astropy and numpy are now dependences, pyfits can be used.

[montefra]

• [x] numpy is a dependancy, there is no need to check it
• [x] astropy is a dependancy, there is no need to check it
• [x] make it possible to use either astropy or pyfits fits representation. Will help for cases like #51

[montefra]

@ericmandel : I have began to add tests.

As far as you know, is there any way to run ds9 without the GUI?

[ericmandel]

As far as you know, is there any way to run ds9 without the GUI?

@montefra You need an X server, but you can use the virtual Xvfb. See:

http://ds9.si.edu/doc/faq.html

for details. @wjoye will have more recent experience with this than I.

[montefra]

Thank you for the pointer. I'll see if I can use this: it might be enough. For now I just want to check that I can send and receive data.

[montefra]

there is a pytest pluging for Xvfb. I'll give a try in the next few days.

[ericmandel]

That would be pretty convenient!

[montefra]

@ericmandel : good news! pytest-xvfb works and I can test ds9. I have added tests for {set,get}_fits. I will add similar tests for {set,get}_pyfits and then I will call this issue closed.

I'll add inline comments as review to clarify the points. Can you please go through them and let me know if something doesn't make sense?

[ericmandel]

@montefra Better than good news, this is great news and great work. The whole Python community owes you a debt ...

The changes looks fine and make sense, although I hasten to admit that if there are subtle details, they may be a bit beyond me, my Python knowledge having leaked away due to non-use.

Again, it's really good work.

[montefra]

The changes looks fine and make sense, although I hasten to admit that if there are subtle details, they may be a bit beyond me, my Python knowledge having leaked away due to non-use.

The main thing is to check that I am not doing something stupid with the way I communicate with ds9, in particular when using xpaget.

Again, it's really good work.

Thank you :blush:

[montefra]

thank you for checking. Sometimes English slips ...

On the {get,set}_pyfits: I tried to keep its behaviour as similar as possible to the old one. If pyfits is not installed, warn the user to use {get,set}_fits. Otherwise {get,set}_pyfits gets or returns a pyfits.HDUList instance.

[montefra]

If you don't have any request or comment, I might merge this PR

[ericmandel]

It looks fine, I'd merge it and take the rest of the day off!

[montefra]

@ericmandel : do you ha ve preferences on how to do the merging?

The merge button shows three options:

1. Merge pull request: I guess that it does a merge or a fast-forward depending on the status
2. Squash and merge: I don't like the idea of squashing
3. Rebase and merge: first make sure that the changes are on top of the master and then do a fast-forward.

I think that for this case 1. and 3. are the same

[ericmandel]

I generally just merge. Both rebasing and squashing can hide the actual flow of development and, in general, I would prefer to have the commits match that flow.

[montefra]

@ericmandel : fine with me.