Skip to content

Migrate random glue_lowering to overload where easy #8061

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
sklam merged 9 commits into from
Aug 17, 2022
Merged

Migrate random glue_lowering to overload where easy #8061

Changes from 1 commit
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
bea6f65
Migrate random `glue_lowering` to `overload` where easy
apmasell May 17, 2022
ba597b0
Fix missing array overrides
apmasell Jun 1, 2022
189c9ae
Relax types on hypergeometric
apmasell Jun 2, 2022
fe6fd8e
Add missing tests and remove unncessary type checks
apmasell Jul 18, 2022
4fa48fa
Fix parameter names and test `size=None`
apmasell Jul 20, 2022
c100b2f
Create special unit tests for gamma distribution over array
apmasell Jul 25, 2022
f8684a9
Ensure logseries works on 32-bit
apmasell Jul 26, 2022
74bfeae
Add test for remaining random array implementations
apmasell Jul 27, 2022
f5c5b8f
Add test for vonmises random array
apmasell Aug 4, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
Relax types on hypergeometric 
Allow floating point types in hypergeometric distribution
  • Loading branch information
@apmasell
apmasell committed Jun 2, 2022
commit 189c9aefc1063f418bc33d6e569aafa92b8d5d0b
21 changes: 11 additions & 10 deletions numba/cpython/randomimpl.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1153,21 +1153,21 @@ def _impl(loc, scale, size):

@overload(np.random.hypergeometric)
def hypergeometric_impl(ngood, nbad, nsamples):
if (isinstance(ngood, types.Integer) and isinstance(nbad, types.Integer)
and isinstance(nsamples, types.Integer)):
if (isinstance(ngood, types.Number) and isinstance(nbad, types.Number)
and isinstance(nsamples, types.Number)):
def _impl(ngood, nbad, nsamples):
"""Numpy's algorithm for hypergeometric()."""
d1 = nbad + ngood - nsamples
d1 = int(nbad) + int(ngood) - int(nsamples)
d2 = float(min(nbad, ngood))

Y = d2
K = nsamples
K = int(nsamples)
while Y > 0.0 and K > 0:
Y -= math.floor(np.random.random() + Y / (d1 + K))
K -= 1
Z = int(d2 - Y)
if ngood > nbad:
return nsamples - Z
return int(nsamples) - Z
else:
return Z

Expand All @@ -1176,11 +1176,12 @@ def _impl(ngood, nbad, nsamples):

@overload(np.random.hypergeometric)
def hypergeometric_impl(ngood, nbad, nsamples, size):
if (isinstance(ngood, types.Integer) and isinstance(nbad, types.Integer) and
isinstance(nsamples, types.Integer) and isinstance(size, types.NoneType)):
return lambda ngood, nbad, nsamples, _size: np.random.hypergeometric(ngood, nbad, nsamples)
if (isinstance(ngood, types.Integer) and isinstance(nbad, types.Integer) and
isinstance(nsamples, types.Integer) and
if (isinstance(ngood, types.Number) and isinstance(nbad, types.Number) and
isinstance(nsamples, types.Number) and isinstance(size, types.NoneType)):
return lambda ngood, nbad, nsamples, _size:\

@sklam sklam Jul 19, 2022

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
return lambda ngood, nbad, nsamples, _size:\
return lambda ngood, nbad, nsamples, size:\

np.random.hypergeometric(ngood, nbad, nsamples)
if (isinstance(ngood, types.Number) and isinstance(nbad, types.Number) and
isinstance(nsamples, types.Number) and
(isinstance(size, types.Integer) or (isinstance(size, types.UniTuple)
and isinstance(size.dtype,
types.Integer)))):
Expand Down