Release 0.1 about reproject

[astrofrog]

I'll see if I can improve the test coverage a little.

[cdeil]

Is #59 a real bug? Or just a misunderstanding or something that has been fixed in the meantime?

[cdeil]

I'm playing with the examples from the docs ... everything works so far.

The example in http://reproject.readthedocs.org/en/latest/healpix.html gives these warnings:

>>> array, footprint = reproject_from_healpix(filename_ligo, target_header)

/Users/deil/Library/Python/3.4/lib/python/site-packages/astropy-1.1.dev12384-py3.4-macosx-10.10-x86_64.egg/astropy/coordinates/angles.py:672: RuntimeWarning: invalid value encountered in less
  if np.any(self_angle < wrap_angle_floor) or np.any(self_angle >= wrap_angle):
/Users/deil/Library/Python/3.4/lib/python/site-packages/astropy-1.1.dev12384-py3.4-macosx-10.10-x86_64.egg/astropy/coordinates/angles.py:672: RuntimeWarning: invalid value encountered in greater_equal
  if np.any(self_angle < wrap_angle_floor) or np.any(self_angle >= wrap_angle):
/Users/deil/Library/Python/3.4/lib/python/site-packages/astropy-1.1.dev12384-py3.4-macosx-10.10-x86_64.egg/astropy/coordinates/angles.py:563: RuntimeWarning: invalid value encountered in less
  if np.any(angles.value < lower) or np.any(angles.value > upper):
/Users/deil/Library/Python/3.4/lib/python/site-packages/astropy-1.1.dev12384-py3.4-macosx-10.10-x86_64.egg/astropy/coordinates/angles.py:563: RuntimeWarning: invalid value encountered in greater
  if np.any(angles.value < lower) or np.any(angles.value > upper):

It's not important, but probably these should be fixed or suppressed?

[cdeil]

Here it should be target_header instead of header_in I think:
https://github.com/astrofrog/reproject/blame/17b20618bd568062dcf3db00fe31bd0ff15743e6/docs/healpix.rst#L107

[cdeil]

There's also this warning which should be suppressed because it occurs frequently in normal usage (I think):

In [8]: array_exact, footprint_exact = reproject_exact(hdu2, hdu1.header)
/Users/deil/Library/Python/3.4/lib/python/site-packages/reproject-0.0.dev214-py3.4-macosx-10.10-x86_64.egg/reproject/spherical_intersect/core.py:190: RuntimeWarning: invalid value encountered in true_divide
  return array_new / weights, weights

[astrofrog]

#59 is a real bug

The warnings due to the NaN values are annoying so this should fix it: https://github.com/astrofrog/reproject/pull/78

The header_in typo has been fixed directly in master with 0f367cb

[astrofrog]

I fixed the other warning you mentioned in #78 too.

[astrofrog]

@cdeil - can you try the latest master version again? I think all the warnings should now be fixed. I don't think I will be able to address #59 any time soon, but I think it's a corner case. Anything else that should be addressed?

[cdeil]

One thing that could be useful for users: say which scipy and healpy version is needed for reproject here:
http://reproject.readthedocs.org/en/latest/#requirements

[cdeil]

For the healpix example from the docs I still get these warnings from wcsaxes.

Possible / worth fixing or suppressing?

from astropy.wcs import WCS
import matplotlib.pyplot as plt

ax = plt.subplot(1,1,1, projection=WCS(target_header))
ax.imshow(array, vmin=0, vmax=1.e-8)
ax.coords.grid(color='white')
ax.coords.frame.set_color('none')

/Users/deil/Library/Python/3.4/lib/python/site-packages/wcsaxes-0.4.dev512-py3.4.egg/wcsaxes/transforms.py:171: RuntimeWarning: invalid value encountered in greater
  invalid = np.any(np.abs(pixel_check - pixel_full) > 1., axis=1)
/Users/deil/Library/Python/3.4/lib/python/site-packages/wcsaxes-0.4.dev512-py3.4.egg/wcsaxes/coordinate_range.py:50: RuntimeWarning: invalid value encountered in greater
  reset = np.abs(wjump) > 180.
/Users/deil/Library/Python/3.4/lib/python/site-packages/wcsaxes-0.4.dev512-py3.4.egg/wcsaxes/coordinate_range.py:58: RuntimeWarning: invalid value encountered in greater
  reset = np.abs(wjump) > 180.
/Users/deil/Library/Python/3.4/lib/python/site-packages/wcsaxes-0.4.dev512-py3.4.egg/wcsaxes/coordinate_range.py:11: RuntimeWarning: invalid value encountered in greater
  values_new[values_new > 180.] -= 360
/Users/deil/Library/Python/3.4/lib/python/site-packages/wcsaxes-0.4.dev512-py3.4.egg/wcsaxes/coordinate_helpers.py:470: RuntimeWarning: invalid value encountered in greater
  reset = (((normal_angle_full - tick_angle) % 360 > 90.) &
/Users/deil/Library/Python/3.4/lib/python/site-packages/wcsaxes-0.4.dev512-py3.4.egg/wcsaxes/coordinate_helpers.py:471: RuntimeWarning: invalid value encountered in greater
  ((tick_angle - normal_angle_full) % 360 > 90.))
/Users/deil/Library/Python/3.4/lib/python/site-packages/wcsaxes-0.4.dev512-py3.4.egg/wcsaxes/coordinate_helpers.py:482: RuntimeWarning: invalid value encountered in greater
  w1[w2 - w1 > 180.] += 360
/Users/deil/Library/Python/3.4/lib/python/site-packages/wcsaxes-0.4.dev512-py3.4.egg/wcsaxes/coordinate_helpers.py:483: RuntimeWarning: invalid value encountered in greater
  w2[w1 - w2 > 180.] += 360
/Users/deil/Library/Python/3.4/lib/python/site-packages/wcsaxes-0.4.dev512-py3.4.egg/wcsaxes/coordinate_helpers.py:513: RuntimeWarning: invalid value encountered in less
  np.nonzero(((t - w1) * (t - w2)) < 0)[0]])
/Users/deil/Library/Python/3.4/lib/python/site-packages/wcsaxes-0.4.dev512-py3.4.egg/wcsaxes/grid_paths.py:40: RuntimeWarning: invalid value encountered in greater
  sep[sep > np.pi] -= 2. * np.pi
/Users/deil/Library/Python/3.4/lib/python/site-packages/wcsaxes-0.4.dev512-py3.4.egg/wcsaxes/grid_paths.py:42: RuntimeWarning: invalid value encountered in greater
  mask = np.abs(sep > ROUND_TRIP_TOL)

[cdeil]

Saving the HEALPIX file I get Astropy deprecation warnings with healpy version 1.8.4.
Are these fixed in healpy master already? Worth a healpy issue or pull request?

from healpy import write_map
write_map('healpix_map.fits', array)

WARNING: AstropyDeprecationWarning: The new_table function is deprecated and may be removed in a future version.
        Use :meth:`BinTableHDU.from_columns` for new BINARY tables or :meth:`TableHDU.from_columns` for new ASCII tables instead. [healpy.fitsfunc]
WARNING:astropy:AstropyDeprecationWarning: The new_table function is deprecated and may be removed in a future version.
        Use :meth:`BinTableHDU.from_columns` for new BINARY tables or :meth:`TableHDU.from_columns` for new ASCII tables instead.
WARNING: AstropyDeprecationWarning: The use of header.update() to add new keywords to a header is deprecated.  Instead, use either header.set() or simply `header[keyword] = value` or `header[keyword] = (value, comment)`.  header.set() is only necessary to use if you also want to use the before/after keyword arguments. [astropy.io.fits.header]
WARNING:astropy:AstropyDeprecationWarning: The use of header.update() to add new keywords to a header is deprecated.  Instead, use either header.set() or simply `header[keyword] = value` or `header[keyword] = (value, comment)`.  header.set() is only necessary to use if you also want to use the before/after keyword arguments.

[cdeil]

@astrofrog – I tried all examples from the docs again. All work, except for #76 for which you said it's an Astropy / CFITSIO / Macports issue. So 👍 to release now ... thank you !!!!

[astrofrog]

@cdeil - could you report the warnings for WCSAxes and healpy in the respective issue trackers? (this is not something we can fix here)

I've added the minimum versions (as far as I can tell, might get updated if people submit bug reports) in 0f6e5c6

[cdeil]

OK, I'll file issues about the warnings from wcsaxes and healpy with as short as possible test cases.

[cdeil]

I've made separate issues for the warnings mentioned above at astrofrog/wcsaxes#159 and healpy/healpy#253.

[astrofrog]

I think we're good to go for 0.1, since the notebook issue is not something that we can easily control

[astrofrog]

Version 0.1 has been released!

@cdeil - thanks for all your help!

[cdeil]

pip install reproject and running the tests works for me.

I think the stable branch is pointing to bc2cdc2 but the 0.1 tag is pointing to ee8adde , which leads to readthedocs showing a dev version when I select "stable". Not a big deal ... probably it's OK if this is resolved with the 0.2 release ....

[astrofrog]

@cdeil - this is because I made a few small changes to the docs after the release (including install instructions). A little annoying esthetically, but will make sure that's not needed next time.