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
eigenvalues and eigenvectors using arb #30393
Comments
Branch: u/mmezzarobba/acb_eig |
Commit: |
This comment has been minimized.
This comment has been minimized.
New commits:
|
comment:2
Note that the signature of that method is changing in #29243. Maybe better to follow that? |
Branch pushed to git repo; I updated commit sha1. New commits:
|
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
|
comment:6
Few comments:
What should happen when the parent is
Can this benefits from your current branch? |
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
|
comment:8
Replying to @seblabbe:
Done. (Though the second point is bugware. What we need IMHO is for each method's documentation to include or link to that of the methods it overrides, and to change the way we write docstrings to take advantage of that. It doesn't make any sense to repeat the same information over and over with no automatic way of keeping the various places in sync.)
[...]
In principle yes (and matrices over RR and CC can benefit too), but at the moment we don't even have a dedicated class for matrices over RBF. I think we should keep that for another ticket, and probably wait until the methods are no longer marked experimental. |
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
|
comment:10
Replying to @mezzarobba:
Thanks.
I agree with your point (out of sync). But, I do hate the documentation of That being said, I think my point in comment 2 was mostly to avoid using the
since it will not be the 1st argument anymore after #29243. If you can force the
I agree. |
comment:11
I get an error while building the documentation:
In fact it seems to be triplicate:
|
Branch pushed to git repo; I updated commit sha1. New commits:
|
comment:14
Replying to @seblabbe:
It may still be slightly better to have it there, to be compatible with the generic method in case someone explicitly passes a second argument equal to |
comment:15
Replying to @seblabbe:
I don't understand what is going on (why should two different inline links necessarily have distinct link texts?!), but let's see if the commit I just pushed solves the issue. |
comment:16
Can you do the following change (x3) ? - See the <http://arblib.org/acb_mat.html#c.acb_mat_eig_multiple for
+ See http://arblib.org/acb_mat.html#c.acb_mat_eig_multiple for |
comment:17
also: - http://arblib.org/acb_mat.html#c.acb_mat_approx_eig_qr_
+ http://arblib.org/acb_mat.html#c.acb_mat_approx_eig_qr |
Reviewer: Sébastien Labbé |
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
|
comment:21
Oops. I went too fast indeed, thanks! New commits:
|
comment:22
I found the correct fix. In ReStructuredText syntax, one underscore at the end creates a target -See the `Arb documentation <http://arblib.org/acb_mat.html#c.acb_mat_eig_multiple>`_
-See the `Arb documentation <http://arblib.org/acb_mat.html#c.acb_mat_approx_eig_qr>`_
-See the `Arb documentation <http://arblib.org/acb_mat.html#c.acb_mat_eig_simple>`_
+See the `Arb documentation <http://arblib.org/acb_mat.html#c.acb_mat_eig_multiple>`__
+See the `Arb documentation <http://arblib.org/acb_mat.html#c.acb_mat_approx_eig_qr>`__
+See the `Arb documentation <http://arblib.org/acb_mat.html#c.acb_mat_eig_simple>`__ Do you want to revert it back to using two underscores? |
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
|
comment:24
Replying to @seblabbe:
Oh, I didn't know about the difference, thank you!
Done (but untested; I don't want to rebuild the doc right now). |
comment:25
and I checked that documentation builds ok. |
comment:26
Replying to @mezzarobba:
I created #30444 to add it to the documentation. |
Changed branch from u/mmezzarobba/acb_eig to |
Changed commit from |
comment:28
While following the examples in the documentation in order to prepare something for the release tours, I noticed that the computed right eigenvectors are wrong. They do not satisfy the eigenvector equation. In fact, it seems to be a problem of transposition. The computed vectors are the rows of the eigenmatrix, but the eigenvectors would be the columns.
|
comment:29
Uh, thanks a lot for catching the issue! I thought I has checked... |
comment:30
Fixed (I hope) at #30568. |
comment:31
Oups, I didn't notice that issue during the review. Thanks. |
comment:32
Another follow up: #33652 |
The wrappers are marked as experimental because the corresponding arb functions are.
Component: linear algebra
Author: Marc Mezzarobba
Branch:
b792cbd
Reviewer: Sébastien Labbé
Issue created by migration from https://trac.sagemath.org/ticket/30393
The text was updated successfully, but these errors were encountered: