dgellerup / laminar Goto Github PK

I love this package! Thanks for all the work developing it thus far.

I added some basic and more complex tests of iter_flow() on a forked repo

1. Running them, one of the things I noticed is that varying the number of cores and the length of the iterable could cause the tests to fail which had otherwised passed. It seems that cause of the failures is the order in which the returned results dictionary returns keys. It does not always return in matching order to the iterable. This can make it hard to reassemble the results if order matters as it does in the tests, written.

I was running in Python 3.6.8 |Anaconda, Inc.| (default, Dec 29 2018, 19:04:46)

pytest test/test_iterflow_complex.py -v

>>>  [print(k) for k in r.keys()]
-------------------------------------------- Captured stdout call --------------------------------------------
data[160-319]
data[0-159] ## <<-------------- **out of order**
data[320-479]
data[480-638]
data[639-797]
data[798-956]
data[957-1115]
data[1116-1274]
_________________

2. In another test I explored whether I could increase the length of tuples being iterated through. At n = 66 (which creates tuple for all coordinates in an upper triangular 66x66 matrix), using cores = 2), iter_flow hangs on my machine, even thought it was perfectly happy with n = 65. I am going to look into this further tomorrow, but you might want to run these tests and see if you have similar issues. It may point to a hidden issue.

Thanks,
kmayerb