def test_nothing():
n = 128
data = np.random.randn(n)
Gf = fit_gaussian.fit_gaussian(data)
assert Gf.amplitude() < 4
assert np.std(Gf.histogram(n)-data)<4
def rate_profile(self,hdr,candidate,profile,std,cache):
if "gaussian" in cache:
G = cache["gaussian"]
else:
G = fit_gaussian.fit_gaussian(profile)
cache["gaussian"] = G
return self.rate_gaussian_profile(hdr,candidate,profile,std,G,cache)
def check_fit_quality(G,noise_amp,n):
true = G.histogram(n)
noise = np.random.randn(n)*noise_amp*G.amplitude()
data = true + noise
Gf = fit_gaussian.fit_gaussian(data)
assert np.abs(Gf.amplitude()-G.amplitude())/G.amplitude()<0.01
def plot_residuals(data):
n = len(data)
xs = np.arange(n)/float(n)
test_ks = np.exp(np.linspace(np.log(0.1),np.log(1000),1000))
G = fit_gaussian.fit_gaussian(data)
plt.figure()
resids = [fit_gaussian.rms_residual(k,data) for k in test_ks]
plt.loglog(test_ks,resids,color="green")
best_k = test_ks[np.argmin(resids)]
plt.axvline(best_k,color="green")
plt.axvline(G.k,color="cyan")
plt.figure()
mue, ae, be = fit_gaussian.fit_all_but_k(best_k, data)
plt.plot(xs, data, color="black", label="data")
plt.plot(xs, (ae*fit_gaussian.vonmises_histogram(best_k,mue,n)+be),
color="green", label="exhaustive best fit")
plt.plot(xs, G.histogram(n), color="cyan", label="best fit")
plt.plot([G.mu,G.mu],[G.b,G.b+G.amplitude(n)], color="red", label="height")
print "Amplitude over RMS:", G.amplitude(n)/np.std(data-G.histogram(n))
plt.legend(loc="best")
return data
