David R. Pugh
David R. Pugh
We definitely do not want people to feel blocked from using QuantEcon because their favorite CompEcon MATLAB routine appears to be unavailable. 1. Have RA's to do a first-pass port...
I have typically followed Robert Kerns suggestion on [SO](http://stackoverflow.com/questions/5836335/consistenly-create-same-random-numpy-array) and created an instance of `numpy.random.RandomState` which returns a numpy-aware PRNG... ``` python import numpy as np prng = np.random.RandomState(42) ```...
@albop From the discussion in the SO post that I linked to I came away with the impression that using `np.random.RandomState` would give you more consistently reproducible randomness across platforms/computers....
@oyamad Not sure what you mean "in the test function"? One possibility would be to create the `prng` instance externally and then access its methods from within the function: ```...
@oyamad Where is the function `test_mc_sample_path_seed`? I can't find it (at least not quickly)...
Given the way that `mc_sample_path` is currently [written](https://github.com/QuantEcon/QuantEcon.py/blob/master/quantecon/mc_tools.py) you will not be able to use `np.random.RandomState(seed)` as suggested in the SO post and in this thread. I wonder if this...
I think if we solve #159 we get this for free. Not sure that we can pass `rv_frozen` objects though...
@jstac I think it might be possible to use Numba to do this. See the very bottom of issue #159 for an example of how to pass jitted functions as...
Sounds good to me.
@dwr-psandhu sounds like the issue is with the `conda env export` command. I have been creating Conda environments from environment files that I create by hand for years and never...