def test_log():
# test log
set_loglevel("WARNING")
assert get_loglevel() == 30
warning_('Ok, this is nicely executing!')
set_loglevel(10)
assert get_loglevel() == 10
def test_logger(caplog):
logger = logging.getLogger("SpectroChemPy")
logger.propagate = True
caplog.set_level(DEBUG)
# We can set the level using strings
set_loglevel("DEBUG")
assert logger.handlers[0].level == INFO # DEBUG only on the file
assert logger.handlers[1].level == DEBUG
set_loglevel(WARNING)
assert logger.handlers[0].level == WARNING
assert logger.handlers[1].level == WARNING
error_("\n" + "*" * 80 + "\n")
debug_("debug in WARNING level - should not appear")
info_("info in WARNING level - should not appear")
warning_("OK this is a Warning")
error_("OK This is an Error")
error_("\n" + "*" * 80 + "\n")
set_loglevel(INFO)
assert logger.handlers[0].level == INFO
assert logger.handlers[1].level == INFO
debug_("debug in INFO level - should not appear on stdout")
info_("OK - info in INFO level")
warning_("OK this is a Warning")
error_("OK This is an Error")
error_("\n" + "*" * 80 + "\n")
set_loglevel("DEBUG")
assert logger.handlers[0].level == INFO
assert logger.handlers[1].level == DEBUG
debug_("OK - debug in DEBUG level")
info_("OK - info in DEBUG level")
assert caplog.records[-1].levelname == "INFO"
assert caplog.records[-1].message.endswith("OK - info in DEBUG level")
warning_("OK this is a Warning")
assert caplog.records[-1].levelname == "WARNING"
assert caplog.records[-1].message.endswith("OK this is a Warning")
error_("OK This is an Error")
assert caplog.records[-1].levelname == "ERROR"
assert caplog.records[-1].message.endswith("OK This is an Error")
def test_logger(caplog):
logger = logging.getLogger('SpectroChemPy')
logger.propagate = True
caplog.set_level(logging.DEBUG)
# We can set the level using strings
set_loglevel("DEBUG")
assert logger.level == logging.DEBUG
set_loglevel(WARNING)
assert logger.level == logging.WARNING
error_('\n' + '*' * 80 + '\n')
debug_('debug in WARNING level - should not appear')
info_('info in WARNING level - should not appear')
warning_('OK this is a Warning')
error_('OK This is an Error')
error_('\n' + '*' * 80 + '\n')
set_loglevel(INFO)
assert logger.level == logging.INFO
debug_('debug in INFO level - should not appear')
info_('OK - info in INFO level')
warning_('OK this is a Warning')
error_('OK This is an Error')
error_('\n' + '*' * 80 + '\n')
set_loglevel('DEBUG')
assert logger.level == logging.DEBUG
debug_('OK - debug in DEBUG level')
info_('OK - info in DEBUG level')
assert caplog.records[-1].levelname == 'INFO'
assert caplog.records[-1].message == 'OK - info in DEBUG level'
warning_('OK this is a Warning')
assert caplog.records[-1].levelname == 'WARNING'
assert caplog.records[-1].message == 'OK this is a Warning'
error_('OK This is an Error')
assert caplog.records[-1].levelname == 'ERROR'
assert caplog.records[-1].message == 'OK This is an Error'
#
# This is a basic behavior of python : on way to avoid. stoppping the execution without displaying a message is :
# %%
try:
from spectrochempy import toto # noqa: F811, F401
except Exception:
pass
# %% [markdown]
# ## API Configuration
#
# Many options of the API can be set up
# %%
scp.set_loglevel(scp.INFO)
# %% [markdown]
# In the above cell, we have set the **log** level to display ``info`` messages, such as this one:
# %%
scp.info_('this is an info message!')
scp.debug_('this is a debug message!')
# %% [markdown]
# Only the info message is displayed, as expected.
#
# If we change it to ``DEBUG``, we should get the two messages
# %%
scp.set_loglevel(scp.DEBUG)
# ```python
# scp.set_loglevel('DEBUG')
# scp.set_loglevel(scp.DEBUG)
# scp.set_loglevel(10)
# ```
# The current loglevel can be obtained with the `scp.get_loglevel()` function.
#
# The following instruction prints the current loglevel
# %%
print(f"Default: {scp.get_loglevel()}") # print the current loglevel
# %% [markdown]
# It yields `30` the numerical value corresponding to the `WARNING` level. Now, the next instruction
# lowers the loglevel to `ÌNFO`:
# %%
scp.set_loglevel(scp.INFO) # set loglevel to 'INFO'
# %% [markdown]
# We see that the API then delivers the INFO message: `changed default log_level to INFO`.
# %% [markdown]
# And finally, the next instructions reset the loglevel to `WARNING` level (default), and print it.
# As seen below, no message `changed default log_level to ...` is delivered
# %%
scp.set_loglevel("WARNING") # reset to default
print(f"New loglevel: {scp.get_loglevel()}")
# %% [markdown]
# It is also possible to issue such messages in scripts. In the cell below, we set the loglevel to `INFO` and try to
# print two types of messages:
def test_MCRALS():
"""Test MCRALS with synthetic data"""
def gaussian(x, h, c, w, noise):
return h * (np.exp(-1 / 2 * ((x.data - c) / w) ** 2)) + noise * np.random.randn(
len(x)
) # a gaussian with added noise
def expon(t, c0, l, noise):
return c0 * (np.exp(l * t.data)) + noise * np.random.randn(len(t.data))
def get_C(C):
return C
n_PS = 2 # number of pure species
n_t = 10
n_wl = 10
h = [1, 1]
c = [250, 750]
w = [100, 100]
noise_spec = [0.0, 0.0]
noise_conc = [0.0, 0.0]
c0 = [10, 1]
l = np.array([-2, 2]) * 1e-2
t_c = Coord(np.arange(0, 100, 100 / n_t), title="time", units="s")
wl_c = Coord(np.arange(0, 1000, 1000 / n_wl), title="wavelength", units="nm")
PS_c = Coord(
range(n_PS),
title="species",
units=None,
labels=["PS#" + str(i) for i in range(n_PS)],
)
St = NDDataset.zeros((n_PS, len(wl_c)), coordset=(PS_c, wl_c))
C = NDDataset.zeros((len(t_c), n_PS), coordset=(t_c, PS_c))
St0 = NDDataset.zeros((n_PS, len(wl_c)), coordset=(PS_c, wl_c))
C0 = NDDataset.zeros((len(t_c), n_PS), coordset=(t_c, PS_c))
for i, id in enumerate((0, 1)):
C.data[:, i] = expon(t_c, c0[id], l[id], noise_conc[id])
St.data[i, :] = gaussian(wl_c, h[id], c[id], w[id], noise_spec[id])
C0.data[:, i] = expon(t_c, c0[id], l[id], 0)
St0.data[i, :] = gaussian(wl_c, h[id], c[id], w[id], 0)
D = dot(C, St)
D.title = "intensity"
#############################
# Test normal functioning
# guess = C0
mcr = MCRALS(D, C0, tol=30.0)
assert "converged !" in mcr.log[-15:]
# test attributes
for attr in [
"log",
"logs",
"params",
"Chard",
"Stsoft",
"St",
"C",
"extOutput",
"fixedC",
"X",
]:
assert hasattr(mcr, attr)
# test plot
mcr.plotmerit()
show()
# test diverging
mcr = MCRALS(
D,
C0,
monoIncConc=[0, 1],
monoIncTol=1.0,
unimodSpec=[0, 1],
normSpec="euclid",
closureConc=[0, 1],
closureMethod="constantSum",
maxdiv=1,
)
assert "Stop ALS optimization" in mcr.log[-40:]
# guess = C0, hard modeling
mcr = MCRALS(D, C0, hardConc=[0, 1], getConc=get_C, argsGetConc=(C0,), tol=30.0)
assert "converged !" in mcr.log[-15:]
# guess = C, test with deprecated parameters
# and few other parameters set to non-default values to improve coverage
mcr = MCRALS(
D,
C0,
# deprecated:
unimodMod="strict",
unimodTol=1.1,
verbose=True,
# other parameters set to non-default values
monoIncConc=[0],
monoDecConc=[1],
closureConc=[0, 1],
nonnegConc=None,
unimodConc=None,
unimodSpec="all",
nonnegSpec=None,
normSpec="max",
maxit=1,
)
set_loglevel("WARN")
# guess = C0.data, test with other parameters
mcr = MCRALS(
D,
C0.data,
normSpec="euclid",
closureConc=[0, 1],
closureMethod="constantSum",
maxit=1,
)
assert "Convergence criterion ('tol')" in mcr.log[-100:]
# guess = St as ndarray
mcr = MCRALS(D, St0.data, tol=15.0)
assert "converged !" in mcr.log[-15:]
#########################
# Text exceptions
# guess with wrong number of dimensions
try:
mcr = MCRALS(
D,
np.random.rand(n_t - 1, n_PS),
)
except ValueError as e:
assert e.args[0] == "the dimensions of guess do not match the data"
# guess with wrong nonnegConc parameter
try:
mcr = MCRALS(D, C, nonnegConc=[2])
except ValueError as e:
assert "please check nonnegConc" in e.args[0]
try:
mcr = MCRALS(D, C, nonnegConc=[0, 1, 1])
except ValueError as e:
assert "please check nonnegConc" in e.args[0]
# guess with wrong unimodConc parameter
try:
mcr = MCRALS(D, C, unimodConc=[2])
except ValueError as e:
assert "please check unimodConc" in e.args[0]
try:
mcr = MCRALS(D, C, unimodConc=[0, 1, 1])
except ValueError as e:
assert "please check unimodConc" in e.args[0]
# wrong closureTarget
try:
mcr = MCRALS(D, C, closureTarget=[0, 1, 1])
except ValueError as e:
assert "please check closureTarget" in e.args[0]
# wrong hardC_to_C_idx
try:
mcr = MCRALS(D, C, hardC_to_C_idx=[2])
except ValueError as e:
assert "please check hardC_to_C_idx" in e.args[0]
try:
mcr = MCRALS(D, C, hardC_to_C_idx=[0, 1, 1])
except ValueError as e:
assert "please check hardC_to_C_idx" in e.args[0]
# wrong unimodSpec
try:
mcr = MCRALS(D, C, unimodSpec=[2])
except ValueError as e:
assert "please check unimodSpec" in e.args[0]
try:
mcr = MCRALS(D, C, unimodSpec=[0, 1, 1])
except ValueError as e:
assert "please check unimodSpec" in e.args[0]
# wrong nonnegSpec
try:
mcr = MCRALS(D, C, nonnegSpec=[2])
except ValueError as e:
assert "please check nonnegSpec" in e.args[0]
try:
mcr = MCRALS(D, C, nonnegSpec=[0, 1, 1])
except ValueError as e:
assert "please check nonnegSpec" in e.args[0]
# -*- coding: utf-8 -*-
from subprocess import Popen, PIPE
# ======================================================================================================================
# Copyright (©) 2015-2021 LCS - Laboratoire Catalyse et Spectrochimie, Caen, France. =
# CeCILL-B FREE SOFTWARE LICENSE AGREEMENT - See full LICENSE agreement in the root directory =
# ======================================================================================================================
import pytest
from spectrochempy import APIref, set_loglevel, get_loglevel, warning_, version
set_loglevel('WARNING')
def test_api():
assert 'EFA' in APIref
assert 'CRITICAL' in APIref
assert 'np' in APIref
assert 'NDDataset' in APIref
assert 'abs' in APIref
def test_datadir():
from spectrochempy.application import DataDir
print(DataDir().listing())
def test_version():
# test version
assert len(version.split('.')) >= 3
X.y.default ######################################################################################################################## # Let's now plot the spectral range of interest. The default coordinate is now used: X_ = X[:, 2250.0:1950.0] print(X_.y.default) _ = X_.plot() _ = X_.plot_map() ############################################################################### # IRIS analysis without regularization # ------------------------------------ ######################################################################################################################## # Perform IRIS without regularization (the loglevel can be set to `INFO` to have information on the running process) scp.set_loglevel(scp.INFO) iris = scp.IRIS(X_, "langmuir", q=[-8, -1, 50]) ######################################################################################################################## # Plots the results iris.plotdistribution() _ = iris.plotmerit() ############################################################################### # With regularization and a manual search # --------------------------------------- ######################################################################################################################## # Perform IRIS with regularization, manual search iris = scp.IRIS(X_, "langmuir", q=[-8, -1, 50], reg_par=[-10, 1, 12]) iris.plotlcurve(title="L curve, manual search")
# %%
St0 = A[3]
_ = St0.plot()
# %% [markdown]
# Note that, again, no information has been given as to the ordinate and abscissa data. We could add them as previously
# but this is not very important. The key point is that the 'wavelength' dimension is compatible with the data 'X',
# which is indeed the case (both have a length of 95). If it was not, an error would be generated in the following.
#
# #### ALS Optimization
# With this guess 'St0' and the dataset 'X' we can create a MCRALS object. At this point of the tutorial, we will use
# all the default parameters. We switch to `log_level = "INFO"` have a summary of the ALS iterations:
# %%
scp.set_loglevel("INFO")
mcr = scp.MCRALS(X, St0)
# %% [markdown]
# The optimization has converged within few iterations. The figures reported for each iteration are defined as follows:
#
# - 'Error/PCA' is the standard deviation of the residuals with respect to data reconstructed by a PCA with as many
# components as pure species (4 in this example),
#
# - 'Error/exp': is the standard deviation of the residuals with respect to the experimental data X,
#
# - '%change': is the percent change of 'Error/exp' between 2 iterations
#
# The default is to stop when this %change between two iteration is negative (so that the solution is improving),
# but with an absolute value lower than 0.1% (so that the improvement is considered negligible). This parameter -
# as well as several other parameters affecting the ALS optimization can be changed by the setting the 'tol' value.
# -*- coding: utf-8 -*-
# flake8: noqa
from subprocess import Popen, PIPE
import pytest
from spectrochempy import APIref, set_loglevel, get_loglevel, warning_, version
set_loglevel("WARNING")
def test_api():
assert "EFA" in APIref
assert "CRITICAL" in APIref
assert "NDDataset" in APIref
assert "abs" in APIref
def test_datadir():
# test print a listing of the testdata directory
from spectrochempy.application import DataDir
print(DataDir().listing())
# or simply
print(DataDir())
assert str(DataDir()).startswith("testdata")