| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|\ \ \ \ \ \ \ \
| | | | | | | | |
| | | | | | | | | |
MAINT: lstsq: compute residuals inside the ufunc
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
This prevents an overly large output array being allocated.
It also means the the residuals can be handled as a separate out argument in future.
|
|\ \ \ \ \ \ \ \ \
| |_|_|_|_|_|/ / /
|/| | | | | | | | |
DOC: Fix NumpyVersion example (closes gh-10935)
|
| | | | | | | | | |
|
|\ \ \ \ \ \ \ \ \
| |/ / / / / / / /
|/| | | | | | | | |
BUG: Return NULL from PyInit_* when exception is raised
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
I don't think this is documented anywhere, but I'm pretty sure module init
functions should return NULL in order to communicate that an exception
occurred during initialization (as is the standard Python/C API convention).
It's clear from the CPython code
[here](https://github.com/python/cpython/blob/master/Python/importdl.c#L162)
that if you don't return NULL, the exception is swallowed and replaced with the
message "initialization of %s raised unreported exception".
Admittedly, this is only useful for people porting Numpy to new platforms where
it is helpful to know where module initialization is failing, but it can't hurt.
|
|\ \ \ \ \ \ \ \ \
| |_|_|_|/ / / / /
|/| | | | | | | | |
Fixes #663: fixed dtype alignment for array of structs in case of converting from tuple descr
|
| | |_|/ / / / /
| |/| | | | | |
| | | | | | | |
| | | | | | | | |
from tuple descr
|
|/ / / / / / /
| | | | | | |
| | | | | | |
| | | | | | | |
Closes #10847
|
| |_|_|/ / /
|/| | | | |
| | | | | |
| | | | | | |
Fixes gh-10780
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | | |
Add missing underscore to prototype in check_embedded_lapack
|
| | |_|_|/ /
| |/| | | | |
|
|\ \ \ \ \ \
| |/ / / / /
|/| | | | | |
DOC: rework documents and silence warnings during sphinx build
|
| | | | | | |
|
| | | | | | |
|
| | | | | | |
|
| | | | | | |
|
| | | | | | |
|
| | | | | | |
|
| | | | | | |
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | | |
BUG: fix savetxt, loadtxt for '+-' in complex
|
| | | | | | | |
|
|\ \ \ \ \ \ \
| | | | | | | |
| | | | | | | | |
TST: reactivate module docstring tests, fix float formatting
|
| | | | | | | | |
|
|\ \ \ \ \ \ \ \
| |/ / / / / / /
|/| | | | | | | |
DOC: Fix link in `See Also` section of `randn` docstring.
|
| | | | | | | | |
|
|\ \ \ \ \ \ \ \
| | | | | | | | |
| | | | | | | | | |
DOC: Cross-reference full/full_like in a few "See Also" sections.
|
| | | | | | | | | |
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
While teaching numpy I was asked the best way to create an array of nan,
and `np.full` seem not be cross linked from many places; In particular
in the documentation of `zeros` and `ones` seam like obvious candidates
to add them.
Reorder all the see-also to be
- empty_like
- ones_like
- zero_like
- full_like
- empty
- ones
- zeros
- full
|
| |/ / / / / / /
|/| | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
* fixes #10886 : Bad link to `np.random.randint`
* Issue #10886: Change from `np.random.randint` to `randint` in 'See Also` doc-string section of mtrand.pyx
|
|\ \ \ \ \ \ \ \
| |_|_|_|_|_|/ /
|/| | | | | | | |
ENH: umath: don't make temporary copies for in-place accumulation
|
| | |/ / / / /
| |/| | | | |
| | | | | | |
| | | | | | | |
Do the operation `ufunc.accumulate(x, out=x)` without temporary copies.
|
|\ \ \ \ \ \ \
| | |_|_|_|/ /
| |/| | | | | |
MAINT: Implement `lstsq` as a `gufunc`
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
This does not yet enable any broadcasting, but makes doing so in future far
easier.
|
| | | | | | | |
|
| | | | | | | |
|
| | |_|/ / /
| |/| | | | |
|
|\ \ \ \ \ \
| |/ / / / /
|/| | | | | |
ENH: Add object loops to the comparison ufuncs
|
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
This helps solve RobotLocomotion/drake#8315
|
| |_|_|_|/
|/| | | | |
|
| | | | |
| | | | |
| | | | | |
Now falls back on sturges estimator when the IQR is zero
|
| | | | | |
|
|\ \ \ \ \
| |_|/ / /
|/| | | | |
BUG: core: fix NPY_TITLE_KEY macro on pypy
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
On Pypy, dictionary keys do not necessarily preserve object identity.
This however was assumed by the NPY_TITLE_KEY macro, which relies on
descriptor.c:568 using the same 'title' object both as a dictionary key
as an entry in the tuple inserted.
Since the items in the field dict are unique, value identity is however
sufficient for the NPY_TITLE_KEY macro. On PyPy, fix the macro by
comparing values instead.
|
|\ \ \ \
| |_|/ /
|/| | | |
BUG: fix wrong inplace vectorization on overlapping arguments
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The input arguments to inplace operations can overlap for accumulate
operatation where out[i+1] = in[i] + out[i]
This breaks the no loop carried dependency assumptions the compiler has
due to the ivdep pragma that is required for GCC to vectorize the loop.
Currently this does not harm as accumulate operations are out of place,
but future enhancements may change that.
Fix this by verifying the arguments do not overlap within a hardware
vector size.
As we do not know the vector size of future machines use an extremely
large value of 1024 bytes (commodity hardware currently has at most 64
bytes)
|
|\ \ \ \
| | | | |
| | | | | |
MAINT: Remove yield tests
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Just replaces yields with normal function calls. Some of these tests
could be parametrized with a bit of reorganization, however, the
comments in the tests would have been lost and they seemed worth
keeping.
|
| | | | | |
|
| | | | | |
|