Add support for RGB tiles

[adl1995]

As discussed in issue #62, I have added a test case of JPG tiles, which is passing. However, I haven't added a test case for PNG tiles, because the current survey only support FITS, JPG tiles.

[tboch]

Could you use the 2MASS6XH sample in hips-extra, which has PNG tiles?

[adl1995]

@tboch The PNG tiles are already available in hips-extra. I was reluctant to add the test case as it would require a separate survey, the current one does not support PNG tiles.

I think a separate function get_test_tiles_png would server the purpose, or would it be better to take the tile path as a parameter?

[tboch]

I have no strong feeling about how to test this. But if the library supports PNG, we should add a test using PNG tiles.

[adl1995]

I just added a test case for PNG tiles, however, it is currently failing. The pixel values sum to 0.0. I suspect this is because it doesn't fall in the sky image.

[cdeil]

You have committed changes to astropy_helpers: https://github.com/hipspy/hips/pull/63/files Please get rid of them (and never commit changes to astropy_helpers in the future).

[cdeil]

I just added a test case for PNG tiles, however, it is currently failing. The pixel values sum to 0.0. I suspect this is because it doesn't fall in the sky image.

Can you pick another survey that has PNG and data at least for part of the sky image?

[tboch]

 Can you pick another survey that has PNG and data at least for part of the sky image?

I'd suggest using HiPS P/AKARI/FIS/Color that has PNG and data in the galactic plane.

[coveralls]

Coverage decreased (-0.2%) to 97.814% when pulling 958dc4d1dc6775004839d9dc5a01ca7465217bb2 on adl1995:rgb_support into 3cd9d60c8d1f723304d351cbdab58644e4886bbe on hipspy:master.

[cdeil]

@adl1995 - I merged https://github.com/hipspy/hips-extra/pull/3 just now. Does it work if you use tiles from that survey?

[cdeil]

@adl1995 - Maybe add a section on making an RGB color image in the high-level docs with a nice example? https://hips.readthedocs.io/en/latest/getting_started.html#more-advanced-examples

The HiPS paper uses "HiPS of the Spitzer GLIMPSE 360 survey" in Figure 4, which looks nice.

[adl1995]

@cdeil I just made some updates. For the example on getting started page, should I duplicate all the steps from above or only the ones that need changing?

[coveralls]

Coverage increased (+0.05%) to 97.613% when pulling 3c02fd09dafd027a85998ed6f130752637c04711 on adl1995:rgb_support into c9b92f1a689fa6799f624e5c710af3483395f8f6 on hipspy:master.

[cdeil]

For the example on getting started page, should I duplicate all the steps from above or only the ones that need changing?

As you like. I think I would re-use the geometry, but otherwise it's new.

Things to explain there:

• That HiPS also has JPEG and PNG, pointing out one example for each
• What the return data array is (it's shape, and that it's float even though int pixel data wend in from the tiles).
• End the example with how RGB images can be plotted with matplotlib / wcsaxes and saved to FITS or PNG / JPEG? (i.e. tell users a bit how they can work with the image they generate)

[coveralls]

Coverage increased (+0.05%) to 97.613% when pulling 3c02fd09dafd027a85998ed6f130752637c04711 on adl1995:rgb_support into c9b92f1a689fa6799f624e5c710af3483395f8f6 on hipspy:master.

[coveralls]

Coverage increased (+0.05%) to 97.613% when pulling 8ca56a7036540d528fb891a811a9bec901086aa3 on adl1995:rgb_support into c9b92f1a689fa6799f624e5c710af3483395f8f6 on hipspy:master.

[adl1995]

@cdeil Please merge this PR. I will make a separate PR for documentation changes.

[cdeil]

OK, merging now.