def process_image(path):
global align
img = cv2.imread(path)
img_n,img_m,_ = img.shape
midline = [8, 27, 28, 29, 51, 57, 62, 66] # All midline points except for two points on the tip of the nose, which are left out
# Rotate until we are within 0.5 dgerees of straight
for i in range(5):
bb = align.getLargestFaceBoundingBox(img)
landmarks = align.findLandmarks(img,bb)
# Find line of symmetry and rotate image accordingly
invslope, invintercept, _, _, _ = orgr.orthoregress([ landmarks[i][1] for i in midline ], [ landmarks[i][0] for i in midline ])
# Rotate
if abs(invslope) < 0.0087: # Approximately 0.5 degrees
# everything is great, no rotation required
break
else:
img = imutils.rotate_bound(img, atan(invslope)*180/pi)
bb = align.getLargestFaceBoundingBox(img)
landmarks = align.findLandmarks(img,bb)
#~ invslope, invintercept, _, _, _ = scipy.stats.linregress([ landmarks[i][1] for i in midline ], [ landmarks[i][0] for i in midline ])
invslope, invintercept, _, _, _ = orgr.orthoregress([ landmarks[i][1] for i in midline ], [ landmarks[i][0] for i in midline ])
assert (abs(atan(invslope)*180/pi) < 0.5)
# Crop the face part
ys = [ p[1] for p in landmarks ]
xs = [ p[0] for p in landmarks ]
miny = min(ys)
maxy = max(ys)
minx = min(xs)
maxx = max(xs)
# Make sure axis of symmetry is in the middle of the image
minx = int(floor(min(minx, 2*invintercept-maxx)))
maxx = int(ceil(max(maxx, 2*invintercept-minx)))
# Do the actual cropping
img = img[miny:maxy, minx:maxx]
# Scale image to width 500
scale_factor = 500./(maxx-minx)
img = cv2.resize(img, None, fx = scale_factor, fy = scale_factor, interpolation = cv2.INTER_CUBIC)
# Re-adjust landmarks
for i in range(len(landmarks)):
x,y = landmarks[i]
landmarks[i] = int(round((x-minx)*scale_factor)), int(round((y-miny)*scale_factor))
if DISPLAY_IMAGE:
# Mark face landmarks on image
for i,point in enumerate(landmarks):
#~ if i not in midline:
#~ continue
cv2.circle(img,point,2,(255,0,0),-1)
# Draw line of symmetry on the face
img_n,img_m,_ = img.shape
cv2.line(img,(img_m/2,0),(img_m/2,img_n),(255,0,0),1)
return landmarks, img
def main():
# This function reformats Bart's file
# hubble constant used throughout
hubbleC = 71
# open the files
if getpass.getuser() == 'frenchd':
filename = '/usr/data/moosejaw/frenchd/GT_update2/returnPA11111111111111.csv'
else:
print 'Could not determine username. Exiting.'
sys.exit()
diameterFile = open(filename, 'rU')
# read in the csv files as dictionary csv files
reader = csv.DictReader(diameterFile)
count = 0
dict = {}
# surveys = [\
# 'r (SDSS Petrosian)',\
# 'r (SDSS Exponential)',\
# 'r (SDSS Isophotal)',\
# 'K_s (2MASS isophotal)',\
# 'K_s (2MASS "total")',\
# 'ESO-LV "Quick Blue" IIa-O',\
# 'POSS1 103a-O',\
# 'r (SDSS deVaucouleurs)',\
# 'r (SDSS de Vaucouleurs)',\
# 'RC3 D_25, R_25 (blue)',\
# 'RC3 D_0 (blue)',\
# 'ESO-Uppsala "Quick Blue" IIa-O',\
# 'ESO-LV IIIa-F',\
# 'RC3 A_e (Johnson B)',\
# 'POSS1 103a-E',\
# 'K_s (LGA/2MASS isophotal)',\
# 'K_s (LGA/2MASS "total")',\
# 'B (Johnson)',\
# 'R (Kron-Cousins)',\
# '408 & 1407 MHz',\
# '1407 MHz']
# survey 2 is always the same, we're comparing survey1 to 'K_s (2MASS "total")' each time
survey2 = 'K_s (2MASS "total")'
# survey2 = 'r (SDSS Petrosian)'
s2List = []
# survey1 = 'ESO-LV "Quick Blue" IIa-O'
# survey1 = 'r (SDSS Petrosian)'
# survey1 = 'r (SDSS Exponential)'
# survey1 = 'r (SDSS Isophotal)'
# survey1 = 'K_s (2MASS isophotal)'
# survey1 = 'K_s (2MASS "total")'
# survey1 = 'POSS1 103a-O'
# survey1 = 'r (SDSS deVaucouleurs)'
# survey1 = 'r (SDSS de Vaucouleurs)'
# survey1 = 'RC3 D_25, R_25 (blue)'
# survey1 = 'RC3 D_0 (blue)'
# survey1 = 'ESO-Uppsala "Quick Blue" IIa-O'
# survey1 = 'ESO-Uppsala "Quick Blue" IIa-O' - none w/ r Petro
# survey1 = 'ESO-LV IIIa-F' - none w/ r Petro
# survey1 = 'ESO-LV IIIa-F'
# survey1 = 'RC3 A_e (Johnson B)'
# survey1 = 'POSS1 103a-E' - none w/ r Petro
# survey1 = 'POSS1 103a-E'
# survey1 = 'K_s (LGA/2MASS isophotal)' - only 1 w/ K total
# survey1 = 'K_s (LGA/2MASS "total")' - only 1 w/ K total
# survey1 = 'B (Johnson)'
survey1 = 'R (Kron-Cousins)'
# survey1 = '408 & 1407 MHz' - none w/ r Petro
# survey1 = '408 & 1407 MHz' - none w/ K total
# survey1 = '1407 MHz' - only 1 w/ r Petro
# survey1 = '1407 MHz' - only 1 w/ K total
s1List = []
# true for major axis comparison, false for minor axis comparison
major = True
dif = 10000
cutoff = 2000000
for line in reader:
count += 1
if count < 455000000:
oldName = line['oldName']
oldCompleteList = eval(line['completeList'])
# switch major and minor if they are mixed up.
completeList = []
if len(str(oldCompleteList)) > 5:
# print 'completeList: ',oldCompleteList
for m in oldCompleteList:
diameters = m[1]
ratio = m[2]
maj, minor = diameters[0], diameters[1]
if isNumber(maj):
if float(maj) < cutoff:
# switch them if the minor axis is larger than the major
if isNumber(maj) and isNumber(minor):
if maj < minor:
minor = float(maj) * float(ratio)
if isNumber(minor) and not isNumber(maj):
maj, minor = minor, maj
completeList.append(
[m[0], (maj, minor), m[2], m[3]])
if len(completeList)>1 and str(completeList).find(survey1)!=-1 and\
str(completeList).find(survey2)!=-1:
# first choose only one of the measurements from each telescope if there
# are multiple available (choose the largest)
innerDict = {}
for m in completeList:
survey = m[0]
diameters = m[1]
maj, minor = diameters[0], diameters[1]
if major:
axis = maj
else:
axis = minor
# check if the survey has been added previously
if isNumber(axis):
if float(axis) > 0:
if innerDict.has_key(survey):
i = innerDict[survey]
prevMaj, prevMin = i[1][0], i[1][1]
if major:
prevAxis = prevMaj
else:
prevAxis = prevMin
if isNumber(prevAxis) and isNumber(axis):
if prevAxis < axis:
innerDict[survey] = m
else:
innerDict[survey] = m
allSurveys = innerDict.keys()
uniqueList = []
for survey in allSurveys:
m = innerDict[survey]
uniqueList.append(m)
# print 'uniqueList: ',uniqueList
len1before = len(s1List)
len2before = len(s2List)
allList = []
for m in uniqueList:
if major:
allList.append(m[1][0])
else:
allList.append(m[1][1])
# print 'allList: ',allList
med = median(allList)
val1 = 'x'
val2 = 'x'
avg = average(allList)
num = len(allList)
for m in uniqueList:
survey = m[0]
diameters = m[1]
ratio = m[2]
pa = m[3]
if survey == survey1 and len(s1List) == len1before:
# if float(diameters[0]) > 53 and float(diameters[0]) <55:
# print 'Oldname: ',oldName
# print completeList
# print
if major:
if isOdd(diameters[0], med, avg, dif, num):
print 'odd: ', m, ' -- ', med
print 'for: ', oldName
print
break
else:
val1 = diameters[0]
else:
if isOdd(diameters[1], med, avg, dif, num):
print 'odd: ', m, ' -- ', med
print 'for: ', oldName
print
break
else:
val1 = diameters[1]
# add this to the list if it has not already been added
if survey == survey2 and len(s2List) == len2before:
if major:
if isOdd(diameters[0], med, avg, dif, num):
print 'odd: ', m, ' -- ', med
print 'for: ', oldName
print
break
else:
val2 = diameters[0]
else:
if isOdd(diameters[1], med, avg, dif, num):
print 'odd: ', m, ' -- ', med
print 'for: ', oldName
print
break
else:
val2 = diameters[1]
if isNumber(val1) and isNumber(val2):
s1List.append(val1)
s2List.append(val2)
print 'len(1): ', len(s1List)
print 'len(2): ', len(s2List)
max1 = max(s1List)
max2 = max(s2List)
maxAll = max(max1, max2)
print 'maxAll: ', maxAll
##########################################################################################
xdata = s1List
ydata = s2List
##########################################################################################
def func(p, x):
b, c = p
return b * x + c
# Run the regression
xerr = np.array(xdata) * 0.05
yerr = np.array(ydata) * 0.05
out = orthoregress.orthoregress(xdata, ydata, xerr, yerr)
m_lr, b_lr, r_value, p_value, std_err = linregress(s1List, s2List)
print
print 'out.beta: ', out.beta
print 'out.sd_beta: ', out.sd_beta
#print fit parameters and 1-sigma estimates
popt = out.beta
perr = out.sd_beta
print
print('fit parameter 1-sigma error')
print
print('---------')
for i in range(len(popt)):
print(str(popt[i]) + ' +- ' + str(perr[i]))
# prepare confidence level curves
# to draw 5-sigma intervals
nstd = 5.
popt_up5 = popt + nstd * perr
popt_dw5 = popt - nstd * perr
# and 1-sigma intervals
popt_up1 = popt + perr
popt_dw1 = popt - perr
x_fit = np.linspace(min(xdata), max(xdata), 100)
fit = func(popt, x_fit)
fit_up5 = func(popt_up5, x_fit)
fit_dw5 = func(popt_dw5, x_fit)
fit_up1 = func(popt_up1, x_fit)
fit_dw1 = func(popt_dw1, x_fit)
# linregress result
fit_lr = func([m_lr, b_lr], x_fit)
##########################################################################################
# round stuff for nice formatting
m_lr2 = round_to_sig(m_lr, sig=3)
b_lr2 = round_to_sig(b_lr, sig=3)
r_value2 = round_to_sig(r_value, sig=3)
p_value2 = round_to_sig(p_value, sig=3)
std_err2 = round_to_sig(std_err, sig=3)
m_odr = round_to_sig(popt[0], sig=3)
b_odr = round_to_sig(popt[1], sig=3)
merr_odr = round_to_sig(perr[0], sig=3)
berr_odr = round_to_sig(perr[1], sig=3)
# plot the fit
# xp = np.linspace(0, max1, 100)
# ax.plot(xp, m2 * xp + b2, '-', label='1D fit: y = {0}*x + {1}'.format(m2,b2))
# ax.plot(xp, (m2 + std_err) * xp + b2, '-',color='red', label='1D fit: y = {0}*x + {1}'.format(m2,b2))
# ax.plot(xp, (m2 - std_err) * xp + b2, '-',color='red', label='1D fit: y = {0}*x + {1}'.format(m2,b2))
fitLabel = '1D fit: y = {0}*x + {1}'.format(m_lr2, b_lr2)
r_valueLabel = 'r_value = {0}'.format(r_value2)
p_valueLabel = 'p_value = {0}'.format(p_value2)
std_errLabel = 'std_err = {0}'.format(std_err2)
#plot
fig = figure(figsize=(7.7, 5.7))
ax = fig.add_subplot(111)
# fig, ax = plt.subplots(1)
fsize = 13
rcParams['font.size'] = fsize
errorbar(xdata,
ydata,
yerr=yerr,
xerr=xerr,
hold=True,
ecolor='k',
fmt='none',
label='data')
xlabel('{0} diameters'.format(survey1), fontsize=fsize)
ylabel('{0} diameters'.format(survey2), fontsize=fsize)
title('fit with error on both axis', fontsize=fsize)
plot(x_fit, fit, 'r', lw=2, label='ODR = {0}*x + {1}'.format(m_odr, b_odr))
plot(x_fit, fit_lr, color='green', lw=2, label='Linregress fit curve')
ax.fill_between(x_fit,
fit_up1,
fit_dw1,
alpha=.45,
label='1-sigma: m+/- {0}, b+/- {1}'.format(
merr_odr, berr_odr))
ax.fill_between(x_fit,
fit_up5,
fit_dw5,
alpha=.20,
label='5-sigma interval')
legend(loc='upper left', fontsize=fsize)
# xlim(0,max(x))
# ylim(0,max(y))
print
print 'm_lr2: ', m_lr2
print 'b_lr2: ', b_lr2
print 'r_value2: ', r_value2
print 'p_value2: ', p_value2
print 'std_err2: ', std_err2
print
ax.set_ylim(0, max2 + max2 * 0.35)
ax.set_xlim(0, max1 + max1 * 0.1)
# show()
# sys.exit()
survey1 = survey1.replace('/', '_')
survey2 = survey2.replace('/', '_')
ans = 'y'
if ans == 'y':
directory = '/usr/data/moosejaw/frenchd/GT_update2/diameterFits_odr/'
# write the figure
fig.savefig('{0}/{1}-{2}_odrreg.pdf'.format(directory, survey1,
survey2),
format='pdf')
# write the information file
fit_result = open(
'{0}/{1}-{2}_odrreg_fits.txt'.format(directory, survey1, survey2),
'wt')
fit_result.write('----Orthogonal Distance Regression results----\n')
fit_result.write('[m,b] = {0}\n'.format(out.beta))
fit_result.write('[m_err,b_err] = {0}\n'.format(out.sd_beta))
fit_result.write('Covariance = {0}\n'.format(out.cov_beta))
fit_result.write('Residual Variance = {0}\n'.format(out.res_var))
fit_result.write('Relative Error = {0}\n'.format(out.rel_error))
fit_result.write('Info = {0}\n'.format(out.info))
fit_result.write('Stop Reason = {0}\n'.format(out.stopreason))
fit_result.write('\n')
fit_result.write('\n')
fit_result.write('----Standard Linear Regression results----\n')
fit_result.write('{0}\n'.format(fitLabel))
fit_result.write('r_value = {0}\n'.format(r_valueLabel))
fit_result.write('p_value = {0}\n'.format(p_valueLabel))
fit_result.write('std_err = {0}\n'.format(std_err))
fit_result.close()
# fig.savefig('/usr/users/frenchd/Galaxy Table code/diameterFix/finalFits/ESO-LV_"Quick Blue"_IIa-O_vs_K_s(LGA_2MASS)_total.pdf',format='pdf')
else:
show()
diameterFile.close()
def main():
# This function reformats Bart's file
# hubble constant used throughout
hubbleC = 71
# open the files
if getpass.getuser() == 'frenchd':
filename = '/usr/data/moosejaw/frenchd/GT_update2/returnPA11111111111111.csv'
else:
print 'Could not determine username. Exiting.'
sys.exit()
diameterFile = open(filename, 'rU')
# read in the csv files as dictionary csv files
reader = csv.DictReader(diameterFile)
count = 0
dict = {}
# surveys = [\
# 'r (SDSS Petrosian)',\
# 'r (SDSS Exponential)',\
# 'r (SDSS Isophotal)',\
# 'K_s (2MASS isophotal)',\
# 'K_s (2MASS "total")',\
# 'ESO-LV "Quick Blue" IIa-O',\
# 'POSS1 103a-O',\
# 'r (SDSS deVaucouleurs)',\
# 'r (SDSS de Vaucouleurs)',\
# 'RC3 D_25, R_25 (blue)',\
# 'RC3 D_0 (blue)',\
# 'ESO-Uppsala "Quick Blue" IIa-O',\
# 'ESO-LV IIIa-F',\
# 'RC3 A_e (Johnson B)',\
# 'POSS1 103a-E',\
# 'K_s (LGA/2MASS isophotal)',\
# 'K_s (LGA/2MASS "total")',\
# 'B (Johnson)',\
# 'R (Kron-Cousins)',\
# '408 & 1407 MHz',\
# '1407 MHz']
# survey 2 is always the same, we're comparing survey1 to 'K_s (2MASS "total")' each time
survey2 = 'K_s (2MASS "total")'
# survey2 = 'r (SDSS Petrosian)'
s2List = []
# survey1 = 'ESO-LV "Quick Blue" IIa-O'
# survey1 = 'r (SDSS Petrosian)'
# survey1 = 'r (SDSS Exponential)'
# survey1 = 'r (SDSS Isophotal)'
# survey1 = 'K_s (2MASS isophotal)'
# survey1 = 'K_s (2MASS "total")'
survey1 = 'ESO-LV "Quick Blue" IIa-O'
# survey1 = 'POSS1 103a-O'
# survey1 = 'r (SDSS deVaucouleurs)'
# survey1 = 'r (SDSS de Vaucouleurs)'
# survey1 = 'RC3 D_25, R_25 (blue)'
# survey1 = 'RC3 D_0 (blue)'
# survey1 = 'ESO-Uppsala "Quick Blue" IIa-O'
# survey1 = 'ESO-Uppsala "Quick Blue" IIa-O' - none w/ r Petro
# survey1 = 'ESO-LV IIIa-F' - none w/ r Petro
# survey1 = 'ESO-LV IIIa-F'
# survey1 = 'RC3 A_e (Johnson B)'
# survey1 = 'POSS1 103a-E' - none w/ r Petro
# survey1 = 'POSS1 103a-E'
# survey1 = 'K_s (LGA/2MASS isophotal)' - only 1 w/ K total
# survey1 = 'K_s (LGA/2MASS "total")' - only 1 w/ K total
# survey1 = 'B (Johnson)'
# survey1 = 'R (Kron-Cousins)'
# survey1 = '408 & 1407 MHz' - none w/ r Petro
# survey1 = '408 & 1407 MHz' - none w/ K total
# survey1 = '1407 MHz' - only 1 w/ r Petro
# survey1 = '1407 MHz' - only 1 w/ K total
s1List = []
# true for major axis comparison, false for minor axis comparison
major = True
dif = 10
cutoff = 2000000
for line in reader:
count += 1
if count < 455000000:
oldName = line['oldName']
oldCompleteList = eval(line['completeList'])
# switch major and minor if they are mixed up.
completeList = []
if len(str(oldCompleteList)) > 5:
# print 'completeList: ',oldCompleteList
for m in oldCompleteList:
diameters = m[1]
ratio = m[2]
maj, minor = diameters[0], diameters[1]
if isNumber(maj):
if float(maj) < cutoff:
# switch them if the minor axis is larger than the major
if isNumber(maj) and isNumber(minor):
if maj < minor:
minor = float(maj) * float(ratio)
if isNumber(minor) and not isNumber(maj):
maj, minor = minor, maj
completeList.append(
[m[0], (maj, minor), m[2], m[3]])
if len(completeList)>1 and str(completeList).find(survey1)!=-1 and\
str(completeList).find(survey2)!=-1:
# first choose only one of the measurements from each telescope if there
# are multiple available (choose the largest)
innerDict = {}
for m in completeList:
survey = m[0]
diameters = m[1]
maj, minor = diameters[0], diameters[1]
if major:
axis = maj
else:
axis = minor
# check if the survey has been added previously
if isNumber(axis):
if float(axis) > 0:
if innerDict.has_key(survey):
i = innerDict[survey]
prevMaj, prevMin = i[1][0], i[1][1]
if major:
prevAxis = prevMaj
else:
prevAxis = prevMin
if isNumber(prevAxis) and isNumber(axis):
if prevAxis < axis:
innerDict[survey] = m
else:
innerDict[survey] = m
allSurveys = innerDict.keys()
uniqueList = []
for survey in allSurveys:
m = innerDict[survey]
uniqueList.append(m)
# print 'uniqueList: ',uniqueList
len1before = len(s1List)
len2before = len(s2List)
allList = []
for m in uniqueList:
if major:
allList.append(m[1][0])
else:
allList.append(m[1][1])
# print 'allList: ',allList
med = median(allList)
val1 = 'x'
val2 = 'x'
avg = average(allList)
num = len(allList)
for m in uniqueList:
survey = m[0]
diameters = m[1]
ratio = m[2]
pa = m[3]
if survey == survey1 and len(s1List) == len1before:
# if float(diameters[0]) > 53 and float(diameters[0]) <55:
# print 'Oldname: ',oldName
# print completeList
# print
if major:
if isOdd(diameters[0], med, avg, dif, num):
print 'odd: ', m, ' -- ', med
print 'for: ', oldName
print
break
else:
val1 = diameters[0]
else:
if isOdd(diameters[1], med, avg, dif, num):
print 'odd: ', m, ' -- ', med
print 'for: ', oldName
print
break
else:
val1 = diameters[1]
# add this to the list if it has not already been added
if survey == survey2 and len(s2List) == len2before:
if major:
if isOdd(diameters[0], med, avg, dif, num):
print 'odd: ', m, ' -- ', med
print 'for: ', oldName
print
break
else:
val2 = diameters[0]
else:
if isOdd(diameters[1], med, avg, dif, num):
print 'odd: ', m, ' -- ', med
print 'for: ', oldName
print
break
else:
val2 = diameters[1]
if isNumber(val1) and isNumber(val2):
s1List.append(val1)
s2List.append(val2)
print 'len(1): ', len(s1List)
print 'len(2): ', len(s2List)
max1 = max(s1List)
max2 = max(s2List)
maxAll = max(max1, max2)
print 'maxAll: ', maxAll
fig = figure()
ax = fig.add_subplot(111)
# do the fit
# f = np.polyfit(s1List, s2List,1,full=True)
# fit, residuals, rank, singular_values, rcond = f
# print
# print 'fit: ',fit
# print 'residuals: ',residuals
# print 'rank: ',rank
# print 'singular_values: ',singular_values
# print 'rcond: ',rcond
# print
# z = np.poly1d(fit)
# xp = np.linspace(0, max1, 100)
# ax.plot(xp,z(xp),'-',label='1D fit: y = {0}*x + {1}'.format(str(round(fit[0],3)),str(round(fit[1],3))))
##########################################################################################
# m, b, r_value, p_value, std_err = linregress(s1List, s2List)
xdata = s1List
ydata = s2List
# pfit, pcov = optimize.curve_fit(lineFunc, xdata, ydata)
##########################################################################################
# These are initial guesses for fits:
pstart = [1.0, 0.0]
# bootstrap method
# def ff(x, p):
# return lineFunc(x, *p)
# These are the true parameters
# p0 = 1.0
# p1 = 40
# p2 = 2.0
# These are initial guesses for fits:
# pstart = [
# p0 + random.random(),
# p1 + 5.*random.random(),
# p2 + random.random()
# ]
# %matplotlib inline
# import matplotlib.pyplot as plt
# xvals = np.linspace(0., 1, 120)
# yvals = f(xvals, p0, p1, p2)
# Generate data with a bit of randomness
# xdata = np.array(xvals)
# np.random.seed(42)
# err_stdev = 0.2
# yvals_err = np.random.normal(0., err_stdev, len(xdata))
# ydata = f(xdata, p0, p1, p2) + yvals_err
# plt.plot(xvals, yvals)
# plt.plot(xdata, ydata, 'o', mfc='None')
# pfit, perr = fit_bootstrap(pstart, xdata, ydata, ff)
#
# print 'pfit: ',pfit
# print
# print 'perr: ',perr
##########################################################################################
# pfit, perr = fit_curvefit(pstart, xdata, ydata, ff)
#
# pfit, pcov = optimize.curve_fit(lineFunc,xdata,ydata,p0=pstart)
# error = []
# for i in range(len(pfit)):
# try:
# error.append(np.absolute(pcov[i][i])**0.5)
# except:
# error.append( 0.00 )
# pfit_curvefit = pfit
# perr_curvefit = np.array(error)
# print "curve_fit: ", pfit_curvefit, perr_curvefit
#
# print("\nFit parameters and parameter errors from curve_fit method :")
# print("pfit = ", pfit)
# print("perr = ", perr)
# error = []
# for i in range(len(pfit)):
# try:
# error.append(absolute(pcov[i][i])**0.5)
# except:
# error.append( 0.00 )
# pfit_curvefit = pfit
# perr_curvefit = array(error)
# print("\nFit parameters and parameter errors from curve_fit method :")
# print("pfit = ", pfit_curvefit)
# print("perr = ", perr_curvefit)
# pfit_curvefit = pfit
# perr_curvefit = array(perr)
#
# m = pfit_curvefit[0]
# b = pfit_curvefit[1]
#
# std_err = perr_curvefit[0]
# r_value = 0
# p_value = 0
##########################################################################################
# def func(p, x):
# a, b, c = p
# return a*x*x + b*x + c
def func(p, x):
b, c = p
return b * x + c
# Model object
quad_model = odr.Model(func)
x = np.array(xdata)
y = np.array(ydata)
noise_x = x * 0.05
noise_y = y * 0.05
# Create a RealData object
data = odr.RealData(x, y, sx=noise_x, sy=noise_y)
# data = odr.Data(x, y)
# Set up ODR with the model and data
# odr2 = odr.ODR(data, quad_model, beta0=[0., 1., 1.])
odr2 = odr.ODR(data, quad_model, beta0=[1., 1.])
# Run the regression
out = odr2.run()
#print fit parameters and 1-sigma estimates
popt = out.beta
perr = out.sd_beta
print
print('fit parameter 1-sigma error')
print
print('---------')
for i in range(len(popt)):
print(str(popt[i]) + ' +- ' + str(perr[i]))
# prepare confidence level curves
# to draw 5-sigma intervals
nstd = 5.
popt_up = popt + nstd * perr
popt_dw = popt - nstd * perr
x_fit = np.linspace(min(x), max(x), 100)
fit = func(popt, x_fit)
fit_up = func(popt_up, x_fit)
fit_dw = func(popt_dw, x_fit)
#plot
fig, ax = plt.subplots(1)
# rcParams['font.size']= 15
# errorbar(x, y, yerr=noise_y, xerr=noise_x, hold=True, ecolor='k', fmt='none', label='data')
xlabel('{0} diameters'.format(survey1), fontsize=15)
ylabel('{0} diameters'.format(survey2), fontsize=15)
title('fit with error on both axis', fontsize=15)
plot(x_fit, fit, 'r', lw=2, label='best fit curve')
# plot(x0, y0, 'k-', lw=2, label='True curve')
# ax.fill_between(x_fit, fit_up, fit_dw, alpha=.25, label='5-sigma interval')
legend(loc='lower right', fontsize=15)
xlim(0, max(x))
ylim(0, max(y))
show()
##########################################################################################
# try orthoregress modules
import orthoregress
xerr = np.array(xdata) * 0.05
yerr = np.array(ydata) * 0.05
out = orthoregress.orthoregress(xdata, ydata, xerr, yerr)
print
print 'out.beta: ', out.beta
print 'out.sd_beta: ', out.sd_beta
print 'out.cov_beta: ', out.cov_beta
print 'out.delta: ', out.delta
print 'out.eps: ', out.eps
print 'out.res_var: ', out.res_var
print 'out.sum_sqare: ', out.sum_sqare
print 'out.sum_square_delta: ', out.sum_square_delta
print 'out.sum_square_eps: ', out.sum_square_eps
print
sys.exit()
# round stuff for nice formatting
m2 = round_to_sig(m, sig=3)
b2 = round_to_sig(b, sig=3)
r_value2 = round_to_sig(r_value, sig=3)
p_value2 = round_to_sig(p_value, sig=3)
std_err2 = round_to_sig(std_err, sig=3)
# plot the fit
xp = np.linspace(0, max1, 100)
ax.plot(xp,
m2 * xp + b2,
'-',
label='1D fit: y = {0}*x + {1}'.format(m2, b2))
ax.plot(xp, (m2 + std_err) * xp + b2,
'-',
color='red',
label='1D fit: y = {0}*x + {1}'.format(m2, b2))
ax.plot(xp, (m2 - std_err) * xp + b2,
'-',
color='red',
label='1D fit: y = {0}*x + {1}'.format(m2, b2))
# plot the data
ax.plot(s1List, s2List, '.', label='Data with dif={0}'.format(dif))
print
print 'm2: ', m2
print 'b2: ', b2
print 'r_value2: ', r_value2
print 'p_value2: ', p_value2
print 'std_err2: ', std_err2
print
xLabelPos = 0.6
yLabelPos = 0.7
yShift = 0.05
lsize = 12
fitLabel = '1D fit: y = {0}*x + {1}'.format(m2, b2)
r_valueLabel = 'r_value = {0}'.format(r_value2)
p_valueLabel = 'p_value = {0}'.format(p_value2)
std_errLabel = 'std_err = {0}'.format(std_err2)
ax.annotate(fitLabel,
xy=(xLabelPos, yLabelPos),
xycoords='axes fraction',
size=lsize)
ax.annotate(r_valueLabel,
xy=(xLabelPos, yLabelPos - yShift),
xycoords='axes fraction',
size=lsize)
ax.annotate(p_valueLabel,
xy=(xLabelPos, yLabelPos - yShift - yShift),
xycoords='axes fraction',
size=lsize)
ax.annotate(std_errLabel,
xy=(xLabelPos, yLabelPos - yShift - yShift - yShift),
xycoords='axes fraction',
size=lsize)
# legend()
ax.set_xlabel('{0} diameters'.format(survey1))
ax.set_ylabel('{0} diameters'.format(survey2))
ax.set_ylim(0, max2 + max2 * 0.3)
ax.set_xlim(0, max1 + max1 * 0.3)
# ans = raw_input("Save (y/n)? ")
# while ans != 'y' and ans != 'n':
# ans = raw_input("Save (y/n)? ")
survey1 = survey1.replace('/', '_')
survey2 = survey2.replace('/', '_')
ans = 'y'
if ans == 'y':
fig.savefig(
'/usr/data/moosejaw/frenchd/GT_update2/diameterFits/{0}-{1}_dif{2}_linreg3.pdf'
.format(survey1, survey2, dif),
format='pdf')
# fig.savefig('/usr/users/frenchd/Galaxy Table code/diameterFix/finalFits/ESO-LV_"Quick Blue"_IIa-O_vs_K_s(LGA_2MASS)_total.pdf',format='pdf')
else:
show()
diameterFile.close()