-
-
Notifications
You must be signed in to change notification settings - Fork 5.1k
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.
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
DOC: update 1.5.0 release notes #12221
Conversation
e0c86ef
to
3c966b6
Compare
3c966b6
to
732aa76
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks Tyler. Looks like another solid release
Generalized QR factorization routines (``?geqrf``) now have full ``_lwork`` | ||
counterparts. | ||
|
||
`scipy.linalg.cossin` Cosine Sine decomposition of unitary matrices has been |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This name is pretty bad, I'd expect something named cossin
to do an elementwise math operation rather than a decomposition. Should we rename it before the release?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we had this discussion but we couldn't came up with anything better and it is kind of what everybody goes with elsewhere. So we counted on the fact that it is under linalg namespace. But yes I would really appreciate a better one.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
either csd
(in analogy with svd
) or a more explicit cossin_decomp
would be better imho
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
csd
is already overloaded in scipy.signal.csd
and some circles already use it for a shortcut to cumulative power spectral density. So that won't fly
Frankly, cossin_decomp
is even uglier than cossin
also becomes a one-off since the others don't have it.
Why I prefer cossin
is that there is nothing called cossin
in any trig libraries. And you can't accidentally fall to that namespace.
``w, v, info`` to ``w, v, m, isuppz, info`` | ||
|
||
The order of output arguments ``w``, ``v`` of ``<sy/he>{gv, gvd, gvx}`` is | ||
swapped. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Noting here for further review: check if these give hard errors; sounds like silently changing numerical values now.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Only the argument order is changed to match the remaining ones. So I don't know if we can raise errors based on that or how.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
They do have different shape at least:
w : rank-1 array('d') with bounds (n)
v : rank-2 array('d') with bounds (n,n) and a storage
Still, it sounds like something to avoid. Was there a reason why it had to be done (e.g. infrastructure for a whole family of functions), or was it simply "let's make it consistent"?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Probably they were hardly ever used. Because their lwork argument was not correct. So both fixed and made consistent.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Okay that makes sense - I can try to add a sentence so that that bit doesn't read like people should go audit their code because we did something unusual.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
These are the PRs that I created. If there is anything I can do, let me know.
bf547f4
to
1cf8d83
Compare
@AtsushiSakai Ok, I condensed the large number of input validation/exception handling improvements to a single broad statement. |
1cf8d83
to
7a638d2
Compare
Ok, I did the "manual sweep" of Also, tried to fix the doc build errors because of |
7a638d2
to
48d4487
Compare
* copy over the release notes from the wiki and reformat/clean up for `rst` format * use `tools/authors.py` to draft in the author list for release `1.5.0`
Co-authored-by: Warren Weckesser <warren.weckesser@gmail.com>
Co-authored-by: Josh Wilson <person142@users.noreply.github.com>
* add some missing release note ENH PR entries * address some reviewer comments * some fixes for doc build errors with `make html-scipyorg` due to the release notes update * add issue/PR lists
48d4487
to
3a12bf2
Compare
If the CI is green after most recent update I will most likely merge. |
We can consider backporting for unresolved matters if we really have to. |
The ``binned_statistic_dd`` function with ``statistic="std"`` performance was | ||
improved by ~4x. | ||
|
||
``scipy.stats.kstest(rvs, cdf,...)`` now handles both one-sample and |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
perhaps put the 'kstest' higher up next to the 'kstwo' item as they both relate to K-S statistics?
The ``axis`` parameter was added to `scipy.stats.f_oneway`, allowing it compute | ||
multiple one-way ANOVA tests for data stored in n-dimensional arrays. Some | ||
performance improvements to ``f_oneway`` as well. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Add the missing word to
, and make the last phrase a sentence.
The ``axis`` parameter was added to `scipy.stats.f_oneway`, allowing it compute | |
multiple one-way ANOVA tests for data stored in n-dimensional arrays. Some | |
performance improvements to ``f_oneway`` as well. | |
The ``axis`` parameter was added to `scipy.stats.f_oneway`, allowing it to | |
compute multiple one-way ANOVA tests for data stored in n-dimensional | |
arrays. The performance of ``f_oneway`` was also improved for some cases. |
copy over the release notes from the wiki and reformat/clean up
for
rst
formatuse
tools/authors.py
to draft in the author list for release1.5.0
TODO:
gh_lists.py
is working: BUG, REL: gh_lists.py compromised scraping #122201.5.0
PRs withENH
label and fill in missing entries (please do help).mailmap
for author name issues brought up in this PR [NO AUTHOR COMPLAINTS AT ALL]