From c4a6313995363a6de3176e383c6f43d139c55e7b Mon Sep 17 00:00:00 2001
From: MaNyh <nyhuis@tfd.uni-hannover.de>
Date: Mon, 28 Mar 2022 17:59:11 +0200
Subject: [PATCH] better fit
---
.../timeseries/stationary_signal_check.py | 29 +++++++-----
tests/test_ntrfc_postprocessing.py | 45 +++++++++----------
2 files changed, 40 insertions(+), 34 deletions(-)
diff --git a/ntrfc/postprocessing/timeseries/stationary_signal_check.py b/ntrfc/postprocessing/timeseries/stationary_signal_check.py
index e9db44f0..2aca1150 100644
--- a/ntrfc/postprocessing/timeseries/stationary_signal_check.py
+++ b/ntrfc/postprocessing/timeseries/stationary_signal_check.py
@@ -1,5 +1,7 @@
import numpy as np
from matplotlib import pyplot as plt
+import matplotlib
+#matplotlib.use('TkAgg')
from ntrfc.postprocessing.timeseries.integral_scales import integralscales
@@ -22,7 +24,7 @@ def parsed_timeseries_analysis(timesteps, signal, resolvechunks=20, verbose=True
if stationarity==True:
csig=signal
ctime=timesteps
- plot_stationarity_analisys(csig, ctime, signal, stationarity_timestep, timesteps)
+ plot_stationarity_analisys(csig, ctime, signal, stationarity_timestep, timesteps,signal_type)
return stationarity, timescale, stationarity_timestep
@@ -30,12 +32,13 @@ def parsed_timeseries_analysis(timesteps, signal, resolvechunks=20, verbose=True
csig = np.concatenate([*checksigchunks[resolvechunks - i:]])
ctime = np.concatenate([*checktimechunks[resolvechunks - i:]])
- signal_type,newstationarity,newstationarity_timestep, newtimescale,newlengthscale = check_signal_stationarity(resolvechunks, csig, ctime)
+ new_signal_type,newstationarity,newstationarity_timestep, newtimescale,newlengthscale = check_signal_stationarity(resolvechunks, csig, ctime)
if newstationarity:
newscales = (timescale, lengthscale)
stationarity=True
scales = newscales
+ signal_type=new_signal_type
stationarity_timestep = newstationarity_timestep
@@ -43,15 +46,15 @@ def parsed_timeseries_analysis(timesteps, signal, resolvechunks=20, verbose=True
# when no further stationarity found, return status
# when done, return last status
- plot_stationarity_analisys(csig, ctime, signal, stationarity_timestep, timesteps)
+ plot_stationarity_analisys(csig, ctime, signal, stationarity_timestep, timesteps,signal_type)
return stationarity, scales, stationarity_timestep
- plot_stationarity_analisys(csig, ctime, signal, stationarity_timestep, timesteps)
+ plot_stationarity_analisys(csig, ctime, signal, stationarity_timestep, timesteps,signal_type)
return stationarity, scales, stationarity_timestep
-def plot_stationarity_analisys(csig, ctime, signal, stationarity_timestep, timesteps):
+def plot_stationarity_analisys(csig, ctime, signal, stationarity_timestep, timesteps, signaltype):
plt.figure()
plt.plot(timesteps, signal)
plt.plot(ctime, csig, color="black", linewidth=4)
@@ -61,6 +64,10 @@ def plot_stationarity_analisys(csig, ctime, signal, stationarity_timestep, times
if ymax-ymin<0.01:
ymax=0.5+np.mean(signal)
ymin =-0.5+np.mean(signal)
+ else:
+ ymin -=1
+ ymax += 1
+
if sts==0.0 or sts:
plt.vlines(sts, ymin=ymin, ymax=ymax,
linewidth=4, color="k", linestyles="dashed")
@@ -69,9 +76,9 @@ def plot_stationarity_analisys(csig, ctime, signal, stationarity_timestep, times
else:
plt.axvspan(0, timesteps[-1],
facecolor='red', alpha=0.5)
-
+ plt.title(signaltype)
plt.ylim(ymin, ymax)
- plt.show() #
+ plt.show()
def check_signal_stationarity(resolvechunks, signal, timesteps, verbose = True):
@@ -93,9 +100,9 @@ def check_signal_stationarity(resolvechunks, signal, timesteps, verbose = True):
# todo: now it is only mean, val and var that is being investigated.
# it makes sense to also investigate the behaviour of the autocorrelation
# but as the signal is divided into chunks, one has to
- const_mean = np.allclose(mean, means,rtol=0.05)
- const_val = np.allclose(mean, signal,rtol=0.05)
- const_var = np.allclose(var,vars,rtol=0.4)
+ const_mean = np.allclose(mean, means,rtol=0.02)
+ const_val = np.allclose(mean, signal,rtol=0.02)
+ const_var = np.allclose(var,vars,rtol=0.02)
#
if const_mean and const_var:
timescale, lengthscale = integralscales(signal, timesteps)
@@ -107,7 +114,7 @@ def check_signal_stationarity(resolvechunks, signal, timesteps, verbose = True):
timescale, lengthscale = 0,0
timescales, lengthscales = np.zeros(resolvechunks),np.zeros(resolvechunks)
- const_tscales = np.allclose(timescales, timescale,rtol=0.1)
+ const_tscales = np.allclose(timescales, timescale,rtol=0.02)
"""
from itertools import product
diff --git a/tests/test_ntrfc_postprocessing.py b/tests/test_ntrfc_postprocessing.py
index 9c4fe577..bd6f2abd 100644
--- a/tests/test_ntrfc_postprocessing.py
+++ b/tests/test_ntrfc_postprocessing.py
@@ -31,7 +31,7 @@ def sine_signal(numtimesteps, time):
def noise_signal(numtimesteps, time):
- mu, sigma = 0, 0.2 # mean and standard deviation
+ mu, sigma = 0, 0.02 # mean and standard deviation
s = np.random.normal(mu, sigma, size=numtimesteps)
Dt = time / numtimesteps
@@ -74,14 +74,14 @@ def tanh_sin_signal(numtimesteps, time):
def tanh_sin_noise_signal(numtimesteps, time):
timesteps = np.linspace(0, time, numtimesteps)
- stationary = time/4
+ stationary = time/10
tanhsignal = np.tanh(timesteps * stationary ** -1)
timesteps = np.linspace(0, 1, numtimesteps) * time
- omega = 1000
- signal = np.sin(timesteps * omega) + 1
+ omega = 100
+ signal = np.sin(timesteps * omega)*0.3 + 1
- mu, sigma = 0, 0.2 # mean and standard deviation
+ mu, sigma = 0, 0.1 # mean and standard deviation
s = np.random.normal(mu, sigma, size=numtimesteps)
Dt = time / numtimesteps
@@ -89,7 +89,7 @@ def tanh_sin_noise_signal(numtimesteps, time):
dt = time / 1000
timescale = int(dt / Dt)
weights = np.repeat(1.0, timescale) / timescale
- noise = np.convolve(s, weights, 'same') + 1
+ noise = np.convolve(s, weights, 'same')
omega = 1000
timesteps = np.linspace(0, time, numtimesteps)
@@ -111,11 +111,11 @@ def sine_abate_signal(numtimesteps, time):
def complex_signal(numtimesteps, time):
timesteps = np.linspace(0, time, numtimesteps)
- stationary = time/4
+ stationary = time/10
tanhsignal = np.tanh(timesteps * stationary ** -1)
timesteps = np.linspace(0, 1, numtimesteps) * time
- omega = 1000
+ omega = 10
signal = np.sin(timesteps * omega) + 1
mu, sigma = 0, 0.2 # mean and standard deviation
@@ -145,32 +145,31 @@ def complex_signal(numtimesteps, time):
def test_analize_stationarity(verbose=True):
- tight_tolerance = 1e-2
loose_tolerance = 1
signals = {"constant": constant_signal(1000, 1),
"ramp": ramp_signal(1000, 1),
"rampconst": ramptoconst_signal(10000, 1),
"sine": sine_signal(10000, 1),
- "noise": noise_signal(10000, 1),
- "convergentsine_signal": convergentsine_signal(10000, 10),
- "tanh":tanh_signal(10000,20),
- "tanh_sin":tanh_sin_signal(10000,10),
- "tanh_sin_noise": tanh_sin_noise_signal(10000, 2.8),
- "sine_abate":sine_abate_signal(20000, 4),
- "complex":complex_signal(40000,60)
+ "noise": noise_signal(100000, 10),
+ "convergentsine_signal": convergentsine_signal(10000, 40),
+ "tanh":tanh_signal(20000,20),
+ "tanh_sin":tanh_sin_signal(20000,32),
+ "tanh_sin_noise": tanh_sin_noise_signal(40000, 480),
+ "sine_abate":sine_abate_signal(30000, 4),
+ "complex":complex_signal(50000,480)
}
expections = {"constant": (True, (0, 0), [0,0]),
"ramp": (False, (0, 0), [-1,-1]),
"rampconst": (True, (0, 0), [0.45,0.55]),
- "sine": (True, (0.0005, 0.0005), [0,0]),
- "noise": (True, (0.0005, 0.0005), [0,0]),
- "convergentsine_signal": (True, (0.0005, 0.0005), [5,5.5]),
- "tanh":(True,(0,0),[3.5,4.5]),
- "tanh_sin":(True,(0.0005,0.0005),[4,4.75]),
- "tanh_sin_noise":(True,(0.0005,0.0005),[0.8,1.2]),
+ "sine": (True, (0.005, 0.005), [0,0]),
+ "noise": (True, (0.005, 0.005), [0,0]),
+ "convergentsine_signal": (True, (0.0005, 0.0005), [15.5,16.5]),
+ "tanh":(True,(0,0),[4.5,5.5]),
+ "tanh_sin":(True,(0.0005,0.0005),[15.5,16.5]),
+ "tanh_sin_noise":(True,(0.0005,0.0005),[110,130]),
"sine_abate":(True,(0.00025,0.00025),[0.35,1]),
- "complex":(True,(0.00025,0.0025),[20,26])
+ "complex":(True,(0.00025,0.0025),[90,130])
}
for name, sig in signals.items():
--
GitLab