Add support for Solar Orbiter PHI instrument

[dgarciabriseno]

This patch adds support for SOLO's PHI FDT-BLOS, FDT-ICNT, HRT-BLOS, and HRT-ICNT images.

TODO:

• Needs a web directory to ingest these images

[dgarciabriseno]

How the test images I was given look on my dev system:

[wafels]

Images may need to be rotated. If north is up, then those active regions are at very high latitudes!

[bogdanni]

Can WCS processing be implemented in the web client? That is, CRVAL/CRPIX/CROTA etc.

[dgarciabriseno]

I think it would be a big change since a lot of code was written with the assumption that the jp2's would already be centered and rotated. It's worth an exploration, but I don't have time to look into it yet.

[bogdanni]

EUI images are also like that, the proper orientation is achieved via WCS keywords.

[wafels]

The discussion on WCS in helioviewer is here:

Helioviewer-Project/helioviewer.org#509

[dgarciabriseno]

I'm able to apply the rotation client side. I need to work on keeping it in sync with the back end.

[dgarciabriseno]

Hit a roadblock trying to support client side rotation. The problem is that to make screenshots, we extract tiles from an unrotated jp2 image. kakadu doesn't support extracting information this way.

I can't think of a way to make this work. Something like

• rotate selection area
• reformat rotated selection to be an unrotated rectangle (required by kakadu)
• re-rotate cutout back to appropriate position (axis of rotation could be outside the image (image magick doesn't support this)

runs into issues.

So I think what we have to do is rotate the images during data ingestion.
I'm able to use glymur + scipy to rotate the images and write out new jp2s. It doesn't retain the so called "quality layers," it does retain the multiple resolution levels.