def chi_sq(trial_params, covinv, coords):
"""Compute chi^2 given a set of trial parameters,
the inverse covariance matrix, and the x-y coordinates to fit.
"""
return fsum([(coords[outer][1]-
fit_func(coords[outer][0], trial_params))*
covinv[outer][inner]*(coords[inner][1]-
fit_func(coords[inner][0],
trial_params))
for outer in range(len(coords))
for inner in range(len(coords))])
def mkplot(coords, cov, result_min):
"""Plot the fitted graph."""
with PdfPages("foo.pdf") as pdf:
XCOORD = [coords[i][0] for i in range(len(coords))]
YCOORD = [coords[i][1] for i in range(len(coords))]
ER2 = np.array([cov[i][i] for i in range(len(coords))])
plt.errorbar(XCOORD, YCOORD, yerr=ER2, linestyle="None")
# the fit function is plotted on a scale 1000x more fine
# than the original data points
XFIT = np.arange(
XCOORD[0], XCOORD[len(XCOORD) - 1], abs((XCOORD[len(XCOORD) - 1] - XCOORD[0])) / 1000.0 / len(XCOORD)
)
# result_min.x is is the array of minimized fit params
YFIT = np.array([fit_func(XFIT[i], result_min.x) for i in range(len(XFIT))])
# only plot fit function if minimizer result makes sense
if result_min.status == 0:
print "Minimizer thinks that it worked. Plotting fit."
plt.plot(XFIT, YFIT)
# todo: figure out a way to generally assign limits to plot
# plt.xlim([XCOORD[0], XMAX+1])
# magic numbers for the problem you're solving
# plt.ylim([0, 0.1])
# add labels, more magic numbers
plt.title("Some Correlation function vs. q^2")
# todo: figure out a way to generally place text on plot
# STRIKE1 = "Energy = " + str(result_min.x[1]) + "+/-" + str(
# ERR_ENERGY)
# STRIKE2 = "Amplitude = " + str(result_min.x[0]) + "+/-" + str(
# ERR_A0)
# X_POS_OF_FIT_RESULTS = XCOORD[3]
# plt.text(X_POS_OF_FIT_RESULTS, YCOORD[3], STRIKE1)
# plt.text(X_POS_OF_FIT_RESULTS, YCOORD[7], STRIKE2)
plt.xlabel("q^2 (GeV/c^2)^2")
plt.ylabel("the function")
# read out into a pdf
pdf.savefig()
# show the plot
plt.show()
return 0