Handle aperture photometry when images linked by WCS for advanced cases #2154

pllim · 2023-04-14T21:40:16Z

Description

This pull request is to investigate and hopefully fix the bug of aperture photometry not able to provide correct result in certain cases, e.g.:

When images have different pixel scales.
When Subset no longer retains original shape after being projected onto a non-reference image.

Blocked by

POC: Circular annulus support in Imviz #2164

TODO

Remove warning added in DOC: Add warning to aperture photometry plugin about apertures #2167 about the aperture (distortion warning should stay).
Implement the fix.
- Can we handle this by passing around sky regions in the background?
- ~~Can we also fix Edit Subset plugin by translating sky region to pixel of the selected image? Would need new dropdown for the plugin.~~
See if we can also address Recenter tool only looks at reference layer #2265
- Add tests

Change log entry

Is a change log needed? If yes, is it added to CHANGES.rst? If you want to avoid merge conflicts,
list the proposed change log here for review and add to CHANGES.rst before merge. If no, maintainer
should add a no-changelog-entry-needed label.

Checklist for package maintainer(s)

This checklist is meant to remind the package maintainer(s) who will review this pull request of some common things to look for. This list is not exhaustive.

Are two approvals required? Branch protection rule does not check for the second approval. If a second approval is not necessary, please apply the trivial label.
Do the proposed changes actually accomplish desired goals? Also manually run the affected example notebooks, if necessary.
Do the proposed changes follow the STScI Style Guides?
Are tests added/updated as required? If so, do they follow the STScI Style Guides?
Are docs added/updated as required? If so, do they follow the STScI Style Guides?
Did the CI pass? If not, are the failures related?
Is a milestone set? Set this to bugfix milestone if this is a bug fix and needs to be released ASAP; otherwise, set this to the next major release milestone.
After merge, any internal documentations need updating (e.g., JIRA, Innerspace)? 🐱 🐱 🐱

pllim · 2023-04-14T21:41:13Z

notebooks/concepts/imviz_advanced_aper_phot.ipynb

+    "axs[1].set_ylim(70, 80)\n",
+    "axs[2].imshow(rect_grp.subsets[2].to_mask(), origin='lower')\n",
+    "axs[2].set_xlim(105, 125)\n",
+    "axs[2].set_ylim(70, 80);"


This would give:

This corresponds to the following images loaded into Imviz and linked by WCS:

@larrybradley , to answer your question, if I draw a rectangle on the third image (down-sampled and rotated), the rectangle comes back warped as if I drew it on the reference image, which is what I suspected would happen.

pllim · 2023-04-17T21:06:14Z

@larrybradley and I discussed several options. The most scientifically correct way is still using sky regions (as suggested by @bmorris3). However, this does disable the plugin for any data without WCS.

Hence, Larry also requested that we put in extra logic to still fall back to the old way of doing things (pixel regions) if image does not have WCS and is the reference image for the Subset. But is that the same as a viewer reference image (have to check what becomes the viewer reference image when you only display a non-reference image in the viewer; answer is still the original reference image)?

While we're at it, I guess technically we can always falls back to using pixel regions if we detect the link is in pixel space whether the data of interest has WCS or not. Keep in mind that even when Imviz is asked to link by WCS and the reference data has WCS, any image without WCS still links to the reference data by pixel.

And since glue-astronomy does not handle sky regions (see glue-viz/glue-astronomy#89), we would have to rely on our own region_translators.py.

pllim · 2023-04-18T15:23:39Z

For completeness, Larry said photutils would ignore distortion anyway (that is, if you pass in a circular aperture and WCS with distortion, it would still be circular when extracting the data array, not a distorted circle shape) so we should not worry about distortion here.

pllim · 2023-04-21T17:23:44Z

For future me: With this, I don't need GWCS.

And since I will add all these as static test data files, hopefully we never have to rerun these again.

To downsample by 2

w2_hdr['CRPIX1'] = 250 / 2
w2_hdr['CRPIX2'] = 150 / 2
w2_hdr['NAXIS1'] = 250
w2_hdr['NAXIS2'] = 150
w2_hdr['CDELT1'] = 2
w2_hdr['CDELT2'] = 2

Goes with:

data_sm = block_reduce(data, 2)

Where:

data = make_100gaussians_image()
w = make_wcs(data.shape)

To rotate

https://github.com/musevlt/mpdaf/blob/321cec4e95dc70c3e7eaf7402a2e3a24ebfb5b8e/lib/mpdaf/obj/coords.py#L1550

def rotate(wcs_in, theta):
    wcs = deepcopy(wcs_in)
    theta = np.deg2rad(theta)
    sinq = np.sin(theta)
    cosq = np.cos(theta)
    mrot = np.array([[cosq, -sinq],
                     [sinq, cosq]])

    if wcs.wcs.has_cd():    # CD matrix
        newcd = np.dot(mrot, wcs.wcs.cd)
        wcs.wcs.cd = newcd
        wcs.wcs.set()
    elif wcs.wcs.has_pc():      # PC matrix + CDELT
        newpc = np.dot(mrot, wcs.wcs.get_pc())
        wcs.wcs.pc = newpc
        wcs.wcs.set()
    else:
        raise TypeError("Unsupported wcs type (need CD or PC matrix)")
    
    return wcs

Goes with:

data_rot, _ = reproject_interp(image_2, w3, shape_out=data_sm.shape)

jdaviz/configs/imviz/tests/data/gauss100_fits_wcs.fits

notebooks/concepts/imviz_advanced_aper_phot.ipynb

codecov · 2023-08-10T03:20:59Z

Codecov Report

Patch coverage: 100.00% and project coverage change: +0.02% 🎉

Comparison is base (0c97b76) 90.70% compared to head (3ab15d3) 90.72%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2154      +/-   ##
==========================================
+ Coverage   90.70%   90.72%   +0.02%     
==========================================
  Files         157      157              
  Lines       17969    18015      +46     
==========================================
+ Hits        16299    16345      +46     
  Misses       1670     1670

Files Changed	Coverage Δ
...igs/default/plugins/subset_plugin/subset_plugin.py	`93.64% <100.00%> (+0.10%)`	⬆️
jdaviz/configs/imviz/helper.py	`97.35% <100.00%> (+0.21%)`	⬆️
...imviz/plugins/aper_phot_simple/aper_phot_simple.py	`92.87% <100.00%> (+0.21%)`	⬆️
jdaviz/configs/imviz/plugins/viewers.py	`89.65% <100.00%> (-0.86%)`	⬇️
...daviz/configs/imviz/tests/test_simple_aper_phot.py	`100.00% <100.00%> (ø)`
jdaviz/core/region_translators.py	`98.92% <100.00%> (-0.09%)`	⬇️

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

bmorris3

Neat!

jdaviz/configs/default/plugins/subset_plugin/subset_plugin.py

jdaviz/configs/default/plugins/subset_plugin/subset_plugin.vue

jdaviz/configs/imviz/helper.py

bmorris3 · 2023-08-10T17:51:15Z

jdaviz/configs/imviz/plugins/viewers.py

+        # TODO: Brett Morris might want to look at this for
+        # https://github.com/spacetelescope/jdaviz/pull/2179
+        #     ref_label = self.state.reference_data ???
+        #
+        # The original links were created against data_collection[0], not necessarily
        # against the current viewer reference_data
        ref_label = self.session.application.data_collection[0].label


If I understand your comment correctly, you're thinking about (near-)future-proofing this by looking up the current reference data's label? If so, I'll suggest that change here.

jdaviz/configs/imviz/plugins/viewers.py

jdaviz/core/region_translators.py

bmorris3

Thanks for the updates 👌🏻

kecnry · 2023-08-11T17:22:04Z

jdaviz/configs/default/plugins/subset_plugin/subset_plugin.vue

+        <v-text-field v-if="item.name === 'Parent'"
+          :label="item.name"
+          :value="item.value"
+          style="padding-top: 0px; margin-top: 0px"
+          :readonly="true"
+          hint="Subset was defined with respect to this reference data (read-only)"
+        ></v-text-field>


I think we'll eventually want to move this out of the subregions once reparenting is implemented (or perhaps remove entirely), but since its read-only, it doesn't hurt to have for now and could be useful for debugging.

kecnry · 2023-08-11T17:23:32Z

jdaviz/configs/imviz/helper.py

+        """Find the type of ``glue`` linking between the given
+        data labels. A link is bi-directional. If there are
+        more than 2 data in the collection, one of the given
+        labels should be the reference data or look-up will fail.


@bmorris3 - does the rotation work remove some of this flexibility (by forbidding mixed linking)? If we might need to change or remove this, should we make it private in the meantime?

I think this method is useful with or without the rotation work. It's likely the case that it will be redundant with the link type attribute after image rotation is merged and WCS-linked in Imviz, but I think it's handy for debugging.

Do you think it should be public or private in the meantime?

I don't see why it should be private. The information is so useful in the meantime that I think it should be public. If it comes a day we don't need it anymore, then we can deprecate it.

I agree that public is fine.

pllim · 2023-08-14T16:02:42Z

There is a conflict now, I will have to rebase and let the CI run again.

advanced aperture photometry. Add tests that will fail without fix WIP: How to actually fix this mess Remove warning from doc [ci skip] [rtd skip]

BUG: Make sure recentering works in Subset Tools WIP: What about the actual bug [ci skip] [rtd skip]

to images linked by WCS if sky projection is different

and remove outdated comment

and remove debug comment

Co-authored-by: Brett M. Morris <[email protected]>

pllim · 2023-08-14T17:01:35Z

Looks like a new deprecation warning from scipy over the weekend. It is unrelated.

`numpy.distutils` is deprecated since NumPy 1.23.0, as a result

pllim · 2023-08-14T18:07:58Z

Thanks, all!

pllim added this to the 3.5 milestone Apr 14, 2023

github-actions bot added the documentation Explanation of code and concepts label Apr 14, 2023

pllim commented Apr 14, 2023

View reviewed changes

pllim force-pushed the aper-subset-not-ref branch from 57de42c to 2e6eb56 Compare April 17, 2023 21:29

pllim force-pushed the aper-subset-not-ref branch from 2e6eb56 to 08a1b27 Compare April 19, 2023 21:04

pllim commented Apr 21, 2023

View reviewed changes

jdaviz/configs/imviz/tests/data/gauss100_fits_wcs.fits Outdated Show resolved Hide resolved

pllim added bug Something isn't working imviz labels Apr 21, 2023

This comment was marked as resolved.

Sign in to view

This was referenced Apr 24, 2023

POC: Circular annulus support in Imviz #2164

Closed

DOC: Add warning to aperture photometry plugin about apertures #2167

Merged

haticekaratay modified the milestones: 3.5, 3.6 May 25, 2023

pllim force-pushed the aper-subset-not-ref branch from 53677ca to 4bc18e0 Compare July 6, 2023 21:30

javerbukh modified the milestones: 3.6, 3.7 Jul 28, 2023

pllim force-pushed the aper-subset-not-ref branch 3 times, most recently from f6a38a9 to cfb3ff1 Compare August 8, 2023 22:04

pllim commented Aug 8, 2023

View reviewed changes

notebooks/concepts/imviz_advanced_aper_phot.ipynb Outdated Show resolved Hide resolved

pllim force-pushed the aper-subset-not-ref branch from 0184fe6 to 901375b Compare August 10, 2023 03:07

pllim marked this pull request as ready for review August 10, 2023 03:54

pllim requested review from duytnguyendtn, rosteen, javerbukh and kecnry as code owners August 10, 2023 03:54

pllim requested review from bmorris3, haticekaratay and cshanahan1 as code owners August 10, 2023 03:54

bmorris3 reviewed Aug 10, 2023

View reviewed changes

pllim force-pushed the aper-subset-not-ref branch from fcb3e91 to 71738c0 Compare August 10, 2023 21:07

bmorris3 self-requested a review August 11, 2023 13:16

bmorris3 approved these changes Aug 11, 2023

View reviewed changes

pllim added the Ready for final review label Aug 11, 2023

kecnry reviewed Aug 11, 2023

View reviewed changes

kecnry approved these changes Aug 14, 2023

View reviewed changes

pllim and others added 8 commits August 14, 2023 12:05

EXP: Build test case for

6221785

advanced aperture photometry. Add tests that will fail without fix WIP: How to actually fix this mess Remove warning from doc [ci skip] [rtd skip]

FEAT: Add parent info to ROI subset

dc827c8

BUG: Make sure recentering works in Subset Tools WIP: What about the actual bug [ci skip] [rtd skip]

BUG: Fix aperture incorrectly applied

9064835

to images linked by WCS if sky projection is different

Fix PEP 8 warning

5fc28a8

and remove outdated comment

Moar fixes and tests

b421485

Add tests for sky bg

cca45af

and remove debug comment

Apply suggestions from code review

444ef5c

Co-authored-by: Brett M. Morris <[email protected]>

Shorten code

3ab15d3

Co-authored-by: Brett M. Morris <[email protected]>

pllim force-pushed the aper-subset-not-ref branch from 71738c0 to 3ab15d3 Compare August 14, 2023 16:27

pllim merged commit 7dcc714 into spacetelescope:main Aug 14, 2023
13 of 14 checks passed

pllim deleted the aper-subset-not-ref branch August 14, 2023 17:04

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Handle aperture photometry when images linked by WCS for advanced cases #2154

Handle aperture photometry when images linked by WCS for advanced cases #2154

pllim commented Apr 14, 2023 •

edited

Loading

pllim Apr 14, 2023

This comment was marked as outdated.

pllim Apr 17, 2023

pllim commented Apr 17, 2023 •

edited

Loading

pllim commented Apr 18, 2023

pllim commented Apr 21, 2023 •

edited

Loading

This comment was marked as resolved.

codecov bot commented Aug 10, 2023 •

edited

Loading

bmorris3 left a comment

bmorris3 Aug 10, 2023 •

edited

Loading

bmorris3 left a comment

kecnry Aug 11, 2023

kecnry Aug 11, 2023

bmorris3 Aug 14, 2023

kecnry Aug 14, 2023

pllim Aug 14, 2023

bmorris3 Aug 14, 2023

pllim commented Aug 14, 2023

pllim commented Aug 14, 2023

pllim commented Aug 14, 2023

Handle aperture photometry when images linked by WCS for advanced cases #2154

Handle aperture photometry when images linked by WCS for advanced cases #2154

Conversation

pllim commented Apr 14, 2023 • edited Loading

Description

Blocked by

TODO

Change log entry

Checklist for package maintainer(s)

Choose a reason for hiding this comment

This comment was marked as outdated.

Choose a reason for hiding this comment

pllim commented Apr 17, 2023 • edited Loading

pllim commented Apr 18, 2023

pllim commented Apr 21, 2023 • edited Loading

To downsample by 2

To rotate

This comment was marked as resolved.

codecov bot commented Aug 10, 2023 • edited Loading

Codecov Report

bmorris3 left a comment

Choose a reason for hiding this comment

bmorris3 Aug 10, 2023 • edited Loading

Choose a reason for hiding this comment

bmorris3 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

pllim commented Aug 14, 2023

pllim commented Aug 14, 2023

pllim commented Aug 14, 2023

pllim commented Apr 14, 2023 •

edited

Loading

pllim commented Apr 17, 2023 •

edited

Loading

pllim commented Apr 21, 2023 •

edited

Loading

codecov bot commented Aug 10, 2023 •

edited

Loading

bmorris3 Aug 10, 2023 •

edited

Loading