert result is None
# Test values
lr.freq = np.array([2, 4, 5, 6, 9])
lr.spect = np.array([0, 1, 2, 3, 4])
xlim = (0, 10)
ylim = (-1, 5)
xshift = 5.3
xscale = 0.5
yscale = 2.1
yshift = 3.9
exlim = tuple([x * xscale + xshift for x in xlim])
eylim = tuple([y * yscale + yshift for y in ylim])
expect_freq = lr.freq * xscale + xshift
exspected = lr.spect * yscale + yshift
# Test modified xlim/ylim and values
ax = lr.plot(do_spect=True,
show=False,
xshift=xshift,
yshift=yshift,
xscale=xscale,
yscale=yscale,
xlim=xlim,
ylim=ylim)
assert np.allclose(exlim, ax.get_xlim())
assert np.allclose(eylim, ax.get_ylim())
assert np.allclose(expect_freq, ax.lines[0].get_xdata())
assert np.allclose(exspected, ax.lines[0].get_ydata())
"""
This script is an example of the minimum needed (in a script) to use LRSpectrum
While many options are available for generating the spectrum, enough defaults
are assumed to allow for generation of the spectrum without any parameters.
For more info on the options available, look in lrspectrum.py or the other
example files.
lrspectrum must be installed (or a symbolic link to ../lrspectrum/lrspectrum.py
must be included in the calling directory)
"""
from lrspectrum import LRSpectrum
lr = LRSpectrum('example_ammonia.log')
# gen_spect must be called before plotting
lr.gen_spect()
# In order to see the plot, specify show=True
lr.plot(show=True)
def test_gen_spect():
""" Test LRSpectrum.gen_spect """
single_root = 'lrspectrum/test/data/single_root.log'
lr = LRSpectrum(single_root, program='Gaussian')
lr.gen_spect(broad=0.5)
# Test error raising
# Test broad
with pytest.raises(TypeError):
lr.gen_spect(broad={'a': 0.5})
# Test wlim
with pytest.raises(TypeError):
lr.gen_spect(wlim=1)
with pytest.raises(TypeError):
lr.gen_spect(wlim='a')
with pytest.raises(IndexError):
lr.gen_spect(wlim=[1])
# Test res
with pytest.raises(TypeError):
lr.gen_spect(res='a')
# Test meth
with pytest.raises(TypeError):
lr.gen_spect(meth=5)
with pytest.raises(ValueError):
lr.gen_spect(meth='ThIs iS nOt A sUpPoRtEd MeThOd')
# Test automatic spectrum generation
expected = np.array([2.9451285853942224, 7.418871414605778], dtype='float')
actual = np.array([i for i in lr.wlim], dtype='float')
assert np.allclose(expected, actual)
# Test values
wlim = (2.9451285853942224, 7.418871414605778)
# Lorentzian
expected = np.loadtxt('lrspectrum/test/data/single_root_spect.txt')
lr.gen_spect(broad=0.5, wlim=wlim, res=100, meth='lorentz')
assert np.allclose(lr.freq, expected[:, 0])
assert np.allclose(lr.spect, expected[:, 1])
# Gaussian
expected = np.loadtxt('lrspectrum/test/data/single_root_gaussian.txt')
lr.gen_spect(broad=0.5, wlim=wlim, res=100, meth='gaussian')
assert np.allclose(lr.freq, expected[:, 0])
assert np.allclose(lr.spect, expected[:, 1])
# Test that a runtime error is raised when automatically generating
# spectral range and all oscillator strengths are zero
lr.roots = {}
try:
lr.gen_spect(broad=0.5)
raise AssertionError(
"Did not raise Runtime Error when determining wlim in " +
"gen_spect() with no roots") # pragma: no cover
except RuntimeError:
pass