使用emcee时函数定义的语法错误

Question

我正在尝试使用emcee模块来重新创建一个发行版。下面是我的代码:

freq,asd = np.loadtxt('noise.csv',delimiter=',',unpack=True)

psd = asd**2
SNRth = 4.5
d = 600
dm = 0.9
#interpolate!
S = interpolate.interp1d(freq,psd)
def SNR2(chirp,f): 
    return 5*np.pi**(-4/3)*chirp**(5/3)/(96*d**2)*integrate.quad(lambda f: f**(-7/3)/S(f), 20, 1500, limit=1000)[0]
def mp(SNR2, x):
    return integrate.quad(lambda x: np.exp(-0.5*(x+SNR2))*special.iv(0,np.sqrt(x*SNR2)),0,SNRth**2)
def pp(SNR2, x):
    return 0.5*np.exp(-0.5*(SNRth**2+SNR2))*special.iv(0,np.sqrt(x*SNR2))
def mm(SNR2, x):
    return 1-np.exp(-SNRth**2/2)
def pm(SNR2, x):
    return (1/(2*dm**2))*np.exp(-SNRth**2/2)
#global variables for the sampler
nwalkers = 500
ndim = 2 
nmcmc = 500 #number of mcmc steps...make it a factor of 2
#variables for the population
mupop = 0.4 #mean
sigpop = 0.1 #standard deviation
#variables for individual events
Ntrig = 37 #number of individual events.
Nnoise = 11
Nsamp = 100 #number of sample points for each individual event
stdmin = 1e-50 #smallest possible standard deviation
stdmax = 0.2 #largest possible standard deviation
#enforce uniform prior on the mean and the variance between (0,1)
def lnprior(theta):
    mu, sig = theta
    if 0 < mu < 1 and 0 < sig < 1:
        return 0
    return -np.inf
#log likelihood for the population
#seps are single event posteriors
def lnlike(theta,*seps):
    mu, sig= theta
    seps = np.asarray(seps).reshape((Ntrig,Nsamp)) #TODO: find a better way to pass a variable number of arguments to emcee
return np.prod(pp(SNR2(chirp,f),seps)+pm(SNR2(chirp,f),seps))*np.prod(mp(SNR2(chirp,f),seps[:Nnoise])+mm(SNR2(chirp,f),seps[:Nnoise])
def lnpop(theta,*seps):
    lp = lnprior(theta)
    if not np.isfinite(lp):
        return -np.inf
    seps = np.asarray(seps).reshape((Ntrig,Nsamp)) #TODO: find a better way to pass a variable number of arguments to emcee
    return lp + lnlike(theta,seps)

它给了我一个语法错误，当我def inpop。知道是什么问题吗?我认为这是定义函数的正确方法.....

Answer 1

本行缺少结束的):

return np.prod(pp(SNR2(chirp,f),seps)+pm(SNR2(chirp,f),seps))*np.prod(mp(SNR2(chirp,f),seps[:Nnoise])+mm(SNR2(chirp,f),seps[:Nnoise])) # <-- missing here

通常语法错误来自于回溯中显示的前一行。