def run(self, npts=25, inv_points=None, access_limited=True, **kwargs):
r"""
Parameters
----------
npts : int (default = 25)
The number of pressure points to apply. The list of pressures
is logarithmically spaced between the lowest and highest throat
entry pressures in the network.
inv_points : array_like, optional
A list of specific pressure point(s) to apply.
"""
if 'inlets' in kwargs.keys():
logger.info('Inlets recieved, passing to set_inlets')
self.set_inlets(pores=kwargs['inlets'])
if 'outlets' in kwargs.keys():
logger.info('Outlets recieved, passing to set_outlets')
self.set_outlets(pores=kwargs['outlets'])
self._AL = access_limited
if inv_points is None:
logger.info('Generating list of invasion pressures')
min_p = sp.amin(self['throat.entry_pressure']) * 0.98 # nudge down
max_p = sp.amax(self['throat.entry_pressure']) * 1.02 # bump up
inv_points = sp.logspace(sp.log10(min_p),
sp.log10(max_p),
npts)
self._npts = sp.size(inv_points)
# Execute calculation
self._do_outer_iteration_stage(inv_points)
def plotHeatmap(fwrap, aclass, algoparams, trials, maxsteps):
""" Visualizing performance across trials and across time
(iterations in powers of 2) """
psteps = int(log2(maxsteps)) + 1
storesteps = [0] + [2 ** x for x in range(psteps)]
ls = lossTraces(fwrap, aclass, dim=trials, maxsteps=maxsteps,
storesteps=storesteps, algoparams=algoparams,
minLoss=1e-10)
initv = mean(ls[0])
maxgain = exp(fwrap.stochfun.maxLogGain(maxsteps) + 1)
maxneggain = (sqrt(maxgain))
M = zeros((psteps, trials))
for sid in range(psteps):
# skip the initial values
winfactors = clip(initv / ls[sid+1], 1. / maxneggain, maxgain)
winfactors[isnan(winfactors)] = 1. / maxneggain
M[sid, :] = log10(sorted(winfactors))
pylab.imshow(M.T, interpolation='nearest', cmap=cm.RdBu, #@UndefinedVariable
aspect=psteps / float(trials) / 1,
vmin= -log10(maxgain), vmax=log10(maxgain),
)
pylab.xticks([])
pylab.yticks([])
return ls
def getAxis(self,X,Y):
"""
return the proper axis limits for the plots
"""
out = []
mM = [(min(X),max(X)),(min(Y),max(Y))]
for i,j in mM:
#YJC: checking if values are negative, if yes, return 0 and break
if j <0 or i <0:
return 0
log_i = scipy.log10(i)
d, I = scipy.modf(log_i)
if log_i < 0:
add = 0.5 *(scipy.absolute(d)<0.5)
else:
add = 0.5 *(scipy.absolute(d)>0.5)
m = scipy.floor(log_i) + add
out.append(10**m)
log_j = scipy.log10(j)
d, I = scipy.modf(log_j)
if log_j < 0:
add = - 0.5 *(scipy.absolute(d)>0.5)
else:
add = - 0.5 *(scipy.absolute(d)<0.5)
m = scipy.ceil(log_j) + add
out.append(10**m)
return tuple(out)
def powerlaw_fit(xdata, ydata, yerr):
# Power-law fitting is best done by first converting
# to a linear equation and then fitting to a straight line.
# y = a * x^b
# log(y) = log(a) + b*log(x)
from scipy import log10
from scipy import optimize
powerlaw = lambda x, amp, index: amp*np.power(x,index)
logx = log10(xdata)
logy = log10(ydata)
logyerr = yerr / ydata
# define our (line) fitting function
fitfunc = lambda p, x: p[0] + p[1] * x
errfunc = lambda p, x, y, err: (y - fitfunc(p, x)) / err
pinit = [1.0, -1.0]
out = optimize.leastsq(errfunc, pinit, args=(logx, logy, logyerr), full_output=1)
pfinal = out[0]
covar = out[1]
#y = amp * x^exponent
exponent = pfinal[1] #index in original
amp = 10.0**pfinal[0]
exponentErr = np.sqrt( covar[0][0] )
ampErr = np.sqrt( covar[1][1] ) * amp
chisq = np.sum((((ydata - powerlaw(xdata, amp, exponent))/yerr)**2),axis=0)
return exponent, amp, chisq
def fit(self, kk=None):
"""
Fit Fourier spectrum with the function set at class instantination
==> NB: fitting is done in logarithmic coordinates
and fills plotting arrays with data
--------
Options:
--------
kk
(k1,k2) <None> spectral interval for function fitting
by default interval [ kk[1], kk[imax__kk] ] will be fitted
==> i.e. k=0 is excluded
"""
# fitting interval
if kk:
ik_min=(self.fft_data.kk[1:self.fft_data.imax__kk]<=kk[0]).nonzero()[0][-1]
ik_max=(self.fft_data.kk[1:self.fft_data.imax__kk]<=kk[1]).nonzero()[0][-1]
else:
ik_min=1;
ik_max=self.fft_data.imax__kk
# do fitting
self.__popt,self.__pcov = scipy.optimize.curve_fit(self.__func_fit,
scipy.log(self.fft_data.kk[ik_min:ik_max]),
scipy.log(self.fft_data.Ik[ik_min:ik_max]) )
# boundaries of fitted interval
self.kmin = self.fft_data.kk[ik_min]
self.kmax = self.fft_data.kk[ik_max]
# fill plot arrays <===============
self.kk_plot=scipy.logspace( scipy.log10(self.kmin),
scipy.log10(self.kmax),
self.nk_plot )
self.Ik_plot=self.fitting_function(self.kk_plot)
def test_permutation(self):
#test permutation function
for dn in self.datasets:
D = data.load(os.path.join(self.dir_name,dn))
perm = SP.random.permutation(D['X'].shape[0])
#1. set permuattion
lmm = dlimix.CLMM()
lmm.setK(D['K'])
lmm.setSNPs(D['X'])
lmm.setCovs(D['Cov'])
lmm.setPheno(D['Y'])
if 1:
#pdb.set_trace()
perm = SP.array(perm,dtype='int32')#Windows needs int32 as long -> fix interface to accept int64 types
lmm.setPermutation(perm)
lmm.process()
pv_perm1 = lmm.getPv().ravel()
#2. do by hand
lmm = dlimix.CLMM()
lmm.setK(D['K'])
lmm.setSNPs(D['X'][perm])
lmm.setCovs(D['Cov'])
lmm.setPheno(D['Y'])
lmm.process()
pv_perm2 = lmm.getPv().ravel()
D2 = (SP.log10(pv_perm1)-SP.log10(pv_perm2))**2
RV = SP.sqrt(D2.mean())
self.assertTrue(RV<1E-6)
def entropyloss(act, pred):
epsilon = 1e-15
pred = sp.maximum(epsilon, pred)
pred = sp.minimum(1-epsilon, pred)
el = sum(act*sp.log10(pred) + sp.subtract(1,act)*sp.log10(sp.subtract(1,pred)))
el = el * -1.0/len(act)
return el
def addqqplotinfo(qnull,M,xl='-log10(P) observed',yl='-log10(P) expected',xlim=None,ylim=None,alphalevel=0.05,legendlist=None,fixaxes=False):
distr='log10'
pl.plot([0,qnull.max()], [0,qnull.max()],'k')
pl.ylabel(xl)
pl.xlabel(yl)
if xlim is not None:
pl.xlim(xlim)
if ylim is not None:
pl.ylim(ylim)
if alphalevel is not None:
if distr == 'log10':
betaUp, betaDown, theoreticalPvals = _qqplot_bar(M=M,alphalevel=alphalevel,distr=distr)
lower = -sp.log10(theoreticalPvals-betaDown)
upper = -sp.log10(theoreticalPvals+betaUp)
pl.fill_between(-sp.log10(theoreticalPvals),lower,upper,color="grey",alpha=0.5)
#pl.plot(-sp.log10(theoreticalPvals),lower,'g-.')
#pl.plot(-sp.log10(theoreticalPvals),upper,'g-.')
if legendlist is not None:
leg = pl.legend(legendlist, loc=4, numpoints=1)
# set the markersize for the legend
for lo in leg.legendHandles:
lo.set_markersize(10)
if fixaxes:
fix_axes()
def testPlot():
""" Get/generate the data to play with """
TIME_INC = 1e-6
NUM_POINTS = 40000
t = timeScale(TIME_INC,NUM_POINTS)
noisy_sig = genData(TIME_INC,True, t)
clean_sig = genData(TIME_INC,False,t)
""" Get FFT of signal and the sampling frequency from the time intervals used to generate the signals"""
freq, s_fft = getFFT(noisy_sig, TIME_INC)
freq2,s_fft2 = getFFT(clean_sig, TIME_INC)
""" Show in 2 subplots the signals and their spectrums"""
plb.subplot(211,axisbg='#FFFFCC')
p.plot(t,clean_sig,'b')
p.hold(True)
p.grid(True)
p.plot(t,noisy_sig,'r')
plb.subplot(212,axisbg='#FFFFCC')
#p.hold(False)
p.plot(freq2, 20*sp.log10(s_fft2),'x-b')
p.hold(True)
p.plot(freq, 20*sp.log10(s_fft), '+-r')
p.xticks([-10e4,-5e4,-4e4,-3e4,-2e4,-1e4,0,1e4,2e4,3e4,4e4,5e4,10e4])
p.xlim([-1e5,1e5])
p.grid(True)
#p.show()
q = ScrollingToolQT(p.gcf())
return q # WARNING: it's important to return this object otherwise
def create_grid(r_in, r_out, nshell, space = 'powerlaw1', end = True):
# function to create grid
if space == 'log10':
from scipy import log10, logspace
# get the exponent of the start- and
# stop-radius in input units
start = [log10(r_in), 0][r_in == 0]
stop = log10(r_out)
radii = logspace(start, stop, num=nshell, endpoint=end)
elif space == "powerlaw1":
from scipy import arange
radii = r_in * (r_out/r_in)**(arange(nshell)/(nshell - 1.0))
elif space == 'linear':
from scipy import linspace
# linearly spaced grid
radii = linspace(r_in, r_out, num=nshell, endpoint=end)
elif space == 'powerlaw2':
from scipy import linspace
# first check if coefficients to the power-law was given
#~ if 'exp' in kwargs:
#~ p_exp = kwargs['exp']
#~ else: # if not, set it to 2, i.e. r^2
#~ p_exp = 2
radii = r_in + (r_out - r_in)*(linspace(r_in, r_out, num=nshell, endpoint=end)/(r_out))**2
#pr_int('Not implemented yet.')
#raise ParError(spaced)
else:
raise Exception(space)
return radii
def testAll(tests, allalgos, tolerant=True):
countgood = 0
for i, algo in enumerate(sorted(allalgos)):
print("%d, %s:" % (i + 1, algo.__name__))
print(' ' * int(log10(i + 1) + 2),)
good = True
messages = []
for t in tests:
try:
res = t(algo)
except Exception, e:
if not tolerant:
raise e
res = e
if res is True:
print('.',)
else:
good = False
messages.append(res)
print('F',)
if good:
countgood += 1
print('--- OK.')
else:
print('--- NOT OK.')
for m in messages:
if m is not None:
print(' ' * int(log10(i + 1) + 2), '->', m)
def residual_lmfit(self, pars, x, y):
a = P4Rm()
self.strain_DW(pars)
res = f_Refl_fit(a.AllDataDict["geometry"], self.Data4f_Refl)
y_cal = convolve(abs(res) ** 2, a.ParamDict["resol"], mode="same")
y_cal = y_cal / y_cal.max() + a.AllDataDict["background"]
return log10(y) - log10(y_cal)
def r_ion_neutral(s,t,Ni,Nn,Ti,Tn):
""" This will calculate resonant ion - neutral reactions collision frequencies. See
table 4.5 in Schunk and Nagy.
Inputs
s - Ion name string
t - neutral name string
Ni - Ion density cm^-3
Nn - Neutral density cm^-3
Ti - Ion tempreture K
Tn - Neutral tempreture K
Outputs
nu_ineu - collision frequency s^-1
"""
Tr = (Ti+Tn)*0.5
sp1 = (s,t)
# from Schunk and Nagy table 4.5
nudict={('H+','H'):[2.65e-10,0.083],('He+','He'):[8.73e-11,0.093], ('N+','N'):[3.84e-11,0.063],
('O+','O'):[3.67e-11,0.064], ('N2+','N'):[5.14e-11,0.069], ('O2+','O2'):[2.59e-11,0.073],
('H+','O'):[6.61e-11,0.047],('O+','H'):[4.63e-12,0.],('CO+','CO'):[3.42e-11,0.085],
('CO2+','CO'):[2.85e-11,0.083]}
A = nudict[sp1][0]
B = nudict[sp1][1]
if sp1==('O+','H'):
nu_ineu = A*Nn*sp.power(Ti/16.+Tn,.5)
elif sp1==('H+','O'):
nu_ineu = A*Nn*sp.power(Ti,.5)*(1-B*sp.log10(Ti))**2
else:
nu_ineu = A*Nn*sp.power(Tr,.5)*(1-B*sp.log10(Tr))**2
return nu_ineu
def plot_median_errors(RefinementLevels):
for i in RefinementLevels[0].cases:
x =[];
y =[];
print "Analyzing median error on: ", i ;
for r in RefinementLevels:
x.append(r.LUT.D_dim*r.LUT.P_dim)
r.get_REL_ERR_SU2(i)
y.append(r.SU2[i].median_ERR*100)
x = sp.array(x)
y = sp.array(y)
y = y[sp.argsort(x)]
x = x[sp.argsort(x)]
LHM = sp.ones((len(x),2))
RHS = sp.ones((len(x),1))
LHM[:,1] = sp.log10(x)
RHS[:,0] = sp.log10(y)
sols = sp.linalg.lstsq(LHM,RHS)
b = -sols[0][1]
plt.loglog(x,y, label='%s, %s'%(i,r'$O(\frac{1}{N})^{%s}$'%str(sp.around(b,2))), basex=10, basey=10, \
subsy=sp.linspace(10**(-5), 10**(-2),20),\
subsx=sp.linspace(10**(2), 10**(5),50))
#for r in RefinementLevels:
# x.append(r.LUT.D_dim*r.LUT.P_dim)
# r.get_REL_ERR_SciPy(i)
# y.append(r.SciPy[i].median_ERR*100)
#plt.plot(x,y, label='SciPy: %s'%i)
plt.grid(which='both')
plt.xlabel('Grid Nodes (N)')
plt.ylabel('Median relative error [%]')
return;
def getLogBins(first_point, last_point, log_step):
"""
get the bin in log scale and the center bin value
Parameters:
----------------
first_point, last_point : number
First and last point of the x-axis
log_step : number
Required log-distance between x-points
Returns:
-----------
xbins : array of the x values at the center (in log-scale) of the bin
bins : array of the x values of the bins
"""
log_first_point = scipy.log10(first_point)
log_last_point = scipy.log10(last_point)
# Calculate the bins as required by the histogram function, i.e. the bins edges including the rightmost one
N_log_steps = scipy.floor((log_last_point - log_first_point) / log_step) + 1.0
llp = N_log_steps * log_step + log_first_point
bins_in_log_scale = np.linspace(log_first_point, llp, N_log_steps + 1)
bins = 10 ** bins_in_log_scale
center_of_bins_log_scale = bins_in_log_scale[:-1] + log_step / 2.0
xbins = 10 ** center_of_bins_log_scale
return xbins, bins
def testAll(tests, allalgos, tolerant=True):
countgood = 0
for i, algo in enumerate(sorted(allalgos)):
print(("%d, %s:" % (i + 1, algo.__name__)))
print((" " * int(log10(i + 1) + 2),))
good = True
messages = []
for t in tests:
try:
res = t(algo)
except Exception as e:
if not tolerant:
raise e
res = e
if res is True:
print((".",))
else:
good = False
messages.append(res)
print(("F",))
if good:
countgood += 1
print("--- OK.")
else:
print("--- NOT OK.")
for m in messages:
if m is not None:
print((" " * int(log10(i + 1) + 2), "->", m))
print()
print(("Summary:", countgood, "/", len(allalgos), "of test were passed."))
def testAll(tests, allalgos, tolerant=True):
countgood = 0
for i, algo in enumerate(sorted(allalgos)):
print "%d, %s:" % (i + 1, algo.__name__)
print " " * int(log10(i + 1) + 2),
good = True
messages = []
for t in tests:
try:
res = t(algo)
except Exception, e:
if not tolerant:
raise e
res = e
if res is True:
print ".",
else:
good = False
messages.append(res)
print "F",
if good:
countgood += 1
print "--- OK."
else:
print "--- NOT OK."
for m in messages:
if m is not None:
print " " * int(log10(i + 1) + 2), "->", m
def degree_distrib(net, deg_type="total", node_list=None, use_weights=True,
log=False, num_bins=30):
'''
Computing the degree distribution of a network.
Parameters
----------
net : :class:`~nngt.Graph` or subclass
the network to analyze.
deg_type : string, optional (default: "total")
type of degree to consider ("in", "out", or "total").
node_list : list or numpy.array of ints, optional (default: None)
Restrict the distribution to a set of nodes (default: all nodes).
use_weights : bool, optional (default: True)
use weighted degrees (do not take the sign into account: all weights
are positive).
log : bool, optional (default: False)
use log-spaced bins.
Returns
-------
counts : :class:`numpy.array`
number of nodes in each bin
deg : :class:`numpy.array`
bins
'''
ia_node_deg = net.get_degrees(node_list, deg_type, use_weights)
ra_bins = sp.linspace(ia_node_deg.min(), ia_node_deg.max(), num_bins)
if log:
ra_bins = sp.logspace(sp.log10(sp.maximum(ia_node_deg.min(),1)),
sp.log10(ia_node_deg.max()), num_bins)
counts,deg = sp.histogram(ia_node_deg, ra_bins)
ia_indices = sp.argwhere(counts)
return counts[ia_indices], deg[ia_indices]
def PlotManhattan(self,
xname=str,
yname=str,
Log=Logger):
LogString = '**** Generating Manhattan plot ...'
print LogString
Log.Write(LogString+'\n')
XName = ''
YName = ''
for Key in self.DataContainers.iterkeys():
if(re.search(xname,Key)):
XName = Key
if(re.search(yname,Key)):
YName = Key
X = []
for Entry in self.DataContainers[XName].GetDataArray():
X.append(int(Entry))
X = scipy.array(X)
Y = []
for Entry in self.DataContainers[YName].GetDataArray():
if(Entry=='-1'):
Y.append(float(1.0))
else:
Y.append(float(Entry))
Y = -scipy.log10(scipy.array(Y))
PylabParameters,\
Rectangle = self.PylabUpdateParams()
pylab.rcParams.update(PylabParameters)
PylabFigure = pylab.figure()
PylabFigure.clf()
PylabAxis = PylabFigure.add_axes(Rectangle)
PylabAxis.scatter(x=X,
y=Y)
XSign = scipy.array(PylabAxis.get_xlim())
YSign = -scipy.log10(scipy.array([5.0e-8,5.0e-8]))
PylabAxis.plot(XSign,
YSign,
linestyle='--',
color='grey',
linewidth=1.0)
XSugg = scipy.array(PylabAxis.get_xlim())
YSugg = -scipy.log10(scipy.array([1.0e-5,1.0e-5]))
PylabAxis.plot(XSugg,
YSugg,
linestyle=':',
color='grey',
linewidth=1.0)
PylabAxis.set_ylim([0.0,PylabAxis.get_ylim()[1]])
PylabFigure.savefig('Manhattan.png')
PylabAxis.clear()
pylab.close(PylabFigure)
del PylabFigure
del PylabAxis
return
def effective_order(x, y):
'''Find slope of log log plot to find our effective order of accuracy'''
logx = log10(x)
logy = log10(y)
out = curve_fit(linear_fit, logx, logy)
return out[0][1]
def makeGrid(minval,maxval,gridpoints=0,log=0,floatornot=0):
"""
Make grid between max and min value.
@param minval: lower boundary of grid range
@type minval: float
@param maxval: upper boundary of grid range
@type maxval: float
@keyword gridpoints: number of grid points, including boundaries. If 0 it
is replaced by 2, if 1 then minval is gridpoint
(default: 0)
@type gridpoints: int
@keyword log: if grid should be calculated in logspace
(default: 0)
@type log: bool
@keyword floatornot: 0 if final gridpoints should be rounded to nearest
integer
(default: 0)
@type floatornot: bool
@return: the grid points including the boundaries
@rtype: array
"""
if gridpoints == 0:
gridpoints = 2
else:
gridpoints = int(gridpoints)
final_grid = []
if log:
minval = float(log10(minval))
maxval = float(log10(maxval))
grid = linspace(minval,maxval,gridpoints)
if floatornot:
for i in xrange(len(grid)):
final_grid.append(int(round(10**(grid[i]))))
else:
for i in xrange(len(grid)):
final_grid.append(10**(grid[i]))
else:
minval = float(minval)
maxval = float(maxval)
grid = linspace(minval,maxval,gridpoints)
if floatornot:
for i in xrange(len(grid)):
final_grid.append(int(round(grid[i])))
else:
for i in xrange(len(grid)):
final_grid.append(grid[i])
return final_grid
def svpice( t) : ''' Returns saturation vapor pressure over ice, in hPa, given temperature in K. The Goff-Gratch equation (Smithsonian Met. Tables, 5th ed., pp. 350, 1984) ''' a = 273.16 / t exponent = -9.09718 * (a - 1.) - 3.56654 * log10(a) + 0.876793 * (1. - 1./a) + log10(6.1071) return 10.0**exponent
def D2_setup_scan(self,min1,max1,step1,min2,max2,step2):
self.P1min = min1
self.P1max = max1
self.P1steps = step1
self.P2min = min2
self.P2max = max2
self.P2steps = step2
self.P1range = scipy.logspace(scipy.log10(min1),scipy.log10(max1),step1)
self.P2range = scipy.logspace(scipy.log10(min2),scipy.log10(max2),step2)
def powerlaw(y): s_y = np.sort(y)[::-1] X = sp.log10(np.arange(1,len(s_y)+1)) Y = sp.log10(s_y) pinit = [10.0, 0.0] out = sp.optimize.leastsq(errfunc, pinit, args=(X, Y), full_output = 1) index = out[0][1] amp = 10**out[0][0] return index
def plot_overlap_ps(result_file, ss_file='/Users/bjarnivilhjalmsson/data/GIANT/GIANT_HEIGHT_Wood_et_al_2014_publicrelease_HapMapCeuFreq.txt',
fig_filename='/Users/bjarnivilhjalmsson/data/tmp/manhattan_combPC_HGT.png', method='combPC',
ylabel='Comb. PC (HIP,WC,HGT,BMI) $-log_{10}(P$-value$)$', xlabel='Height $-log_{10}(P$-value$)$', p_thres=0.00001):
# Parse results ans SS file
res_table = pandas.read_table(result_file)
ss_table = pandas.read_table(ss_file)
# Parse
res_sids = sp.array(res_table['SNPid'])
if method == 'MVT':
comb_ps = sp.array(res_table['pval'])
elif method == 'combPC':
comb_ps = sp.array(res_table['combPC'])
if 'MarkerName' in ss_table.keys():
ss_sids = sp.array(ss_table['MarkerName'])
elif 'SNP' in ss_table.keys():
ss_sids = sp.array(ss_table['SNP'])
else:
raise Exception("Don't know where to look for rs IDs")
marg_ps = sp.array(ss_table['p'])
# Filtering boring p-values
res_p_filter = comb_ps < p_thres
res_sids = res_sids[res_p_filter]
comb_ps = comb_ps[res_p_filter]
# ss_p_filter = marg_ps<p_thres
# ss_sids = ss_sids[ss_p_filter]
# marg_ps = marg_ps[ss_p_filter]
common_sids = sp.intersect1d(res_sids, ss_sids)
print 'Found %d SNPs in common' % (len(common_sids))
ss_filter = sp.in1d(ss_sids, common_sids)
res_filter = sp.in1d(res_sids, common_sids)
ss_sids = ss_sids[ss_filter]
res_sids = res_sids[res_filter]
marg_ps = marg_ps[ss_filter]
comb_ps = comb_ps[res_filter]
print 'Now sorting'
ss_index = sp.argsort(ss_sids)
res_index = sp.argsort(res_sids)
marg_ps = -sp.log10(marg_ps[ss_index])
comb_ps = -sp.log10(comb_ps[res_index])
with plt.style.context('fivethirtyeight'):
plt.plot(marg_ps, comb_ps, 'b.', alpha=0.2)
(x_min, x_max) = plt.xlim()
(y_min, y_max) = plt.ylim()
plt.plot([x_min, x_max], [y_min, y_max], 'k--', alpha=0.2)
plt.ylabel(ylabel)
plt.xlabel(xlabel)
plt.tight_layout()
plt.savefig(fig_filename)
plt.clf()
def test_lmm2(self):
"""another test, establishing an lmm-equivalent by a design matrix choice"""
for dn in self.datasets:
D = data.load(os.path.join(self.dir_name,dn))
#construct Kronecker LMM model which has the special case of standard LMM
#covar1: genotype matrix
N = D['K'].shape[0]
P = 3
K1r = D['K']
#K1c = SP.zeros([2,2])
#K1c[0,0] = 1
K1c = SP.eye(P)
K2r = SP.eye(N)
K2c = SP.eye(P)
#A = SP.zeros([1,2])
#A[0,0] =1
A = SP.eye(P)
Acov = SP.eye(P)
Xcov = D['Cov'][:,SP.newaxis]
X = D['X']
Y = D['Y'][:,SP.newaxis]
Y = SP.tile(Y,(1,P))
lmm = dlimix.CKroneckerLMM()
lmm.setK1r(K1r)
lmm.setK1c(K1c)
lmm.setK2r(K2r)
lmm.setK2c(K2c)
lmm.setSNPs(X)
#add covariates
lmm.addCovariates(Xcov,Acov)
#add SNP design
lmm.setSNPcoldesign(A)
lmm.setPheno(Y)
lmm.setNumIntervalsAlt(0)
lmm.setNumIntervals0(100)
lmm.process()
#get p-values with P-dof:
pv_Pdof = lmm.getPv().ravel()
#transform in P-values with a single DOF:
import scipy.stats as st
lrt = st.chi2.isf(pv_Pdof,P)/P
pv = st.chi2.sf(lrt,1)
#compare with single DOF P-values:
D2= ((SP.log10(pv)-SP.log10(D['pv']))**2)
RV = SP.sqrt(D2.mean())
#print "\n"
#print pv[0:10]
#print D['pv'][0:10]
#print RV
#pdb.set_trace()
self.assertTrue(RV<1E-6)
def residual_square(self, p, E_min, nb_minima):
a = P4Rm()
P4Rm.ParamDict["_fp_min"] = p
self.strain_DW()
res = f_Refl_fit(a.AllDataDict["geometry"], self.Data4f_Refl)
y_cal = convolve(abs(res) ** 2, a.ParamDict["resol"], mode="same")
y_cal = y_cal / y_cal.max() + a.AllDataDict["background"]
y_obs = a.ParamDict["Iobs"]
self.on_pass_data_to_thread(y_cal, p, E_min, nb_minima)
return ((log10(y_obs) - log10(y_cal)) ** 2).sum() / len(y_cal)
def to_log(x1, x2, x1err, x2err):
"""
Take linear measurements and uncertainties and transform to log values.
"""
logx1 = scipy.log10(scipy.array(x1))
logx2 = scipy.log10(scipy.array(x2))
x1err = scipy.log10(scipy.array(x1)+scipy.array(x1err)) - logx1
x2err = scipy.log10(scipy.array(x2)+scipy.array(x2err)) - logx2
return logx1, logx2, x1err, x2err
def residual(params, theory, data, linlog, sigma=None, logResidual=False):
"""Calculate residual for fitting"""
residuals = np.array([])
if sigma is None: sigma = 1.
P = theory.Y(data.X, params)
if not logResidual:
res = (P - data.Y)/sigma
else:
res = (sp.log10(P) - sp.log10(data.Y))/sigma
residuals = np.concatenate((residuals, res))
return residuals
def plot(self,truex,truey):
f,a = plt.subplots(3)
#print self.X.shape
a[0].plot(self.X[:,0].flatten(),'b')
a[0].set_ylabel("augx")
a[0].twinx().plot(self.C,'r')
a[1].plot(sp.log10(self.Rreg.flatten()-truey))
a[1].set_ylabel("regret")
a[2].plot([sum(self.C[:j]) for j in xrange(len(self.C))],sp.log10(self.Rreg.flatten()-truey))
a[2].set_ylabel("regret/c")
return
KS = SP.zeros((Y.shape[0], Y.shape[0]))
for iph in range(Y.shape[0]):
for jph in range(Y.shape[0]):
if SP.bitwise_and(phase_vec[iph] == phase_vec[jph],
phase_vec[iph] == 2):
KS[iph, jph] = 1
KG2M = SP.zeros((Y.shape[0], Y.shape[0]))
for iph in range(Y.shape[0]):
for jph in range(Y.shape[0]):
if SP.bitwise_and(phase_vec[iph] == phase_vec[jph],
phase_vec[iph] == 3):
KG2M[iph, jph] = 1
#intra-phase variations in cell size
sfCellSize = SP.log10(f['ratioEndo'][:])
sfCellSize -= sfCellSize.mean()
sfCellSize = sfCellSize.reshape(1, sfCellSize.shape[0])
Ksize = SP.dot(sfCellSize.transpose(), sfCellSize)
Ksize /= Ksize.diagonal().mean()
# filter cell cycle genes
idx_cell_cycle = SP.union1d(cellcyclegenes_filter,
cellcyclegenes_filterCB600)
Ymean2 = Y.mean(0)**2 > 0
idx_cell_cycle_noise_filtered = SP.intersect1d(
idx_cell_cycle, SP.array(SP.where(Ymean2.ravel() > 0)))
Ycc = Y[:, idx_cell_cycle_noise_filtered]
#Fit GPLVM to data
k = 1 # number of latent factors
def plotFigure4(subfolder, FigFolder, rank, subFigName, Indx, oversample,
FigNumber, FigRows, FigCols):
font = {'family': 'sans-serif', 'weight': 'normal', 'size': 14}
matplotlib.rc('font', **font)
# set tick width
matplotlib.rcParams['xtick.major.size'] = 7
matplotlib.rcParams['xtick.major.width'] = 1.0
matplotlib.rcParams['xtick.minor.size'] = 3
matplotlib.rcParams['xtick.minor.width'] = 1.0
matplotlib.rcParams['ytick.major.size'] = 7
matplotlib.rcParams['ytick.major.width'] = 1.0
matplotlib.rcParams['ytick.minor.size'] = 3
matplotlib.rcParams['ytick.minor.width'] = 1.0
matplotlib.rcParams['axes.linewidth'] = 1.0 #set the value globally
# Make LATEX font the same as text font:
matplotlib.rcParams['mathtext.fontset'] = 'custom'
matplotlib.rcParams['mathtext.rm'] = 'Bitstream Vera Sans'
matplotlib.rcParams['mathtext.it'] = 'Bitstream Vera Sans:italic'
matplotlib.rcParams['mathtext.bf'] = 'Bitstream Vera Sans:bold'
# Get Spectral errors:
MeanErrOnSD = sp.load(subfolder + '/' + 'MeanSpecErrOnJ_' +
str(int(oversample)) + 'over.npy')
MeanErrOnSDT = sp.load(subfolder + '/' + 'MeanSpecErrOnJT_' +
str(int(oversample)) + 'over.npy')
MeanErrNystrom = sp.load(subfolder + '/' + 'MeanSpecErrNystrom_' +
str(int(oversample)) + 'over.npy')
MeanErrPinched = sp.load(subfolder + '/' + 'MeanSpecErrPinched_' +
str(int(oversample)) + 'over.npy')
# Get sketch sizes:
ViewVec = sp.load(subfolder + '/' + 'ViewVec.npy')
# Get min and max values:
Errs = [MeanErrOnSD, MeanErrOnSDT, MeanErrNystrom, MeanErrPinched]
ymin = 1.e99
ymax = -1.e99
for Err in Errs:
emax = max(Err)
if emax > ymax:
ymax = emax
emin = min(Err)
if emin < ymin:
ymin = emin
fig = plt.figure(1, figsize=(12, 12))
ax = plt.subplot(FigRows, FigCols, FigNumber)
plt.plot(ViewVec,
MeanErrNystrom,
'k-.',
markersize=11,
markerfacecolor='k',
markeredgecolor='k',
markeredgewidth=2,
linewidth=2.0)
plt.plot(ViewVec,
MeanErrPinched,
'k--',
markersize=11,
markerfacecolor='g',
markeredgecolor='g',
markeredgewidth=2,
linewidth=2.0)
plt.plot(ViewVec,
MeanErrOnSD,
'rv',
markersize=11,
markerfacecolor='r',
markeredgecolor='r',
linewidth=1.5)
plt.plot(ViewVec,
MeanErrOnSDT,
'bo',
markersize=11,
markerfacecolor='b',
markeredgecolor='b',
linewidth=1.5)
plt.yscale('log')
# For y-ticks:
ytksMin = floor(sp.log10(MeanErrNystrom[4] * 0.5))
ytksMax = ceil(sp.log10(MeanErrPinched[0] * 2.))
plt.xlabel('Views, $v$', labelpad=10)
plt.ylabel('Relative Spectral Error', labelpad=10)
# Set y axis limits:
if sp.absolute(ytksMin - ytksMax) < 12.5:
plt.yticks(sp.logspace(-16, 4, 11))
else:
plt.yticks(sp.logspace(-16, 4, 6))
plt.ylim(10.**(ytksMin), 10.**(ytksMax))
plt.xlim(1.5, 6.5)
ax.xaxis.set_major_formatter(ScalarFormatter())
ax.minorticks_off()
ax.set_title('$S_\mathrm{D}$' + ' ($p=$' + str(rank) + ', $l = $' +
str(oversample) + ')',
y=1.02)
ax.text(-0.3,
1.1,
'(' + string.ascii_lowercase[Indx] + ')',
transform=ax.transAxes,
weight='bold')
if Indx == 7:
ax.legend(['Prolonged', 'Pinched', 'Alg. 2 on $J$', 'Alg. 2 on $J^*$'],
numpoints=1,
loc=9,
bbox_to_anchor=(0.4, -0.24),
ncol=5,
frameon=False)
fig.subplots_adjust(left=0.01,
bottom=0.01,
right=0.99,
top=0.99,
wspace=0.4,
hspace=0.4)
fig.savefig(FigFolder + '/Figure4.pdf', format='pdf', bbox_inches='tight')
def get_freqs_db(signal, sample_rate):
freqs, fft_abs = get_freqs_fft(signal, sample_rate)
fft_db = 20 * scipy.log10(fft_abs)
return freqs, fft_db
def main():
tstart = time.time()
tb = pfb_top_block()
tb.run()
tend = time.time()
print "Run time: %f" % (tend - tstart)
if 1:
fig_in = pylab.figure(1, figsize=(16, 9), facecolor="w")
fig1 = pylab.figure(2, figsize=(16, 9), facecolor="w")
fig2 = pylab.figure(3, figsize=(16, 9), facecolor="w")
Ns = 1000
Ne = 10000
fftlen = 8192
winfunc = scipy.blackman
fs = tb._fs
# Plot the input signal on its own figure
d = tb.snk_i.data()[Ns:Ne]
spin_f = fig_in.add_subplot(2, 1, 1)
X, freq = mlab.psd(d,
NFFT=fftlen,
noverlap=fftlen / 4,
Fs=fs,
window=lambda d: d * winfunc(fftlen),
scale_by_freq=True)
X_in = 10.0 * scipy.log10(abs(X))
f_in = scipy.arange(-fs / 2.0, fs / 2.0, fs / float(X_in.size))
pin_f = spin_f.plot(f_in, X_in, "b")
spin_f.set_xlim([min(f_in), max(f_in) + 1])
spin_f.set_ylim([-200.0, 50.0])
spin_f.set_title("Input Signal", weight="bold")
spin_f.set_xlabel("Frequency (Hz)")
spin_f.set_ylabel("Power (dBW)")
Ts = 1.0 / fs
Tmax = len(d) * Ts
t_in = scipy.arange(0, Tmax, Ts)
x_in = scipy.array(d)
spin_t = fig_in.add_subplot(2, 1, 2)
pin_t = spin_t.plot(t_in, x_in.real, "b")
pin_t = spin_t.plot(t_in, x_in.imag, "r")
spin_t.set_xlabel("Time (s)")
spin_t.set_ylabel("Amplitude")
Ncols = int(scipy.floor(scipy.sqrt(tb._M)))
Nrows = int(scipy.floor(tb._M / Ncols))
if (tb._M % Ncols != 0):
Nrows += 1
# Plot each of the channels outputs. Frequencies on Figure 2 and
# time signals on Figure 3
fs_o = tb._fs / tb._M
Ts_o = 1.0 / fs_o
Tmax_o = len(d) * Ts_o
for i in xrange(len(tb.snks)):
# remove issues with the transients at the beginning
# also remove some corruption at the end of the stream
# this is a bug, probably due to the corner cases
d = tb.snks[i].data()[Ns:Ne]
sp1_f = fig1.add_subplot(Nrows, Ncols, 1 + i)
X, freq = mlab.psd(d,
NFFT=fftlen,
noverlap=fftlen / 4,
Fs=fs_o,
window=lambda d: d * winfunc(fftlen),
scale_by_freq=True)
X_o = 10.0 * scipy.log10(abs(X))
f_o = scipy.arange(-fs_o / 2.0, fs_o / 2.0, fs_o / float(X_o.size))
p2_f = sp1_f.plot(f_o, X_o, "b")
sp1_f.set_xlim([min(f_o), max(f_o) + 1])
sp1_f.set_ylim([-200.0, 50.0])
sp1_f.set_title(("Channel %d" % i), weight="bold")
sp1_f.set_xlabel("Frequency (Hz)")
sp1_f.set_ylabel("Power (dBW)")
x_o = scipy.array(d)
t_o = scipy.arange(0, Tmax_o, Ts_o)
sp2_o = fig2.add_subplot(Nrows, Ncols, 1 + i)
p2_o = sp2_o.plot(t_o, x_o.real, "b")
p2_o = sp2_o.plot(t_o, x_o.imag, "r")
sp2_o.set_xlim([min(t_o), max(t_o) + 1])
sp2_o.set_ylim([-2, 2])
sp2_o.set_title(("Channel %d" % i), weight="bold")
sp2_o.set_xlabel("Time (s)")
sp2_o.set_ylabel("Amplitude")
pylab.show()
def snr_est_simple(signal):
s = scipy.mean(abs(signal)**2)
n = 2 * scipy.var(abs(signal))
snr_rat = s / n
return 10.0 * scipy.log10(snr_rat), snr_rat
print('sc.sqrt(a) = ', sc.sqrt(a))
print('sc.sqrt(z) = ', sc.sqrt(z))
c = np.sqrt(a)
w = np.sqrt(z)
print(c**2)
print(w**2)
c = sc.sqrt(a)
w = sc.sqrt(z)
print(c**2)
print(w**2)
print(np.exp(a))
print(np.sin(a))
print(np.log(a))
print(np.log10(a))
print(np.exp(a))
print(np.sin(z))
print(np.log(z))
print(np.log10(z))
print(sc.exp(a))
print(sc.sin(a))
print(sc.log(a))
print(sc.log10(a))
print(sc.exp(z))
print(sc.sin(z))
print(sc.log(z))
print(sc.log10(z))
def heatmap_from_bam(chrm,
start,
stop,
files,
subsample=0,
verbose=False,
bins=None,
log=False,
ax=None,
ymax=0,
outfile=None,
frm='pdf',
xlim=None,
title=None,
xoff=None,
yoff=None,
intron_cov=False,
cmap=None,
col_idx=None):
"""This function takes a list of bam files and a set of coordinates (chrm, start, stop), to
plot a coverage heatmap over all files in that region."""
### subsampling
if subsample > 0 and len(files) > subsample:
npr.seed(23)
files = sp.array(files)
files = npr.choice(files, subsample)
### augment chromosome name
#chr_name = 'chr%s' % chrm
chr_name = chrm
(counts, intron_counts, intron_list) = _get_counts(chr_name,
start,
stop,
files,
intron_cov,
verbose=verbose,
collapsed=False)
if ax is None:
fig = plt.figure(figsize=(10, 4))
ax = fig.add_subplot(111)
if intron_cov:
data = intron_counts
else:
data = counts
if col_idx is not None:
data = data[:, col_idx]
if log:
data = sp.log10(data + 1)
if cmap is not None:
ax.matshow(data, cmap=cmap, aspect='auto')
else:
ax.matshow(data, aspect='auto')
if outfile is not None:
plt.savefig(outfile, dpi=300, format=frm)
def porosimetry(im, sizes=25, inlets=None, access_limited=True, mode='hybrid'):
r"""
Performs a porosimetry simulution on the image
Parameters
----------
im : ND-array
An ND image of the porous material containing True values in the
pore space.
sizes : array_like or scalar
The sizes to invade. If a list of values of provided they are used
directly. If a scalar is provided then that number of points spanning
the min and max of the distance transform are used.
inlets : ND-array, boolean
A boolean mask with True values indicating where the invasion
enters the image. By default all faces are considered inlets,
akin to a mercury porosimetry experiment. Users can also apply
solid boundaries to their image externally before passing it in,
allowing for complex inlets like circular openings, etc. This argument
is only used if ``access_limited`` is ``True``.
access_limited : Boolean
This flag indicates if the intrusion should only occur from the
surfaces (``access_limited`` is True, which is the default), or
if the invading phase should be allowed to appear in the core of
the image. The former simulates experimental tools like mercury
intrusion porosimetry, while the latter is useful for comparison
to gauge the extent of shielding effects in the sample.
mode : string
Controls with method is used to compute the result. Options are:
*'hybrid'* - (default) Performs a distance tranform of the void space,
thresholds to find voxels larger than ``sizes[i]``, trims the resulting
mask if ``access_limitations`` is ``True``, then dilates it using the
efficient fft-method to obtain the non-wetting fluid configuration.
*'dt'* - Same as 'hybrid', except uses a second distance transform,
relative to the thresholded mask, to find the invading fluid
configuration. The choice of 'dt' or 'hybrid' depends on speed, which
is system and installation specific.
*'mio'* - Using a single morphological image opening step to obtain the
invading fluid confirguration directly, *then* trims if
``access_limitations`` is ``True``. This method is not ideal and is
included mostly for comparison purposes. The morphological operations
are done using fft-based method implementations.
Returns
-------
An ND-image with voxel values indicating the sphere radius at which it
becomes accessible from the inlets. This image can be used to find
invading fluid configurations as a function of applied capillary pressure
by applying a boolean comparison: ``inv_phase = im > r`` where ``r`` is
the radius (in voxels) of the invading sphere. Of course, ``r`` can be
converted to capillary pressure using your favorite model.
See Also
--------
fftmorphology
"""
def trim_blobs(im, inlets):
temp = sp.zeros_like(im)
temp[inlets] = True
labels, N = spim.label(im + temp)
im = im ^ (clear_border(labels=labels) > 0)
return im
dt = spim.distance_transform_edt(im > 0)
if inlets is None:
inlets = get_border(im.shape, mode='faces')
inlets = sp.where(inlets)
if isinstance(sizes, int):
sizes = sp.logspace(start=sp.log10(sp.amax(dt)), stop=0, num=sizes)
else:
sizes = sp.sort(a=sizes)[-1::-1]
if im.ndim == 2:
strel = ps_disk
else:
strel = ps_ball
imresults = sp.zeros(sp.shape(im))
if mode == 'mio':
pw = int(sp.floor(dt.max()))
impad = sp.pad(im, mode='symmetric', pad_width=pw)
imresults = sp.zeros(sp.shape(impad))
for r in tqdm(sizes):
imtemp = fftmorphology(impad, strel(r), mode='opening')
if access_limited:
imtemp = trim_blobs(imtemp, inlets)
if sp.any(imtemp):
imresults[(imresults == 0) * imtemp] = r
if im.ndim == 2:
imresults = imresults[pw:-pw, pw:-pw]
else:
imresults = imresults[pw:-pw, pw:-pw, pw:-pw]
elif mode == 'dt':
for r in tqdm(sizes):
imtemp = dt >= r
if access_limited:
imtemp = trim_blobs(imtemp, inlets)
if sp.any(imtemp):
imtemp = spim.distance_transform_edt(~imtemp) < r
imresults[(imresults == 0) * imtemp] = r
elif mode == 'hybrid':
for r in tqdm(sizes):
imtemp = dt >= r
if access_limited:
imtemp = trim_blobs(imtemp, inlets)
if sp.any(imtemp):
imtemp = fftconvolve(imtemp, strel(r), mode='same') > 0.0001
imresults[(imresults == 0) * imtemp] = r
else:
raise Exception('Unreckognized mode ' + mode)
return imresults
def __init__(self,
ll,
fl,
iv,
thid,
ra,
dec,
zqso,
plate,
mjd,
fid,
order,
diff=None,
reso=None,
mmef=None):
qso.__init__(self, thid, ra, dec, zqso, plate, mjd, fid)
if not self.ebv_map is None:
corr = unred(10**ll, self.ebv_map[thid])
fl /= corr
iv *= corr**2
## cut to specified range
bins = sp.floor((ll - forest.lmin) / forest.dll + 0.5).astype(int)
ll = forest.lmin + bins * forest.dll
w = (ll >= forest.lmin)
w = w & (ll < forest.lmax)
w = w & (ll - sp.log10(1. + self.zqso) > forest.lmin_rest)
w = w & (ll - sp.log10(1. + self.zqso) < forest.lmax_rest)
w = w & (iv > 0.)
if w.sum() == 0:
return
bins = bins[w]
ll = ll[w]
fl = fl[w]
iv = iv[w]
## mmef is the mean expected flux fraction using the mock continuum
if mmef is not None:
mmef = mmef[w]
if diff is not None:
diff = diff[w]
if reso is not None:
reso = reso[w]
## rebin
cll = forest.lmin + sp.arange(bins.max() + 1) * forest.dll
cfl = sp.zeros(bins.max() + 1)
civ = sp.zeros(bins.max() + 1)
if mmef is not None:
cmmef = sp.zeros(bins.max() + 1)
ccfl = sp.bincount(bins, weights=iv * fl)
cciv = sp.bincount(bins, weights=iv)
if mmef is not None:
ccmmef = sp.bincount(bins, weights=iv * mmef)
if diff is not None:
cdiff = sp.bincount(bins, weights=iv * diff)
if reso is not None:
creso = sp.bincount(bins, weights=iv * reso)
cfl[:len(ccfl)] += ccfl
civ[:len(cciv)] += cciv
if mmef is not None:
cmmef[:len(ccmmef)] += ccmmef
w = (civ > 0.)
if w.sum() == 0:
return
ll = cll[w]
fl = cfl[w] / civ[w]
iv = civ[w]
if mmef is not None:
mmef = cmmef[w] / civ[w]
if diff is not None:
diff = cdiff[w] / civ[w]
if reso is not None:
reso = creso[w] / civ[w]
## Flux calibration correction
if not self.correc_flux is None:
correction = self.correc_flux(ll)
fl /= correction
iv *= correction**2
if not self.correc_ivar is None:
correction = self.correc_ivar(ll)
iv /= correction
self.T_dla = None
self.ll = ll
self.fl = fl
self.iv = iv
self.mmef = mmef
self.order = order
#if diff is not None :
self.diff = diff
self.reso = reso
# else :
# self.diff = sp.zeros(len(ll))
# self.reso = sp.ones(len(ll))
# compute means
if reso is not None: self.mean_reso = sum(reso) / float(len(reso))
err = 1.0 / sp.sqrt(iv)
SNR = fl / err
self.mean_SNR = sum(SNR) / float(len(SNR))
lam_lya = constants.absorber_IGM["LYA"]
self.mean_z = (sp.power(10., ll[len(ll) - 1]) +
sp.power(10., ll[0])) / 2. / lam_lya - 1.0
def _scale_psf(self, input_irf_file, config):
"""
This internal method scales the IRF PSF extension.
Parameters
----------
input_irf_file: pyfits.HDUList
Open pyfits IRF file, which contains the PSF that should be scaled.
config: dict
A dictionary with the scaling settings. Must have following keys defined:
"energy_scaling": dict
Contains setting for the energy scaling (see the structure below).
"angular_scaling": dict
Contains setting for the off-center angle scaling (see the structure below).
In both cases, internally the above dictionaries should contain:
"err_func_type": str
The name of the scaling function to use. Accepted values are: "constant",
"gradient" and "step".
If err_func_type == "constant":
scale: float
The scale factor. passing 1.0 results in no scaling.
If err_func_type == "gradient":
scale: float
The scale factor. passing 0.0 results in no scaling.
range_min: float
The x value (energy or off-center angle), that corresponds to -1 scale.
range_max: float
The x value (energy or off-center angle), that corresponds to +1 scale.
If err_func_type == "step":
scale: float
The scale factor. passing 0.0 results in no scaling.
transition_pos: list
The list of x values (energy or off-center angle), at which
step-like transitions occur. If scaling the energy dependence,
values must be in TeVs, if angular - in degrees.
transition_widths: list
The list of step-like transition widths, that correspond to transition_pos.
For energy scaling the widths must be in log10 scale.
Returns
-------
None
"""
# Find all "sigma" values - tells how many PSF components we have in the IRF file
column_names = [
col.name.lower()
for col in input_irf_file['POINT SPREAD FUNCTION'].columns
]
sigma_columns = list(
filter(lambda s: "sigma" in s.lower(), column_names))
# --------------------------
# Reading the PSF parameters
self._psf = dict()
self._psf['Elow'] = input_irf_file['POINT SPREAD FUNCTION'].data[
'Energ_lo'][0].copy()
self._psf['Ehigh'] = input_irf_file['POINT SPREAD FUNCTION'].data[
'Energ_hi'][0].copy()
self._psf['ThetaLow'] = input_irf_file['POINT SPREAD FUNCTION'].data[
'Theta_lo'][0].copy()
self._psf['ThetaHi'] = input_irf_file['POINT SPREAD FUNCTION'].data[
'Theta_hi'][0].copy()
for i in range(0, len(sigma_columns)):
sigma_name = 'sigma_{:d}'.format(i + 1)
self._psf[sigma_name] = input_irf_file[
'POINT SPREAD FUNCTION'].data[sigma_name][0].transpose().copy(
)
self._psf['E'] = scipy.sqrt(self._psf['Elow'] * self._psf['Ehigh'])
self._psf['Theta'] = (self._psf['ThetaLow'] +
self._psf['ThetaHi']) / 2.0
# --------------------------
# Creating the energy-theta mesh grid
energy, theta = scipy.meshgrid(self._psf['E'],
self._psf['Theta'],
indexing='ij')
# ---------------------------------
# Scaling the PSF energy dependence
# Constant error function
if config['energy_scaling']['err_func_type'] == "constant":
scale_params = config['energy_scaling']["constant"]
# Constant scaling. Loop over all "sigma" values and scale them by the same factor.
for sigma_column in sigma_columns:
self._psf[
sigma_column +
'_new'] = scale_params['scale'] * self._psf[sigma_column]
# Gradients error function
elif config['energy_scaling']['err_func_type'] == "gradient":
scale_params = config['energy_scaling']["gradient"]
for sigma_column in sigma_columns:
self._psf[sigma_column + '_new'] = self._psf[sigma_column] * (
1 + scale_params['scale'] *
gradient(scipy.log10(energy),
scipy.log10(scale_params['range_min']),
scipy.log10(scale_params['range_max'])))
# Step error function
elif config['energy_scaling']['err_func_type'] == "step":
scale_params = config['energy_scaling']["step"]
break_points = list(
zip(scipy.log10(scale_params['transition_pos']),
scale_params['transition_widths']))
for sigma_column in sigma_columns:
self._psf[sigma_column + '_new'] = self._psf[sigma_column] * (
1 + scale_params['scale'] *
step(scipy.log10(energy), break_points))
else:
raise ValueError("Unknown PSF scaling function {:s}".format(
config['energy_scaling']['err_func_type']))
# ---------------------------------
# ---------------------------------
# Scaling the PSF angular dependence
# Constant error function
if config['angular_scaling']['err_func_type'] == "constant":
scale_params = config['angular_scaling']["constant"]
# Constant scaling. Loop over all "sigma" values and scale them by the same factor.
for sigma_column in sigma_columns:
# input_irf_file['POINT SPREAD FUNCTION'].data[sigma_column] *= scale_params['scale']
self._psf[sigma_column +
'_new'] = scale_params['scale'] * self._psf[
sigma_column + '_new']
# Gradients error function
elif config['angular_scaling']['err_func_type'] == "gradient":
scale_params = config['angular_scaling']["gradient"]
for sigma_column in sigma_columns:
self._psf[sigma_column +
'_new'] = self._psf[sigma_column + '_new'] * (
1 + scale_params['scale'] *
gradient(theta, scale_params['range_min'],
scale_params['range_max']))
# Step error function
elif config['angular_scaling']['err_func_type'] == "step":
scale_params = config['angular_scaling']["step"]
break_points = list(
zip(scale_params['transition_pos'],
scale_params['transition_widths']))
for sigma_column in sigma_columns:
self._psf[
sigma_column +
'_new'] = self._psf[sigma_column + '_new'] * (
1 + scale_params['scale'] * step(theta, break_points))
else:
raise ValueError("Unknown PSF scaling function {:s}".format(
config['angular_scaling']['err_func_type']))
# ---------------------------------
# Recording the scaled PSF
for i in range(0, len(sigma_columns)):
sigma_name = 'sigma_{:d}'.format(i + 1)
input_irf_file['POINT SPREAD FUNCTION'].data[sigma_name][
0] = self._psf[sigma_name + '_new'].transpose()
def pwrlaw_plot (xdata, ydata, yerr):
from scipy import linspace, randn, log10, optimize, sqrt
powerlaw = lambda x, amp, index: amp * (x ** index)
logx = log10(xdata)
logy = log10(ydata)
logyerr = yerr / ydata
# define our (line) fitting function
fitfunc = lambda p, x: p[0] + p[1] * x
errfunc = lambda p, x, y, err: (y - fitfunc(p, x)) / err
pinit = [1.0, -1.0]
out = optimize.leastsq(errfunc, pinit,
args=(logx, logy, logyerr), full_output=1)
pfinal = out[0]
covar = out[1]
print pfinal
print covar
index = pfinal[1]
amp = 10.0 ** pfinal[0]
indexErr = sqrt(covar[0][0])
ampErr = sqrt(covar[1][1]) * amp
print index
# ########
# plotting
# ########
# ax.plot(ydata)
# ax.plot(pl_sequence)
fig, axs = plt.subplots(2, 1)
axs[0].plot(xdata, powerlaw(xdata, amp, index)) # Fit
axs[0].errorbar(xdata, ydata, yerr=yerr, fmt='k.') # Data
(yh1, yh2) = (axs[0].get_ylim()[1] * .9, axs[0].get_ylim()[1] * .8)
xh = axs[0].get_xlim()[0] * 1.1
print axs[0].get_ylim()
print (yh1, yh2)
axs[0].text(xh, yh1, 'Ampli = %5.2f +/- %5.2f' % (amp, ampErr))
axs[0].text(xh, yh2, 'Index = %5.2f +/- %5.2f' % (index, indexErr))
axs[0].set_title('Best Fit Power Law')
axs[0].set_xlabel('X')
axs[0].set_ylabel('Y')
# xlim(1, 11)
#
# subplot(2, 1, 2)
axs[1].loglog(xdata, powerlaw(xdata, amp, index))
axs[1].errorbar(xdata, ydata, yerr=yerr, fmt='k.') # Data
axs[1].set_xlabel('X (log scale)')
axs[1].set_ylabel('Y (log scale)')
import datetime
figfname = datetime.datetime.now().strftime("%d%b%y") + "_pl"
plt.savefig(figfname, bbox_inches='tight')
return figfname
def _scale_edisp(self, input_irf_file, config):
"""
This internal method scales the IRF energy dispersion through the Migration Matrix.
Two scalings can be applied: (1) vs energy and (2) vs off-axis angle. In both cases
the scaling function is taken as (1 + scale * tanh((x-x0)/dx)). In case (1) the scaling
is performed in log-energy.
Parameters
----------
input_irf_file: pyfits.HDUList
Open pyfits IRF file, which contains the Migration Matrix that should be scaled.
config: dict
A dictionary with the scaling settings. Must have following keys defined:
"energy_scaling": dict
Contains setting for the energy scaling (see the structure below).
"angular_scaling": dict
Contains setting for the off-center angle scaling (see the structure below).
In both cases, internally the above dictionaries should contain:
"err_func_type": str
The name of the scaling function to use. Accepted values are: "constant",
"gradient" and "step".
If err_func_type == "constant":
scale: float
The scale factor. passing 1.0 results in no scaling.
If err_func_type == "gradient":
scale: float
The scale factor. passing 0.0 results in no scaling.
range_min: float
The x value (energy or off-center angle), that corresponds to -1 scale.
range_max: float
The x value (energy or off-center angle), that corresponds to +1 scale.
If err_func_type == "step":
scale: float
The scale factor. passing 0.0 results in no scaling.
transition_pos: list
The list of x values (energy or off-center angle), at which
step-like transitions occur. If scaling the energy dependence,
values must be in TeVs, if angular - in degrees.
transition_widths: list
The list of step-like transition widths, that correspond to transition_pos.
For energy scaling the widths must be in log10 scale.
Returns
-------
None
"""
# Reading the MATRIX parameters
self._edisp = dict()
self._edisp['Elow'] = input_irf_file['ENERGY DISPERSION'].data[
'ETRUE_LO'][0].copy()
self._edisp['Ehigh'] = input_irf_file['ENERGY DISPERSION'].data[
'ETRUE_HI'][0].copy()
self._edisp['ThetaLow'] = input_irf_file['ENERGY DISPERSION'].data[
'THETA_LO'][0].copy()
self._edisp['ThetaHi'] = input_irf_file['ENERGY DISPERSION'].data[
'THETA_HI'][0].copy()
self._edisp['Mlow'] = input_irf_file['ENERGY DISPERSION'].data[
'MIGRA_LO'][0].copy()
self._edisp['Mhigh'] = input_irf_file['ENERGY DISPERSION'].data[
'MIGRA_HI'][0].copy()
self._edisp['Matrix_'] = input_irf_file['ENERGY DISPERSION'].data[
'MATRIX'][0].transpose().copy()
self._edisp['E'] = scipy.sqrt(self._edisp['Elow'] *
self._edisp['Ehigh'])
self._edisp['M'] = (self._edisp['Mlow'] + self._edisp['Mhigh']) / 2.0
self._edisp['T'] = (self._edisp['ThetaLow'] +
self._edisp['ThetaHi']) / 2.0
# Creating the energy-migration matix-theta mesh grid
energy, migration, theta = scipy.meshgrid(self._edisp['E'],
self._edisp['M'],
self._edisp['T'],
indexing='ij')
# -------------------------------------------
# Scaling the Matrix energy dependence
# Constant error function
if config['energy_scaling']['err_func_type'] == "constant":
self._edisp['Matrix_new'] = self._edisp['Matrix_'] * config[
'energy_scaling']['constant']['scale']
# Gradients error function
elif config['energy_scaling']['err_func_type'] == "gradient":
scaling_params = config['energy_scaling']['gradient']
self._edisp['Matrix_new'] = self._edisp['Matrix_'] * (
1. + scaling_params['scale'] *
gradient(scipy.log10(energy),
scipy.log10(scaling_params['range_min']),
scipy.log10(scaling_params['range_max'])))
# Step error function
elif config['energy_scaling']['err_func_type'] == "step":
scaling_params = config['energy_scaling']['step']
break_points = list(
zip(scipy.log10(scaling_params['transition_pos']),
scaling_params['transition_widths']))
self._edisp['Matrix_new'] = self._edisp['Matrix_'] * (
1 + scaling_params['scale'] *
step(scipy.log10(energy), break_points))
else:
raise ValueError(
"Edisp energy scaling: unknown scaling function type '{:s}'".
format(config['energy_scaling']['err_func_type']))
# -------------------------------------------
# Scaling the Matrix off-axis angle dependence
# Constant error function
if config['angular_scaling']['err_func_type'] == "constant":
self._edisp['Matrix_new'] = self._edisp['Matrix_new'] * config[
'angular_scaling']['constant']['scale']
# Gradients error function
elif config['angular_scaling']['err_func_type'] == "gradient":
scaling_params = config['angular_scaling']['gradient']
self._edisp['Matrix_new'] = self._edisp['Matrix_new'] * (
1. + scaling_params['scale'] *
gradient(scipy.log10(theta),
scipy.log10(scaling_params['range_min']),
scipy.log10(scaling_params['range_max'])))
# Step error function
elif config['angular_scaling']['err_func_type'] == "step":
scaling_params = config['angular_scaling']['step']
break_points = list(
zip(scipy.log10(scaling_params['transition_pos']),
scaling_params['transition_widths']))
self._edisp['Matrix_new'] = self._edisp['Matrix_new'] * (
1 + scaling_params['scale'] *
step(scipy.log10(theta), break_points))
else:
raise ValueError(
"Edisp angular scaling: unknown scaling function type '{:s}'".
format(config['energy_scaling']['err_func_type']))
# ------------------------------------------
# Recording the scaled Matrix
input_irf_file['ENERGY DISPERSION'].data['MATRIX'][0] = self._edisp[
'Matrix_new'].transpose()
def _scale_bkg(self, input_irf_file, config):
"""
This internal method scales the IRF Background Rate, N.
Two scalings can be applied: (1) vs energy and (2) vs off-axis angle. In both cases
the scaling function is taken as (1 + scale * tanh((x-x0)/dx)). In case (1) the scaling
is performed in log-energy.
Parameters
----------
input_irf_file: pyfits.HDUList
Open pyfits IRF file, which contains the Aeff that should be scaled.
config: dict
A dictionary with the scaling settings. Must have following keys defined:
"energy_scaling": dict
Contains setting for the energy scaling (see the structure below).
"angular_scaling": dict
Contains setting for the off-center angle scaling (see the structure below).
In both cases, internally the above dictionaries should contain:
"err_func_type": str
The name of the scaling function to use. Accepted values are: "constant",
"gradient" and "step".
If err_func_type == "constant":
scale: float
The scale factor. passing 1.0 results in no scaling.
If err_func_type == "gradient":
scale: float
The scale factor. passing 0.0 results in no scaling.
range_min: float
The x value (energy or off-center angle), that corresponds to -1 scale.
range_max: float
The x value (energy or off-center angle), that corresponds to +1 scale.
If err_func_type == "step":
scale: float
The scale factor. passing 0.0 results in no scaling.
transition_pos: list
The list of x values (energy or off-center angle), at which
step-like transitions occur. If scaling the energy dependence,
values must be in TeVs, if angular - in degrees.
transition_widths: list
The list of step-like transition widths, that correspond to transition_pos.
For energy scaling the widths must be in log10 scale.
Returns
-------
None
"""
# Reading the Background parameters.
self._bkg = dict()
self._bkg['dxlow'] = input_irf_file['BACKGROUND'].data['detx_lo'][
0].copy()
self._bkg['dxhigh'] = input_irf_file['BACKGROUND'].data['detx_hi'][
0].copy()
self._bkg['dylow'] = input_irf_file['BACKGROUND'].data['dety_lo'][
0].copy()
self._bkg['dyhigh'] = input_irf_file['BACKGROUND'].data['dety_hi'][
0].copy()
self._bkg['Elow'] = input_irf_file['BACKGROUND'].data['energ_lo'][
0].copy()
self._bkg['Ehigh'] = input_irf_file['BACKGROUND'].data['energ_hi'][
0].copy()
self._bkg['bckgnd'] = input_irf_file['BACKGROUND'].data['bgd'][
0].transpose().copy()
self._bkg['detx'] = (self._bkg['dxlow'] + self._bkg['dxhigh']) / 2.0
self._bkg['dety'] = (self._bkg['dylow'] + self._bkg['dyhigh']) / 2.0
self._bkg['enrg'] = scipy.sqrt(self._bkg['Elow'] * self._bkg['Ehigh'])
# Creating the [detx - dety - energy] mesh grid.
detx, dety, energy = scipy.meshgrid(self._bkg['detx'],
self._bkg['dety'],
self._bkg['enrg'],
indexing='ij')
# ----------------------------------
# Scaling the Background Energy dependence.
# Constant error function
if config['energy_scaling']['err_func_type'] == "constant":
self._bkg['bckgnd_new'] = self._bkg['bckgnd'] * config[
'energy_scaling']['constant']['scale']
# Gradients error function
elif config['energy_scaling']['err_func_type'] == "gradient":
scaling_params = config['energy_scaling']['gradient']
self._bkg['bckgnd_new'] = self._bkg['bckgnd'] * (
1 + scaling_params['scale'] *
gradient(scipy.log10(energy),
scipy.log10(scaling_params['range_min']),
scipy.log10(scaling_params['range_max'])))
# Step error function
elif config['energy_scaling']['err_func_type'] == "step":
scaling_params = config['energy_scaling']['step']
break_points = list(
zip(scipy.log10(scaling_params['transition_pos']),
scaling_params['transition_widths']))
self._bkg['bckgnd_new'] = self._bkg['bckgnd'] * (
1 + scaling_params['scale'] *
step(scipy.log10(energy), break_points))
else:
raise ValueError(
"Background energy scaling: unknown scaling function type '{:s}'"
.format(config['energy_scaling']['err_func_type']))
# -------------------------------------------
# Scaling the Background Angular dependence
# Angular (FOV coordinate X-axis binning)
# Constant error function
if config['angular_scaling']['err_func_type'] == "constant":
self._bkg['bckgnd_new'] = self._bkg['bckgnd_new'] * config[
'angular_scaling']['constant']['scale']
# Gradients error function
elif config['angular_scaling']['err_func_type'] == "gradient":
scaling_params = config['angular_scaling']['gradient']
self._bkg['bckgnd_new'] = self._bkg['bckgnd_new'] * (
1. + scaling_params['scale'] *
gradient(detx, scaling_params['range_min'],
scaling_params['range_max']))
# Step error function
elif config['angular_scaling']['err_func_type'] == "step":
scaling_params = config['angular_scaling']['step']
break_points = list(
zip(scaling_params['transition_pos'],
scaling_params['transition_widths']))
self._bkg['bckgnd_new'] = self._bkg['bckgnd_new'] * (
1 + scaling_params['scale'] * step(detx, break_points))
else:
raise ValueError(
"Background angular scaling: unknown scaling function type '{:s}'"
.format(config['angular_scaling']['err_func_type']))
# Angular (FOV coordinate Y-axis binning)
# Constant error function
if config['angular_scaling']['err_func_type'] == "constant":
self._bkg['bckgnd_new'] = self._bkg['bckgnd_new'] * config[
'angular_scaling']['constant']['scale']
# Gradients error function
elif config['angular_scaling']['err_func_type'] == "gradient":
scaling_params = config['angular_scaling']['gradient']
self._bkg['bckgnd_new'] = self._bkg['bckgnd_new'] * (
1. + scaling_params['scale'] *
gradient(dety, scaling_params['range_min'],
scaling_params['range_max']))
# Step error function
elif config['angular_scaling']['err_func_type'] == "step":
scaling_params = config['angular_scaling']['step']
break_points = list(
zip(scaling_params['transition_pos'],
scaling_params['transition_widths']))
self._bkg['bckgnd_new'] = self._bkg['bckgnd_new'] * (
1 + scaling_params['scale'] * step(dety, break_points))
else:
raise ValueError(
"Background angular scaling: unknown scaling function type '{:s}'"
.format(config['angular_scaling']['err_func_type']))
# ----------------------------------
# Recording the scaled Background
input_irf_file['BACKGROUND'].data['bgd'][0] = self._bkg[
'bckgnd_new'].transpose()
freqs = syfp.fftfreq(len(x), x[1] - x[0])
plt.semilogy(freqs, abs(fft))
plt.axvline(1, color="red", linestyle="--")
plt.show()
import scipy as sy
import scipy.fftpack as syfp
dt = 0.02071
t = np.linspace(0, 10, 1000) ## time at the same spacing of your data
freq = 10
u = np.sin(2 * np.pi * t * freq) ## Note freq=0.01 Hz
# Do FFT analysis of array
FFT = sy.fft.fft(u)
# Getting the related frequencies
freqs = syfp.fftfreq(len(u), t[1] -
t[0]) ## added dt, so x-axis is in meaningful units
# Create subplot windows and show plot
plt.subplot(211)
plt.plot(t, u)
plt.xlabel('Time')
plt.ylabel('Amplitude')
plt.subplot(212)
plt.plot(freqs, sy.log10(abs(FFT)),
'.') ## it's important to have the abs here
plt.xlim(-2 * freq, 2 * freq) ## zoom into see what's going on at the peak
plt.show()
def roomcomp(impresp, filter, target, ntaps, mixed_phase, opformat, trim,
nsthresh, noplot):
"""Primary function.
Determine a room compensation impulse response from a measured room impulse response.
"""
print("Loading impulse response")
# Read impulse response
Fs, data = wavfile.read(impresp)
data = norm(np.hstack(data))
if trim:
print("Removing leading silence")
for spos, sval in enumerate(data):
if abs(sval) > nsthresh:
lzs = max(spos - 1, 0)
print('Impulse starts at position ', spos, '/', len(data))
print('Trimming ',
float(lzs) / float(Fs), ' seconds of silence')
data = data[lzs:len(
data)] # remove everything before sample at spos
break
print("\nSample rate = ", Fs)
print("\nGenerating correction filter")
# Number of taps
if not ntaps:
ntaps = len(data)
# Logarithmic pole positioning
fplog = np.hstack(
(sp.logspace(sp.log10(20.), sp.log10(200.),
14.), sp.logspace(sp.log10(250.), sp.log10(20000.), 13.)))
plog = freqpoles(fplog, Fs)
# Preparing data
# making the measured response minumum-phase
cp, minresp = rceps(data)
# Impulse response
imp = np.zeros(len(data), dtype=np.float64)
imp[0] = 1.0
# Target
outf = []
if target is 'flat':
# Make the target output a bandpass filter
Bf, Af = sig.butter(4, 30 / (Fs / 2), 'high')
outf = sig.lfilter(Bf, Af, imp)
else:
# load target file
t = np.loadtxt(target)
frq = t[:, 0]
pwr = t[:, 1]
# calculate the FIR filter via windowing method
fir = sig.firwin2(5001,
frq,
np.power(10, pwr / 20.0),
fs=(frq[-1] * 2))
# Minimum phase, zero padding
cp, outf = rceps(np.append(fir, np.zeros(len(minresp) - len(fir))))
# Filter design
# Parallel filter design
(Bm, Am, FIR) = parfiltid(minresp, outf, plog)
# equalized loudspeaker response - filtering the
# measured transfer function by the parallel filter
equalizedresp = parfilt(Bm, Am, FIR, data)
# Equalizer impulse response - filtering a unit pulse
equalizer = norm(parfilt(Bm, Am, FIR, imp))
# Windowing with a half hanning window in time domain
han = np.hanning(ntaps * 2)[-ntaps:]
equalizer = han * equalizer[:ntaps]
"""
Mixed-phase compensation
Based on the paper "Mixed Time-Frequency approach for Multipoint
Room Rosponse Equalization," by A. Carini et al.
To use this feature, your Room Impulse Response should have all
the leading zeros removed.
"""
if mixed_phase is True:
# prototype function
hp = norm(np.real(equalizedresp))
# time integration of the human ear is ~24ms
# See "Measuring the mixing time in auditoria," by Defrance & Polack
hop_size = 0.024
samples = hop_size * Fs
bins = np.int(np.ceil(len(hp) / samples))
tmix = 0
# Kurtosis method
for b in range(bins):
start = np.int(b * samples)
end = np.int((b + 1) * samples)
k = kurtosis(hp[start:end])
if k <= 0:
tmix = b * hop_size
break
# truncate the prototype function
taps = np.int(tmix * Fs)
print("\nmixing time(secs) = ", tmix, "; taps = ", taps)
if taps > 0:
# Time reverse the array
h = hp[:taps][::-1]
# create all pass filter
phase = np.unwrap(np.angle(h))
h = np.exp(1j * phase)
# convert from db to linear
mixed = np.power(10, np.real(h) / 20.0)
# create filter's impulse response
mixed = np.real(ifft(mixed))
# convolve and window to desired length
equalizer = conv(equalizer, mixed)
equalizer = han * equalizer[:ntaps]
else:
print("zero taps; skipping mixed-phase computation")
if opformat in ('wav'):
wav_format = 'WAV'
subtype = 'PCM_16'
elif opformat in ('wav24'):
wav_format = 'WAV'
subtype = 'PCM_24'
elif opformat in ('wav32'):
wav_format = 'WAV'
subtype = 'PCM_32'
elif opformat in ('bin'):
wav_format = 'RAW'
subtype = 'FLOAT'
else:
print('Output format not recognized, no file generated.')
# Write data
wavwrite(filter, Fs, norm(np.real(equalizer)), wav_format, subtype)
print('\nOutput format is ' + opformat)
print('Output filter length =', len(equalizer), 'taps')
print('Output filter written to ' + filter)
print(
'\nUse sox to convert output .wav to raw 32 bit IEEE floating point if necessary,'
)
print('or to merge left and right channels into a stereo .wav')
print('\nExample (convert): sox leq48.wav -t f32 leq48.bin')
print(' (merge): sox -M le148.wav re48.wav output.wav\n')
# Plots
if not noplot:
data *= 500
# original loudspeaker-room response
plot(data, fs=Fs, avg='abs')
# 1/3 Octave smoothed
plot(data, fs=Fs, color='r', plots=True)
# equalizer transfer function
plot(0.75 * equalizer, fs=Fs, color='g')
# indicating pole frequencies
plt.vlines(fplog, -2, 2, color='k', linestyles='solid')
# equalized loudspeaker-room response
plot(equalizedresp * 0.01, fs=Fs, avg='abs')
# 1/3 Octave smoothed
plot(equalizedresp * 0.01, fs=Fs, color='r', plots=True)
# Add labels
# May need to reposition these based on input data
plt.text(325, 30, 'Unequalized loudspeaker-room response')
plt.text(100, -15, 'Equalizer transfer function')
plt.text(100, -21, '(Black lines: pole locations)')
plt.text(130, -70, 'Equalized loudspeaker-room response')
a = plt.gca()
a.set_xlim([20, 20000])
a.set_ylim([-80, 80])
plt.ylabel('Amplitude (dB)', color='b')
plt.xlabel('Frequency (Hz)')
plt.grid()
plt.show()
def _Sublimation_Pressure(cls, T):
"""Special sublimation pressure correlation"""
# Use decimal logarithm
P = 10**(-43.39/T+2.5*log10(T)+2.047)
return unidades.Pressure(P, "mmHg")
def main():
tb = pfb_top_block()
tstart = time.time()
tb.run()
tend = time.time()
print "Run time: %f" % (tend - tstart)
if 1:
fig1 = pylab.figure(1, figsize=(16,9))
fig2 = pylab.figure(2, figsize=(16,9))
Ns = 10000
Ne = 10000
fftlen = 8192
winfunc = scipy.blackman
fs = tb._fs
# Plot the input to the decimator
d = tb.snk_i.data()[Ns:Ns+Ne]
sp1_f = fig1.add_subplot(2, 1, 1)
X,freq = mlab.psd(d, NFFT=fftlen, noverlap=fftlen/4, Fs=fs,
window = lambda d: d*winfunc(fftlen),
scale_by_freq=True)
X_in = 10.0*scipy.log10(abs(fftpack.fftshift(X)))
f_in = scipy.arange(-fs/2.0, fs/2.0, fs/float(X_in.size))
p1_f = sp1_f.plot(f_in, X_in, "b")
sp1_f.set_xlim([min(f_in), max(f_in)+1])
sp1_f.set_ylim([-200.0, 50.0])
sp1_f.set_title("Input Signal", weight="bold")
sp1_f.set_xlabel("Frequency (Hz)")
sp1_f.set_ylabel("Power (dBW)")
Ts = 1.0/fs
Tmax = len(d)*Ts
t_in = scipy.arange(0, Tmax, Ts)
x_in = scipy.array(d)
sp1_t = fig1.add_subplot(2, 1, 2)
p1_t = sp1_t.plot(t_in, x_in.real, "b")
p1_t = sp1_t.plot(t_in, x_in.imag, "r")
sp1_t.set_ylim([-tb._decim*1.1, tb._decim*1.1])
sp1_t.set_xlabel("Time (s)")
sp1_t.set_ylabel("Amplitude")
# Plot the output of the decimator
fs_o = tb._fs / tb._decim
sp2_f = fig2.add_subplot(2, 1, 1)
d = tb.snk.data()[Ns:Ns+Ne]
X,freq = mlab.psd(d, NFFT=fftlen, noverlap=fftlen/4, Fs=fs_o,
window = lambda d: d*winfunc(fftlen),
scale_by_freq=True)
X_o = 10.0*scipy.log10(abs(fftpack.fftshift(X)))
f_o = scipy.arange(-fs_o/2.0, fs_o/2.0, fs_o/float(X_o.size))
p2_f = sp2_f.plot(f_o, X_o, "b")
sp2_f.set_xlim([min(f_o), max(f_o)+1])
sp2_f.set_ylim([-200.0, 50.0])
sp2_f.set_title("PFB Decimated Signal", weight="bold")
sp2_f.set_xlabel("Frequency (Hz)")
sp2_f.set_ylabel("Power (dBW)")
Ts_o = 1.0/fs_o
Tmax_o = len(d)*Ts_o
x_o = scipy.array(d)
t_o = scipy.arange(0, Tmax_o, Ts_o)
sp2_t = fig2.add_subplot(2, 1, 2)
p2_t = sp2_t.plot(t_o, x_o.real, "b-o")
p2_t = sp2_t.plot(t_o, x_o.imag, "r-o")
sp2_t.set_ylim([-2.5, 2.5])
sp2_t.set_xlabel("Time (s)")
sp2_t.set_ylabel("Amplitude")
pylab.show()
def cov_from_segments(gene,
seg_counts,
edge_counts,
edge_idx,
ax,
sample_idx=None,
log=False,
cmap_seg=None,
cmap_edg=None,
xlim=None,
grid=False,
order='C'):
"""This function takes a gene and its corresponding segment and edge counts to
produce a coverage overview plot."""
if sample_idx is None:
sample_idx = sp.arange(seg_counts.shape[1])
norm = plt.Normalize(0, sample_idx.shape[0])
if cmap_seg is None:
cmap_seg = plt.get_cmap('jet')
if cmap_edg is None:
cmap_edg = plt.get_cmap('jet')
### iterate over samples
for ii, i in enumerate(sample_idx):
### collect count information and add segment patches
for j in range(gene.segmentgraph.segments.shape[1]):
s = gene.segmentgraph.segments[:, j]
if log:
counts = sp.log10(seg_counts[j, i] + 1)
else:
counts = seg_counts[j, i]
#ax.add_patch(patches.Rectangle((s[0], 0), s[1] - s[0], counts, fill=cmap_seg(norm(ii)),
# edgecolor='none', alpha=0.5))
ax.plot(s, [counts, counts],
'-',
color=cmap_seg(norm(ii)),
linewidth=2)
for j in range(edge_idx.shape[0]):
[s, t] = sp.unravel_index(edge_idx[j],
gene.segmentgraph.seg_edges.shape,
order=order)
if log:
counts = sp.log10(edge_counts[j, i] + 1)
else:
counts = edge_counts[j, i]
add_intron_patch2(ax,
gene.segmentgraph.segments[1, s],
gene.segmentgraph.segments[0, t],
counts,
color=cmap_edg(norm(ii)))
if xlim is not None:
ax.set_xlim(xlim)
### draw grid
if grid:
ax.grid(b=True,
which='major',
linestyle='--',
linewidth=0.2,
color='#222222')
ax.xaxis.grid(False)
ax.set_ylim([0, ax.get_ylim()[1]])
# construction du vecteur fréquence
f = dF * arange(N)
# définition de la matrice M
n = arange(N)
k = n.reshape((N, 1))
M = exp(-2j * pi * k * n / N)
# calcul de la transformée de Fourier discrète par simple produit
# matriciel X = M.s
tdebut = time.time()
X = dot(M, s)
print 'Temps de calcul (s) : ', time.time() - tdebut
# calcul du spectre SP normalisé
FacteurNorm = 2.0 / N
SP = FacteurNorm * abs(X)
# pour affichage en dB
SP1 = 20 * log10(SP / SP.max())
# affichage des résultats
fig = plt.figure(figsize=(14, 8))
# affichage du spectre
plt.xlabel('frequence', fontsize=16)
plt.ylabel('amplitude', fontsize=16)
plt.plot(f, SP1)
plt.show()
'Path to file to previously produced picca_delta.py file to correct for multiplicative errors in the flux calibration'
)
parser.add_argument(
'--ivar-calib',
type=str,
default=None,
required=False,
help=
'Path to previously produced picca_delta.py file to correct for multiplicative errors in the pipeline inverse variance calibration'
)
args = parser.parse_args()
### forest args
forest.lmin = sp.log10(args.lambda_min)
forest.lmax = sp.log10(args.lambda_max)
forest.lmin_rest = sp.log10(args.lambda_rest_min)
forest.lmax_rest = sp.log10(args.lambda_rest_max)
forest.rebin = args.rebin
forest.dll = args.rebin * 1e-4
forest.dla_mask = args.dla_mask
### Get Healpy pixel of the given QSO
objs = {}
ra, dec, zqso, thid, plate, mjd, fid = io.read_drq(args.drq,
0.,
1000.,
keep_bal=True)
cut = (plate == args.plate) & (mjd == args.mjd) & (fid == args.fiberid)
if cut.sum() == 0:
def cov_from_bam(chrm,
start,
stop,
files,
subsample=0,
verbose=False,
bins=None,
log=False,
ax=None,
ymax=0,
outfile=None,
frm='pdf',
xlim=None,
title=None,
xoff=None,
yoff=None,
intron_cov=False,
intron_cnt=False,
marker_pos=None,
col_idx=None,
color_cov='blue',
color_intron_cov='red',
color_intron_edge='green',
grid=False,
strand=None,
highlight=None,
highlight_color='magenta',
highlight_label=None,
min_intron_cnt=0,
return_legend_handle=False,
label=None):
"""This function takes a list of bam files and a set of coordinates (chrm, start, stop), to
plot a coverage overview of that files in that region."""
### subsampling
if subsample > 0 and len(files) > subsample:
npr.seed(23)
files = sp.array(files)
files = npr.choice(files, subsample)
### augment chromosome name
#chr_name = 'chr%s' % chrm
chr_name = chrm
(counts, intron_counts, intron_list) = _get_counts(chr_name,
start,
stop,
files,
intron_cov,
intron_cnt,
verbose,
collapsed=True)
### get mean counts over all bam files
counts /= len(files)
if intron_cov:
intron_counts /= len(files)
if intron_cnt:
for intron in intron_list:
intron_list[intron] = math.ceil(intron_list[intron] /
float(len(files)))
if min_intron_cnt > 0:
intron_list = dict([(x, intron_list[x]) for x in intron_list
if intron_list[x] >= min_intron_cnt])
if col_idx is not None:
counts = counts[col_idx]
if intron_cov:
intron_counts = intron_counts[col_idx]
if intron_cnt:
print >> sys.stderr, 'ERROR: column subsetting is currently not implemented for intron edges'
sys.exit(1)
### bin counts according to options
if bins is None:
bins = counts.shape[0]
bin_counts = counts
bin_intron_counts = intron_counts
if col_idx is not None:
counts_x = sp.arange(col_idx.shape[0])
else:
counts_x = range(start, stop + 1)
else:
if verbose:
print >> sys.stdout, '... binning counts ...'
bin_counts = sp.zeros((bins, ))
bin_intron_counts = sp.zeros((bins, ))
binsize = int(sp.ceil(float(counts.shape[0]) / bins))
for ii, i in enumerate(xrange(0, counts.shape[0], binsize)):
bin_counts[ii] = sp.sum(
counts[i:min(i + binsize, counts.shape[0] - 1)]) / binsize
if intron_cov:
bin_intron_counts[ii] = sp.sum(
intron_counts[i:min(i + binsize, intron_counts.shape[0] -
1)]) / binsize
if col_idx is not None:
counts_x = sp.linspace(0, col_idx.shape[0], num=bins)
else:
counts_x = sp.linspace(start, stop, num=bins)
### use log if chosen
if log:
bin_counts = sp.log10(bin_counts + 1)
bin_intron_counts = sp.log10(bin_intron_counts + 1)
if intron_cnt:
for intron in intron_list:
if intron_list[intron] > 0:
intron_list[intron] = sp.log10(intron_list[intron] + 1)
if ax is None:
fig = plt.figure(figsize=(10, 4))
ax = fig.add_subplot(111)
if intron_cov:
ax.fill_between(counts_x,
bin_intron_counts,
facecolor=color_intron_cov,
edgecolor='none',
alpha=0.5)
ax.fill_between(counts_x,
bin_counts,
facecolor=color_cov,
edgecolor='none',
alpha=0.5)
#ax.set_xticklabels([str(int(x)) for x in sp.linspace(start, stop, num = len(ax.get_xticklabels()))])
ax.set_xlabel('Position on contig %s' % chrm)
### draw strand
if strand == '+':
ax.arrow(0.05,
0.9,
0.2,
0,
head_width=0.05,
head_length=0.02,
fc='#cccccc',
ec='#cccccc',
transform=ax.transAxes)
elif strand == '-':
ax.arrow(0.25,
0.9,
-0.2,
0,
head_width=0.05,
head_length=0.02,
fc='#cccccc',
ec='#cccccc',
transform=ax.transAxes)
### draw grid
if grid:
ax.grid(b=True,
which='major',
linestyle='--',
linewidth=0.2,
color='#222222')
ax.xaxis.grid(False)
if marker_pos is not None:
ax.plot(0, marker_pos, 'or')
if log:
ax.set_ylabel('Read Coverage (log10)')
else:
ax.set_ylabel('Read Coverage')
if ymax > 0:
ax.set_ylim([0, ymax])
if highlight is not None:
highlight_x(ax,
highlight,
highlight_color=highlight_color,
label=highlight_label)
if xlim is not None:
ax.set_xlim(xlim)
ylim = ax.get_ylim()
ax.set_ylim([0, ylim[1]])
if title is not None:
ax.set_title(title)
if xoff:
ax.axes.get_xaxis().set_visible(False)
if yoff:
ax.axes.get_yaxis().set_visible(False)
if intron_cnt:
for intron in intron_list:
add_intron_patch2(ax,
start + intron[0],
start + intron[1] + intron[0],
intron_list[intron],
color=color_intron_edge)
if outfile is not None:
plt.savefig(outfile, dpi=1200, format=frm)
if return_legend_handle:
if label is not None:
return patches.Patch(color=color_cov, alpha=0.5, label=label)
else:
return patches.Patch(color=color_cov,
alpha=0.5,
label='Expression')
def plot_manhattan(ax,
df,
pv_thr=None,
colors=None,
offset=None,
callback=None):
"""
Utility function to make manhattan plot
Parameters
----------
ax : pyplot plot
subplot
df : pandas.DataFrame
pandas DataFrame with chrom, pos and pv
colors : list
colors to use in the manhattan plot
offset : float
offset between in chromosome expressed as fraction of the
length of the longest chromosome (default is 0.2)
callback : function
callback function that takes as input df
Examples
--------
.. doctest::
>>> from matplotlib import pyplot as plt
>>> from limix_lmm.plot import plot_manhattan
>>> import scipy as sp
>>> import pandas as pd
>>> n_chroms = 5
>>> n_snps_per_chrom = 10000
>>> chrom = sp.kron(sp.arange(1, n_chroms + 1), sp.ones(n_snps_per_chrom))
>>> pos = sp.kron(sp.ones(n_chroms), sp.arange(n_snps_per_chrom))
>>> pv = sp.rand(n_chroms * n_snps_per_chrom)
>>> df = pd.DataFrame({'chrom': chrom, 'pos': pos, 'pv': pv})
>>>
>>> ax = plt.subplot(111)
>>> plot_manhattan(ax, df)
"""
if colors is None:
colors = ["k", "Gray"]
if offset is None:
offset = 0.2
dx = offset * df["pos"].values.max()
_x = 0
xticks = []
for chrom_i in sp.unique(df["chrom"].values):
_df = df[df["chrom"] == chrom_i]
if chrom_i % 2 == 0:
color = colors[0]
else:
color = colors[1]
ax.plot(_df["pos"] + _x, -sp.log10(_df["pv"]), ".", color=color)
if callback is not None:
callback(_df)
xticks.append(_x + 0.5 * _df["pos"].values.max())
_x += _df["pos"].values.max() + dx
ax.set_xticks(xticks)
ax.set_xticklabels(sp.unique(df["chrom"].values))
def snr_est_m2m4(signal):
M2 = scipy.mean(abs(signal)**2)
M4 = scipy.mean(abs(signal)**4)
snr_rat = scipy.sqrt(2 * M2 * M2 - M4) / (M2 -
scipy.sqrt(2 * M2 * M2 - M4))
return 10.0 * scipy.log10(snr_rat), snr_rat
def main():
gr_estimators = {
"simple": digital.SNR_EST_SIMPLE,
"skew": digital.SNR_EST_SKEW,
"m2m4": digital.SNR_EST_M2M4,
"svr": digital.SNR_EST_SVR
}
py_estimators = {
"simple": snr_est_simple,
"skew": snr_est_skew,
"m2m4": snr_est_m2m4,
"svr": snr_est_svr
}
parser = OptionParser(option_class=eng_option, conflict_handler="resolve")
parser.add_option(
"-N",
"--nsamples",
type="int",
default=10000,
help="Set the number of samples to process [default=%default]")
parser.add_option("",
"--snr-min",
type="float",
default=-5,
help="Minimum SNR [default=%default]")
parser.add_option("",
"--snr-max",
type="float",
default=20,
help="Maximum SNR [default=%default]")
parser.add_option("",
"--snr-step",
type="float",
default=0.5,
help="SNR step amount [default=%default]")
parser.add_option("-t",
"--type",
type="choice",
choices=gr_estimators.keys(),
default="simple",
help="Estimator type {0} [default=%default]".format(
gr_estimators.keys()))
(options, args) = parser.parse_args()
N = options.nsamples
xx = scipy.random.randn(N)
xy = scipy.random.randn(N)
bits = 2 * scipy.complex64(scipy.random.randint(0, 2, N)) - 1
#bits =(2*scipy.complex64(scipy.random.randint(0, 2, N)) - 1) + \
# 1j*(2*scipy.complex64(scipy.random.randint(0, 2, N)) - 1)
snr_known = list()
snr_python = list()
snr_gr = list()
# when to issue an SNR tag; can be ignored in this example.
ntag = 10000
n_cpx = xx + 1j * xy
py_est = py_estimators[options.type]
gr_est = gr_estimators[options.type]
SNR_min = options.snr_min
SNR_max = options.snr_max
SNR_step = options.snr_step
SNR_dB = scipy.arange(SNR_min, SNR_max + SNR_step, SNR_step)
for snr in SNR_dB:
SNR = 10.0**(snr / 10.0)
scale = scipy.sqrt(2 * SNR)
yy = bits + n_cpx / scale
print "SNR: ", snr
Sknown = scipy.mean(yy**2)
Nknown = scipy.var(n_cpx / scale)
snr0 = Sknown / Nknown
snr0dB = 10.0 * scipy.log10(snr0)
snr_known.append(float(snr0dB))
snrdB, snr = py_est(yy)
snr_python.append(snrdB)
gr_src = blocks.vector_source_c(bits.tolist(), False)
gr_snr = digital.mpsk_snr_est_cc(gr_est, ntag, 0.001)
gr_chn = channels.channel_model(1.0 / scale)
gr_snk = blocks.null_sink(gr.sizeof_gr_complex)
tb = gr.top_block()
tb.connect(gr_src, gr_chn, gr_snr, gr_snk)
tb.run()
snr_gr.append(gr_snr.snr())
f1 = pylab.figure(1)
s1 = f1.add_subplot(1, 1, 1)
s1.plot(SNR_dB, snr_known, "k-o", linewidth=2, label="Known")
s1.plot(SNR_dB, snr_python, "b-o", linewidth=2, label="Python")
s1.plot(SNR_dB, snr_gr, "g-o", linewidth=2, label="GNU Radio")
s1.grid(True)
s1.set_title('SNR Estimators')
s1.set_xlabel('SNR (dB)')
s1.set_ylabel('Estimated SNR')
s1.legend()
f2 = pylab.figure(2)
s2 = f2.add_subplot(1, 1, 1)
s2.plot(yy.real, yy.imag, 'o')
pylab.show()
def _update_canvas(self):
"""
Update the figure when the user changes and input value.
:return:
"""
# Get the parameters from the form
prf_stagger = float(self.prf_stagger.text())
# Get the PRF type
prf_type = self.prf_type.currentText()
# Set up the normalized frequency space
frequency = linspace(0, 4, 1000)
# Clear the axes for the updated plot
self.axes1.clear()
# Calculate response based on PRF type
if prf_type == 'Single':
response = countermeasures.delay_line(frequency) / 4.0
# Display the results
self.axes1.plot(frequency, 10 * log10(response + finfo(float).eps),
'')
elif prf_type == 'Stagger':
response_prf1 = countermeasures.delay_line(frequency) / 4.0
response_prf2 = countermeasures.delay_line(
prf_stagger * frequency) / 4.0
response = 0.5 * (response_prf1 + response_prf2)
# Display the results
self.axes1.plot(frequency,
10 * log10(response_prf1 + finfo(float).eps),
'',
label='PRF 1')
self.axes1.plot(frequency,
10 * log10(response_prf2 + finfo(float).eps),
'--',
label='PRF 2')
self.axes1.plot(frequency,
10 * log10(response + finfo(float).eps),
':',
label='PRF Staggered')
# Place the legend
self.axes1.legend(loc='lower left', prop={'size': 10})
# Set the plot title and labels
self.axes1.set_title('Delay Line Response', size=14)
self.axes1.set_xlabel('Normalized Frequency (f / PRF)', size=12)
self.axes1.set_ylabel('Amplitude (dB)', size=12)
# Set the y-axis lim
self.axes1.set_ylim([-30, 1])
# Turn on the grid
self.axes1.grid(linestyle=':', linewidth=0.5)
# Set the tick label size
self.axes1.tick_params(labelsize=12)
# Update the canvas
self.my_canvas.draw()
def LogVTKIntersectiondata(self):
print("Logging VTK intersection")
# log for vtk intersect
for r in range(self.vtkintersection.shape[0]):
self.vtkintersection[r, :] = scipy.log10(self.intersection[r, :])
def LogVTKdata(self):
print("Logging VTK coords")
# log for vtk
for r in range(self.vtkcoords.shape[0]):
self.vtkcoords[r, :self.coords.shape[1]] = scipy.log10(
self.vtkcoords[r, :self.coords.shape[1]])
def cont_fit(self):
lmax = forest.lmax_rest + sp.log10(1 + self.zqso)
lmin = forest.lmin_rest + sp.log10(1 + self.zqso)
try:
mc = forest.mean_cont(self.ll - sp.log10(1 + self.zqso))
except ValueError:
raise Exception
if not self.T_dla is None:
mc *= self.T_dla
var_lss = forest.var_lss(self.ll)
eta = forest.eta(self.ll)
fudge = forest.fudge(self.ll)
def model(p0, p1):
line = p1 * (self.ll - lmin) / (lmax - lmin) + p0
return line * mc
def chi2(p0, p1):
m = model(p0, p1)
var_pipe = 1. / self.iv / m**2
## prep_del.variance is the variance of delta
## we want here the we = ivar(flux)
var_tot = variance(var_pipe, eta, var_lss, fudge)
we = 1 / m**2 / var_tot
# force we=1 when use-constant-weight
# TODO: make this condition clearer, maybe pass an option
# use_constant_weights?
if (eta == 0).all():
we = sp.ones(len(we))
v = (self.fl - m)**2 * we
return v.sum() - sp.log(we).sum()
p0 = (self.fl * self.iv).sum() / self.iv.sum()
p1 = 0
mig = iminuit.Minuit(chi2,
p0=p0,
p1=p1,
error_p0=p0 / 2.,
error_p1=p0 / 2.,
errordef=1.,
print_level=0,
fix_p1=(self.order == 0))
fmin, _ = mig.migrad()
self.co = model(mig.values["p0"], mig.values["p1"])
self.p0 = mig.values["p0"]
self.p1 = mig.values["p1"]
self.bad_cont = None
if not fmin.is_valid:
self.bad_cont = "minuit didn't converge"
if sp.any(self.co <= 0):
self.bad_cont = "negative continuum"
## if the continuum is negative, then set it to a very small number
## so that this forest is ignored
if self.bad_cont is not None:
self.co = self.co * 0 + 1e-10
self.p0 = 0.
self.p1 = 0.
N2 = inversion(input_inten_wls, pump)
N1 = 1 - N2
# P_p and P_s
totalSignal = []
# For example and illustration, solve power equations for 1 m
for k in range(0, len(wl_grid)):
p_init = input_inten_wls[k]
solution = odeint(power_signal, p_init, z, args=(N1, N2, k))
totalSignal.append(solution[:, 0])
pump_solution = odeint(power_pump, pump, z, args=(N1, N2))
pump_absorption = -10 * sp.log10(
max(pump_solution[1000]) / max(pump_solution[0])
) #check to see if pump absorption matches Data Sheet
print(pump_absorption)
signalPlotting = sp.asarray(totalSignal)
plt.figure()
plt.plot(z, solution[:, 0], '--', label='signal')
plt.legend(loc='best')
plt.xlabel('z')
plt.ylabel('power (W)')
plt.grid()
plt.figure()
plt.plot(z, pump_solution[:, 0], '--', label='pump')
def map_phenotype(p_i, phed, snps_data_file, mapping_method, trans_method, p_dict):
phenotype_name = phed.getPhenotypeName(p_i)
phen_is_binary = phed.isBinary(p_i)
file_prefix = _get_file_prefix_(p_dict['run_id'], p_i, phed.getPhenotypeName(p_i),
mapping_method, trans_method, p_dict['remove_outliers'])
result_name = "%s_%s_%s" % (phenotype_name, mapping_method, trans_method)
res = None
sd = dataParsers.parse_snp_data(snps_data_file , format=p_dict['data_format'], filter=p_dict['debug_filter'])
num_outliers = gwa.prepare_data(sd, phed, p_i, trans_method, p_dict['remove_outliers'])
if p_dict['remove_outliers']:
assert num_outliers != 0, "No outliers were removed, so it makes no sense to go on and perform GWA."
phen_vals = phed.getPhenVals(p_i)
snps = sd.getSnps()
if mapping_method in ['emmax']:
#Load genotype file (in binary format)
sys.stdout.write("Retrieving the Kinship matrix K.\n")
sys.stdout.flush()
k_file = env['data_dir'] + "kinship_matrix_cm" + str(p_dict['call_method_id']) + ".pickled"
kinship_file = p_dict['kinship_file']
if not kinship_file and os.path.isfile(k_file): #Check if corresponding call_method_file is available
kinship_file = k_file
if kinship_file: #Kinship file was somehow supplied..
print 'Loading supplied kinship'
k = lm.load_kinship_from_file(kinship_file, sd.accessions)
else:
print "No kinship file was found. Generating kinship file:", k_file
sd = dataParsers.parse_snp_data(snps_data_file , format=p_dict['data_format'])
snps = sd.getSnps()
k_accessions = sd.accessions[:]
if p_dict['debug_filter']:
import random
snps = random.sample(snps, int(p_dict['debug_filter'] * len(snps)))
k = lm.calc_kinship(snps)
f = open(k_file, 'w')
cPickle.dump([k, sd.accessions], f)
f.close()
num_outliers = gwa.prepare_data(sd, phed, p_i, trans_method, p_dict['remove_outliers'])
k = lm.filter_k_for_accessions(k, k_accessions, sd.accessions)
sys.stdout.flush()
sys.stdout.write("Done!\n")
if p_dict['remove_outliers']:
assert num_outliers != 0, "No outliers were removed, so it makes no sense to go on and perform GWA."
#Check whether result already exists.
if p_dict['use_existing_results']:
print "\nChecking for existing results."
result_file = file_prefix + ".pvals"
if os.path.isfile(result_file):
res = gwaResults.Result(result_file=result_file, name=result_name, snps=snps)
pvals = True
else:
result_file = file_prefix + ".scores"
if os.path.isfile(result_file):
res = gwaResults.Result(result_file=result_file, name=result_name, snps=snps)
pvals = False
if res:
print "Found existing results.. (%s)" % (result_file)
sys.stdout.flush()
if not res: #If results weren't found in a file... then do GWA.
sys.stdout.write("Finished loading and handling data!\n")
print "FIRST STEP: Applying %s to data. " % (mapping_method)
sys.stdout.flush()
kwargs = {}
additional_columns = []
if mapping_method in ['emmax']:
res = lm.emmax(snps, phen_vals, k)
elif mapping_method in ['lm']:
res = lm.linear_model(snps, phen_vals)
else:
print "Mapping method", mapping_method, 'was not found.'
sys.exit(2)
if mapping_method in ['lm', 'emmax']:
kwargs['genotype_var_perc'] = res['var_perc']
betas = map(list, zip(*res['betas']))
kwargs['beta0'] = betas[0]
kwargs['beta1'] = betas[1]
additional_columns.append('genotype_var_perc')
additional_columns.append('beta0')
additional_columns.append('beta1')
pvals = res['ps']
sys.stdout.write("Done!\n")
sys.stdout.flush()
kwargs['correlations'] = calc_correlations(snps, phen_vals)
additional_columns.append('correlations')
res = gwaResults.Result(scores=pvals, snps_data=sd, name=result_name, **kwargs)
if mapping_method in ["emmax", 'lm']:
result_file = file_prefix + ".pvals"
else:
result_file = file_prefix + ".scores"
res.write_to_file(result_file, additional_columns)
print "Generating a GW plot."
sys.stdout.flush()
png_file = file_prefix + "_gwa_plot.png"
#png_file_max30 = file_prefix+"_gwa_plot_max30.png"
if mapping_method in ['lm', "emmax"]:
res.neg_log_trans()
if mapping_method in ["kw", "ft"]:# or p_dict['data_format'] != 'binary':
#res.plot_manhattan(png_file=png_file_max30,percentile=90,type="pvals",ylab="$-$log$_{10}(p)$",
# plot_bonferroni=True,max_score=30)
res.plot_manhattan(png_file=png_file, percentile=90, type="pvals", ylab="$-$log$_{10}(p)$",
plot_bonferroni=True)
else:
if res.filter_attr("mafs", p_dict['mac_threshold']) > 0:
#res.plot_manhattan(png_file=png_file_max30,percentile=90,type="pvals",ylab="$-$log$_{10}(p)$",
# plot_bonferroni=True,max_score=30)
res.plot_manhattan(png_file=png_file, percentile=90, type="pvals", ylab="$-$log$_{10}(p)$",
plot_bonferroni=True)
else:
pass
print "plotting histogram"
hist_file_prefix = _get_file_prefix_(p_dict['run_id'], p_i, phenotype_name, trans_method, p_dict['remove_outliers'])
hist_png_file = hist_file_prefix + "_hist.png"
phed.plot_histogram(p_i, pngFile=hist_png_file)
else:
res.neg_log_trans()
assert res.filter_attr("mafs", p_dict['mac_threshold']), 'All SNPs have MAC smaller than threshold'
print "SECOND STEP:"
res.filter_top_snps(p_dict['second_step_number'])
snps = res.snps
positions = res.positions
chromosomes = res.chromosomes
#Checking res_file exists
file_prefix = _get_file_prefix_(p_dict['run_id'], p_i, phed.getPhenotypeName(p_i),
mapping_method, trans_method, p_dict['remove_outliers'], p_dict['second_step_number'])
res_file = file_prefix + '_res.cpickled'
if p_dict['use_existing_results'] and os.path.isfile(res_file):
print 'Found existing results for the second step... loading.'
f = open(res_file, 'rb')
second_res = cPickle.load(f)
f.close()
else:
if mapping_method == 'lm':
second_res = lm.linear_model_two_snps(snps, phen_vals)
if mapping_method == 'emmax':
second_res = lm.emmax_two_snps(snps, phen_vals, k)
#Pickling results..
print 'Saving results as pickled file:', res_file
f = open(res_file, 'wb')
cPickle.dump(second_res, f, protocol=2)
f.close()
#Plotting second step plots:
score_array = -sp.log10(second_res['ps'])
p3_score_array = -sp.log10(second_res['p3_ps'])
p4_score_array = -sp.log10(second_res['p4_ps'])
import plotResults as pr
pr.plot_snp_pair_result(chromosomes, positions, score_array, file_prefix + '_scatter')
pr.plot_snp_pair_result(chromosomes, positions, p3_score_array, file_prefix + '_p3_scatter')
pr.plot_snp_pair_result(chromosomes, positions, p4_score_array, file_prefix + '_p4_scatter')
if p_dict['region_plots']:
import regionPlotter as rp
regions_results = res.get_top_region_results(p_dict['region_plots'])
plotter = rp.RegionPlotter()
print "Starting region plots..."
for reg_res in regions_results:
chromosome = reg_res.chromosomes[0]
caption = phenotype_name + "_c" + str(chromosome) + "_" + mapping_method
png_file = file_prefix + "_reg_plot_c" + str(chromosome) + "_s" + str(reg_res.positions[0]) \
+ "_e" + str(reg_res.positions[-1]) + ".png"
tair_file = file_prefix + "_reg_plot_c" + str(chromosome) + "_s" + str(reg_res.positions[0]) \
+ "_e" + str(reg_res.positions[-1]) + "_tair_info.txt"
plotter.plot_small_result([reg_res], png_file=png_file, highlight_gene_ids=tair_ids,
caption=caption, tair_file=tair_file)
#Plot Box-plot
png_file = file_prefix + "_reg_plot_c" + str(chromosome) + "_s" + str(reg_res.positions[0]) \
+ "_e" + str(reg_res.positions[-1]) + "_box_plot.png"
(marker, score, chromosome, pos) = reg_res.get_max_snp()
marker_accessions = sd.accessions
phed.plot_marker_box_plot(p_i, marker=marker, marker_accessions=marker_accessions, \
png_file=png_file, title="c" + str(chromosome) + "_p" + str(pos), \
marker_score=score, marker_missing_val=sd.missing_val)
