Mark/support loading swapped axes images

[markccchiang]

Description
This is intended to solve #1953, plotting coordinates for swapped-axes image cubes. It has to work with the backend PR 1194. To test it we may need more sample images with swapped-axes cubes.

Checklist

• changelog updated / no changelog update needed
• e2e test passing / added corresponding fix
• protobuf updated to the latest dev commit / no protobuf update needed

[kswang1029]

Based on my report

Almost the problems are all solved. About PV diagram for the swapped image. Here I have a test, if I set a line region (x coordinate starts from 400.5 and ends to -0.5; y coordinate starts from 20 to ends to 20) in "ngc3079-swap-rfd.fits" and making a PV diagram (e.g. ngc3079-swap-rfd_pv-ConvertToChannel20Image.fits), then the generated PV diagram should be a 2D image in "ngc3079-swap-rdf.fits" of channel of 20 (e.g. ngc3079-swap-rdf-Channel20Image-drop.fits). And the generated PV is consistent to 2D image of channel 20 (two images subtracted and all pixels are 0). This is because the line region length is odd number (if line region length is not odd number then, the generated PV may not 100% match to the 2D image), this part requires more investigation (my feeling is more related to PV rather than supporting swapped image)

I have tested this backend to the current ICD-RxJS, there is one test needs update, once the corresponding backend has been merged, then I can update the test (FILEINFO_FITS_MULTIHDU.test.ts) code very soon.

Please go ahead for the code refactoring and code reviewing.

Again, for the current purpose, we don’t focus on any region related analyses. It is not the scope here, whether the results are correct or not. This is a doc capturing all affected analyses https://docs.google.com/document/d/1eNLZt-VxSrzeYaOFjajJqbKXzIOo-u29xL7g3Cstz2c/edit

[kswang1029]

@markccchiang when a swapped cube is loaded, we will still need to display wcs: ( , ) in the cursor info bar according to the displayed axes. If it is RA vs freq, then we display something like wcs: (12:30:30, 1400.0), for example. This is to be consistent with what we do for PV images.

[kswang1029]

feature-wise this is looking good for the current scope - supporting image visualization of a swapped-axes cube. 👍 Please proceed code review.

[YuHsuan-Hwang]

The changes are looking good! I only have two minor comments.

[kswang1029]

for some reason e49334d introduces test failures for regions. world coordinates of the region are not updating. There are errors from AST.

[kswang1029]

looking good. new e2e tests are implemented.