DOC: np.random documentation cleanup and expansion. #13849
Conversation
| from numpy.random import Generator, PCG64 | ||
| rg = Generator(PCG64()) | ||
| from numpy.random import default_gen | ||
| rg = default_gen(12345) |
There was a problem hiding this comment.
Oops. This depends on #13840
|
Test failure looks unrelated. |
|
According to #13829 (comment)
would help forestall PRs with unacceptable changes |
|
Does this supersede #13675? |
|
Yes. |
|
|
||
| .. [1] The algorithm is carefully designed to eliminate a number of possible | ||
| ways to collide. For example, if one only does one level of spawning, it | ||
| is guaranteed that all states will be unique. But it's easier to |
There was a problem hiding this comment.
Ahh, here is the guarantee. Stream of comments...
| estimate the naive upper bound on a napkin and take comfort knowing | ||
| that the probability is actually lower. | ||
|
|
||
| .. [2] In this calculation, we can ignore the amount of numbers drawn from each |
There was a problem hiding this comment.
Same "ahh" here. Should do a review.
|
LGTM. Minor comments that were mostly already addressed. |
DOC: fix reference links
Also, make the jump step odd to make it a proper Weyl generator. No, this will never, ever, ever, ever matter. But it feels nicer.
| """ | ||
| step = 0x9e3779b97f4a7c15f39cc0605cedc834 | ||
| step = 0x9e3779b97f4a7c15f39cc0605cedc835 |
There was a problem hiding this comment.
I need a LGTM on this one last non-trivial bit, then we can merge this. @bashtage ?
There was a problem hiding this comment.
Feel free to round it which ever way that you prefer. LGTM.
There was a problem hiding this comment.
Thank you for your indulgence. :-)
|
@charris This is ready to go once the last CI finishes. |
|
Thanks Robert. |
I've corrected some errors, filled in some gaps, and extended some of the commentary.