#19 converter

[r-yamada1998]

No description provided.

[KaoruNishikawa]

Quick suggestions:

• FITS header conversion method in the following line should be tostring, not to_string.

  xarray-cube/xarray_cube/converter.py

  Line 13 in 7141b93

  header=cube.hdu.header.to_string(),

• The name of FITS ImageHDU class seems invalid (should start with capital letter).

  xarray-cube/xarray_cube/converter.py

  Line 23 in 7141b93

  cube = sc.SpectralCube.read(fits.imageHDU(data, header))

[KaoruNishikawa]

SpectralCube automatically modifies FITS header like below; are these behaviour OK for this project?
Please let me know what implementation will be employed as the differences affect the test script's implementation.

• The hdu property of SpectralCube modifies the FITS header comments like below

  xarray-cube/xarray_cube/converter.py

  Line 13 in 85dbf94

  header=cube.hdu.header.tostring(),

  >>> cube.hdu.header
  SIMPLE  =                    T / conforms to FITS standard
  ...
  >>> cube.header
  SIMPLE  =                    T /Standard FITS  # same as original hdu
  ...

  If this change isn't acceptable, please use cube.header.

• The string expression of unit can be modified

  >>> hdu.header["BUNIT"]  # original
  'Jy/beam'
  >>> cube.header["BUNIT"]
  'beam-1 Jy'
  >>> cube.hdu.header["BUNIT"]
  'beam-1 Jy'

  I couldn't find the way to keep the expression.

Dismissing, since new implementation problems are found.

[astropenguin]

It seems that cube.hdu creates a new primary HDU while cube.header does not. So I think we should use the latter.

• https://github.com/radio-astro-tools/spectral-cube/blob/9681cca50c1ba9c23a9eee33380afa5c9f1a9f0d/spectral_cube/spectral_cube.py#L2441
• https://github.com/radio-astro-tools/spectral-cube/blob/9681cca50c1ba9c23a9eee33380afa5c9f1a9f0d/spectral_cube/spectral_cube.py#L2456

[r-yamada1998]

cube.headerを用いた場合は、headerの中身によってはheaderを書き換えたものを返すように思われます。特に、cunit3のvalueがない場合などは、自動的にheaderにcunit3のvalueを追加する仕様です。いずれの方法を選んでもオリジナルから変更が行われるという点では大差ないように思われますが、どのようにいたしましょうか？

[astropenguin]

Thank you for your contribution! It looks good to me.