WIP: initial tweaks to ImageSet / avoid dask duplicating work

[tjgalvin]

Sorry I didn't mean to raise this as a PR. I thought I had it as a branch.

I am just going through one step at a time. I think I will tweak this image set concept a little bit.

I'd hazard a guess that imaging is young to be the dominate factor in all of this.

On Tue, 9 May 2023, 5:36 pm Alec Thomson, @.***> wrote:

@.**** commented on this pull request.

In arrakis/imager.py https://github.com/AlecThomson/arrakis/pull/19#discussion_r1188286180:

@@ -70,7 +73,7 @@ def get_prefix( return out_dir / prefix

@.() @.(nout=3)

Since this now returns a single ImageSet, I think this should be nout=1

In arrakis/imager.py https://github.com/AlecThomson/arrakis/pull/19#discussion_r1188287946:

@@ -385,32 +411,69 @@ def smooth_image(image, common_beam_pkl, cutoff=None): return sm_image

@.***() +def smooth_images_in_imageset(image_set: ImageSet, common_beam_pkl, cutoff=None) -> ImageSet:

• Smooth image

• Deserialise the beam

• with open(common_beam_pkl, "rb") as f:
• common_beam = pickle.load(f)
• images = []
• sm_images = {}
• for pol, pol_images in image_set.image_lists.items():

This will really hurt parallelism

— Reply to this email directly, view it on GitHub https://github.com/AlecThomson/arrakis/pull/19#pullrequestreview-1418110728, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACQOAJTG6F43UC3Y4WFFLNDXFIF2LANCNFSM6AAAAAAX3AOHVQ . You are receiving this because you were assigned.Message ID: @.***>

[AlecThomson]

Imaging is looking good for now - lets merge into imager and work on integration