-
Notifications
You must be signed in to change notification settings - Fork 0
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
Fixes for shiny Ska3 #317
Fixes for shiny Ska3 #317
Conversation
@taldcroft do you want this for the matlab release or don't care? |
Not for the MATLAB release. Still need to check on performance and this is a lower priority. |
That's what I figured. So I think proseco, chandra_aca, sparkles are good to have releases cut now. |
@jeanconn - I believe this is ready now. I've done performance testing and things look OK, see the description. |
Somewhere between numpy 1.12 and 1.15 the behavior of argsort changed so that np.argsort([0] * 20) might not be stable. Table was relying on that (upstream fix needed), but in the meantime adding columns like this is now efficient in astropy Table.
Previously astropy Table was making a deep copy of meta attributes but this changed for improved efficiency.
@@ -507,7 +507,8 @@ def plot(self, ax=None, **kwargs): | |||
kwargs.setdefault('bad_stars', np.zeros(len(kwargs['stars']), dtype=bool)) | |||
|
|||
fig = plot_stars(attitude=self.att, ax=ax, **kwargs) | |||
plt.show() | |||
if 'agg' not in plt.get_backend().lower(): | |||
plt.show() |
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.
In shiny (new matplotlib) one gets a warning in tests about calling plt.show()
for a non-interactive backend.
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 looks good to me. For proseco changes I'm also wondering if trouble will be saved by testing on Windows before merging (I haven't started up that VM in a while but suppose I need to get more facility with it anyway).
I think that testing individual package updates by hand will not save trouble (i.e. reduce level of effort). Windows-only fails are pretty rare, so until we get to a point where tests are automated or really easy to run, it will be less effort to fix rare problems at the end than always test up front. Note that despite efforts in that direction, Windows testing is a bit of a pain and we are not testing the exact environment that the FOT will be using anyway. |
Description
This has some changes that are required to pass tests in ska3-shiny (Python 3.6 version). Details are given in the commit messages.
This branch passes tests in current ska3-flight. The only catch is that 186897d will be slower in ska3-flight.
Timing
Running
proseco/tests/timing.py : time_get_aca_catalog(n_samples=100)
So this is a little slower in current Ska3-flight, but will be faster in shiny.