update getViewWCS to adapt to projection

[ManonMarchand]

This is an adaptation of the method getViewWCS so that is also works for projections that are not "SIN".

What's new

new in API:

aladin.getCooFrame()

This new function returns the current cooFrame of the view. The retruned values can be "J2000", "J2000d", "Galactic" and are taken from the label returned by aladin.view.cooframe

What changed

aladin.getViewWCS()

now adapts to the projection and coordinate system in the keys CTYPE1 and CTYPE2 rather than always returning RA---SIN and DEC

What needs to be reviewed

Check that this is the expected output. Example:

"Galactic"

CD1_1: 0.018111371257799755
CD1_2: 0
​CD2_1: 0
​CD2_2: 0.018111371257799755
​CRPIX1: 1280
​CRPIX2: 378
​CRVAL1: 326.8736386214195
​CRVAL2: 7.499153148417243
​CTYPE1: "GLON-AIT"
​CTYPE2: "GLAT-AIT"
​NAXIS: 2
​NAXIS1: 2560
​NAXIS2: 756
​RADECSYS: "ICRS"

"J2000d"

CD1_1: 0.018111371257799755
CD1_2: 0
CD2_1: 0
CD2_2: 0.018111371257799755
CRPIX1: 1280
CRPIX2: 378
CRVAL1: 326.8736386214195
CRVAL2: 7.4991531484172445
CTYPE1: "RA---STG"
CTYPE2: "DEC--STG"
NAXIS: 2
NAXIS1: 2560
NAXIS2: 756
RADECSYS: "ICRS"

[ManonMarchand]

Also possible to add in this PR:

• I wanted to add a test but I did not manage to create a fake aladin instance with happy-dom
• we might need a setCooFrame method too?

[tboch]

Also possible to add in this PR:

• I wanted to add a test but I did not manage to create a fake aladin instance with happy-dom

I don't know, usage of happy-dom is new, perhaps @bmatthieu3 can help.

• we might need a setCooFrame method too?

There is already a setFrame method actually. For symmetry, I suggest keeping setFrame (in order not to break existing API) and adding setCooFrame as an alias.

[ManonMarchand]

There is already a setFrame method actually. For symmetry, I suggest keeping setFrame (in order not to break existing API) and adding setCooFrame as an alias.

Or we could just rename getCooFrame in getFrame?

[tboch]

There is already a setFrame method actually. For symmetry, I suggest keeping setFrame (in order not to break existing API) and adding setCooFrame as an alias.

Or we could just rename getCooFrame in getFrame?

This involve less changes, thus let's do that!

[ManonMarchand]

Should be all good now? :)

EDIT: Nope. We have issues with a few projections.

[ManonMarchand]

New commits add the following stuff:

• tooltip on projections in the selector (cause I got bored of looking for their names all the time)
• implements planetary fits standard as described in https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2018EA000388 and that will be understood in astropy once this PR gets merged Implement transformation to / from CoordinateFrame from / to related WCS astropy/astropy#14820

[bmatthieu3]

I would have chosen maybe something like: 'name' or 'label' instead of 'longName'

[ManonMarchand]

I was kinda un-inspired. "label" sounds perfect thanks 🙂

[tboch]

In table 12 of WCS FITS paper ( https://arxiv.org/pdf/astro-ph/0207413.pdf ), the 3-letters code is called code and the name is called projection name 😅

[ManonMarchand]

Yes, but then I'd need to rename the code into "code" because everywhere else it's "name" for us. Do we do that?

[tboch]

ok, let's keep it as you update it.

[ManonMarchand]

I think we can merge ?

[bmatthieu3]

I am merging this. It would have been great to run some test asserting some wcs are good (a planetary one, with different projection). But this involve to much deep diving into how jsdom/happy-dom can mock webgl2 canvas. This is for another future PR