-
Notifications
You must be signed in to change notification settings - Fork 4
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
Add missing parameters to nlos-z.yaml
and refactor tal.render.render
module
#4
Conversation
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.
Wow, thanks a lot for the PR. The tal.render.render
module was just one huge function and now it feels much better. I'll need to use tal these days so I'll be able to check if the render function works with different capture types (e.g. confocal or exhaustive). I'll merge the PR when I check that it also works well for these cases.
I added some comments/curiosities of mine, overall I like the result. Thanks again for the work :D
tal/render/render.py
Outdated
# NOTE: Windows does not support multiprocessing | ||
run_mitsuba_f() | ||
else: | ||
# HACK: on macOS "spawn" method, which is the default since 3.8, |
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.
AFAIK you are the first macOS user of tal so it's nice to know that it works. Windows gave some problems to other guys in our lab
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.
And do you remember, by any chance, what those problems were? If it's the same issue, maybe it's better to make a proper fix then?..
|
||
|
||
@dataclass | ||
class Point2D: |
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.
Just out of curiosity: is there any special advantage of using Point2D instead of just a tuple, apart from the nicer class name?
Seems fine for me, I'm just asking in case this would be useful in other parts of my projects
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.
Yeah, this is a bit controversial, initially I made it into a named tuple along the lines of
Point2D = namedtuple('Point2D', 'x y')
But then got an impression that it would be convenient to have numpy conversion methods directly on instances. But it doesn't seem to be the case, so it's probably better to revert to named tuples since they
- would provide the convenience accessors like
pt.x
, - be compatible with the regular tuples, and
- require less code (literally a line per type :)
Co-authored-by: Diego Royo <diego.roymen@gmail.com>
…id 'context has already been set' exception
Co-authored-by: Diego Royo <diego.roymen@gmail.com>
Thanks a lot for the PR. Recently I had to make some changes to |
Hi! Please note it is more of a draft/request for comments than a polished change. E.g. it was only tested with
scan_type == "single"
, so likely there are problems in other regimes