'Data' object has no attribute 'write_step'

[Sefa76]

Hello everyone,

i am new here and i dont have much expieriance with the code and ive run into an error that i dont know how to resolve. When i try to run the code with my likelyhood i get this errormassage:

Traceback (most recent call last): File "montepython/MontePython.py", line 40, in sys.exit(run()) File "/mnt/c/Users/sefa7/Desktop/Bachlorarbeit/Code/montepython_public-3.4/montepython/run.py", line 45, in run sampler.run(cosmo, data, command_line) File "/mnt/c/Users/sefa7/Desktop/Bachlorarbeit/Code/montepython_public-3.4/montepython/sampler.py", line 46, in run mcmc.chain(cosmo, data, command_line) File "/mnt/c/Users/sefa7/Desktop/Bachlorarbeit/Code/montepython_public-3.4/montepython/mcmc.py", line 841, in chain if acc % data.write_step == 0: AttributeError: 'Data' object has no attribute 'write_step'

I dont really understand why this shows up, because i am able to run the code with the provided input/likelihoodmethod example.param so the error should be on my side, but there is no refference in the error to my code. Thank you for your time

with kind regards, Sefa

[shan1525]

What param file you are using? How are you running your code? Which likelihood code you are using? Make an elaborate question so that we can figure out what's happening. I wanted to help but I couldn't understand the question properly.

Best, Shan.

[Sefa76]

Hello Shan

im running the code with time python3 montepython/MontePython.py run -p input/hm_et.param -o chains/hm_et -N 10 i could run the example.param just fine with that

my .param is very simple

data.experiments=['fake_et'] data.parameters['H0'] = [68,60,80,3,1,'cosmo'] the 'fake_et' is a likelihood method i wrote myself, its very simple it just compares values for luminocity distance with values from a .csv

this is the init , the .data only has the path to the .csv ` import os from montepython.likelihood_class import Likelihood import pandas as pd import numpy as np

class fake_et(Likelihood):

    def __init__(self, path, data, command_line):
            #super
            Likelihood.__init__(self, path, data, command_line)
            #define array for values of z and dl and sigma
            dataframe = pd.read_csv(self.datapath).to_numpy()
            self.z = dataframe[0]
            self.d = dataframe[1]
            self.sigma = dataframe[2]

    def loglkl(self, cosmo, data):
            #cosmological model for dl(z)
            d=[]
            for i in range(len(self.z)):
                    d.append(cosmo.luminosity_distance(self.z[i]))
            d=np.array(d)
            #calculate chiq
            loglkl = -0.5 * (d - self.d) ** 2 / (self.sigma ** 2)
            return np.sum(loglkl) 

` thank you for your time with kind regards Sefa

[lesgourg]

Hi @Sefa, I think that I see the point. We forgot to ever define a default value of the variable data.write_step in montepython, because of the habit to end up all .param files with the usual lines (like in all the examples of input files):

#------ Mcmc parameters ----
# Number of steps taken, by default (overwritten by the -N command)
data.N=10
# Number of accepted steps before writing to file the chain. Larger means less
# access to disc, but this is not so much time consuming.
data.write_step=5

You did not have these lines at the end of your .param file, so the code complained. You just need to add them. But I agree that, on our side, we should implement a default value of write_step in the code, such that these lines could in principle be skipped in the .param !

[Sefa76]

Hello everyone

thanks the code seems to be working now

with kind regards Sefa

[brinckmann]

Added default value for data.write_step in version 3.5 in case it's not specified in the param file.

Best, Thejs