-
Notifications
You must be signed in to change notification settings - Fork 141
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
Extensions of FileIO interface #373
Conversation
This was only for ImageMagick, see JuliaIO/ImageMagick.jl#7
@timholy the Imagine port is incomplete because: - don't know file ending - don't know magic byte - would need to implement submodule loading in Fileio (own packige for Imagine?) you should probably just do the changes you feel to be appropriate.
This reverts commit 6e7afc0.
Awesome!!! |
OK. @rsrock, I'm going to take the liberty of moving the SIF reader over to JuliaIO, and make sure you have push privileges. |
https://github.com/JuliaIO/AndorSIF.jl Once this passes its tests, I'm merging 😄. |
Ha! That would be quite something :) |
Win64 0.3 is off by one: Expression: raw(res) --> raw(map(ScaleAutoMinMax(Ufixed16),raw(imgr)))
Expected:
Uint16[65535 43131
49330 23688
35780 41243]
Uint16[40516 36945
43380 18967
28964 0]
Occurred:
Uint16[65535 43131
49330 23688
35780 41243]
Uint16[40516 36945
43379 18967
28964 0] This must be a rounding error due floating point precision problems? |
I want to try to debug it first, so for now let's have it generate enough output to reproduce the problem. |
The writemime test failure is expected until JuliaIO/FileIO.jl#37 gets merged. |
No problem, I'll catch up with all these changes when I have the chance. I assume osxnative is moving somewhere else too? Let me know where that will go, I'd like to add image writing too. I doubt that Andor SIF is used much. At one point I tried to grab a small test image, but it was with the wrong version of the software and the format was different. Of course. Ronald S. Rock, Jr., Associate Professor
|
Little checklist:
Probably incomplete ;) |
OK, I think we're basically good to go here. This couples to the following PRs:
I think lots of travis tests will fail (surprisingly, this one succeeded) until packages are registered and tagged. OK to go ahead? |
(I'm not sure what the "rounding error" means above.) |
This one: Expression: raw(res) --> raw(map(ScaleAutoMinMax(Ufixed16),raw(imgr)))
Expected:
Uint16[65535 43131
49330 23688
35780 41243]
Uint16[40516 36945
43380 18967
28964 0]
Occurred:
Uint16[65535 43131
49330 23688
35780 41243]
Uint16[40516 36945
43379 18967
28964 0] |
Ah, thanks for reminding me. Yes, let's wait for it to happen again and see what the debugging output tells us. |
Extensions of FileIO interface
Aside from a weird problem with QuartzImageIO, things are looking good. I'm interested in tagging a new (minor) release very soon. A key question: should Images (temporarily?) add
to its REQUIRE file? Also, if folks can test I'd be grateful. |
Wow, is it Christmas already!? :)
|
💯 Whoo |
A But with the new system, I'm back to having ImageMagick problems:
A Edit to clarify: I'm not sure it's this specific patchset that introduces the regression or any other since the last release, so the problem might well be unrelated! |
Thanks for checking. Right now you have to say |
Ah, right, I read that comment but comitted it to memory with a bitflip :) Adding the |
Awesome! Thanks for the report. I think we can take this as evidence that we should add it to the REQUIRE 😄. |
That, or to the README.md if you want to force people to choose their loader. |
I excised ec2dec5, reverted the commit (#370) that caused the test failure, created some deprecations, and moved Imagine into its own package.
Only AndorSIF to go, right? Do you want me to tackle that?