[AstroPy] A quick survey of PyFITS users

Tom Aldcroft aldcroft at head.cfa.harvard.edu
Thu May 26 08:54:33 EDT 2011


Hi Erik,

Congratulations on the new 3.0 release and the Python 3 compatibility!
 This is good news for PyFITS users and the Python community.  I just
have some documentation requests:

- Could you put a link to the API docs on the main page or somewhere
obvious?  I never even knew they existed online until I just googled
"pyfits api".

- The current online API docs refer to version 2.2.2, not 2.4.0.
Since this is a widely used package, it might be worth the effort to
host API docs for all releases, e.g.
http://stsdas.stsci.edu/stsci_python_epydoc/pyfits/2.4.0/index.html.

- Could you put a version number in the PyFITS Handbook in addition to
a release date?

- Is there any chance of making an online HTML version of the PyFITS
Handbook instead of only PDF?  For me, HTML would be much easier to
use and I wouldn't end up with 10 copies of the manual sitting in my
downloads directory.

Thanks,
Tom

On Wed, May 25, 2011 at 7:11 PM, Erik Bray <embray at stsci.edu> wrote:
> Dear AstroPy community,
>
>     Obviously, if you're not a PyFITS user you can skip this :)  As
> some of you know I've been working on the next version of PyFITS, which
> will hopefully be released within the next couple months.  I've called
> it version 3.0 for two primary reasons:  It's the first version to
> support Python 3, and it contains a pretty large overhaul to the source
> code.
>
>     Despite the large changes, I've tried, at least for this release,
> to maintain as much backwards-compatibility as possible with the last
> release (2.4.0).  So most users hopefully shouldn't even notice much of
> a difference.  I've already had a handful of unofficial beta-testers.
> They've been very helpful in helping me squash some bugs, but haven't
> otherwise pointed out any serious API incompatibilities.
>
>     However, there are some large changes to the 'private' API
> (anything prefixed with an underscore), as well as to a few nominally
> 'public' methods, but that aren't of much use to most users.  In
> addition, there are a few interfaces that are still supported, but that
> I would like to deprecate.
>
>     So I just wanted to try to get a sense of what interfaces users are
> actually using, and if anybody is likely to experience significant
> difficulties with this new version.  Here are my questions:
>
> 1) Are you using any 'private' variables, classes, functions, etc. from
> PyFITS?  That is, is your code using anything imported from pyfits that
> begins with an underscore?
>
> 2) Do you use the classExtensions feature at all?
>
> 3) Do you have any custom subclasses of any classes in pyfits?
>
> 4) Do you use the _ValidHDU.req_cards() method or the Card.fromstring()
> method at all?
>
> 5) Are there any other odd PyFITS interfaces you directly access that
> are not documented in the PyFITS Users' Manual?
>
>
> I hope to deal with as many problems as I can now, so as to not get
> inundated with "bug" reports once users switch to this new version.  Of
> course, not everybody who uses PyFITS is on this mailing list, and that
> will probably happen anyways.  But any information I can gather to
> mitigate that will be greatly helpful.
>
> Thanks,
>
> Erik
> _______________________________________________
> AstroPy mailing list
> AstroPy at scipy.org
> http://mail.scipy.org/mailman/listinfo/astropy
>
>



More information about the AstroPy mailing list