randint accept ints for 'size'
#916
Conversation
Codecov Report
@@ Coverage Diff @@
## master #916 +/- ##
=======================================
Coverage 95.49% 95.49%
=======================================
Files 64 64
Lines 9793 9796 +3
=======================================
+ Hits 9352 9355 +3
Misses 441 441
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
There was a problem hiding this comment.
Thanks a lot @mtar for looking into this right away! I found another inconsistency vs. numpy, see comments.
| shape = (int(size),) | ||
| else: | ||
| if not all(ele > 0 for ele in shape): | ||
| raise ValueError("negative dimensions are not allowed") | ||
|
|
There was a problem hiding this comment.
I was thinking of addressing this like we do for shape in ht.reshape():
def reshape(a: DNDarray, *shape: Union[int, Tuple[int, ...]], **kwargs) -> DNDarray:
...Default size should be 0, not None. Here a numpy example:
>>> d = np.random.randint(0,50)
>>> d
1
>>> d = np.random.randint(0,50, 1)
>>> d
array([0])
>>> e = ht.random.randint(0,50)
>>> e
DNDarray([9], dtype=ht.int32, device=cpu:0, split=None) # ndim=1 but should be 0!| # size parameter allows int arguments | ||
| a = ht.random.randint(1, size=10, split=0, dtype=ht.int64) | ||
| self.assertTrue(ht.equal(a, b)) | ||
|
|
There was a problem hiding this comment.
Add tests where no size is specified and scalar is returned?
|
run tests |
Description
Issue/s resolved: #913
Changes proposed:
Type of change
Memory requirements
n/a
Performance
n/a
Due Diligence
Does this change modify the behaviour of other functions? If so, which?
no