richteague/bettermoments

A robust way to calculate line-of-sight velocities.

bettermoments's Issues

Channel width in the zeroth moment maps

I think the variable chan in methods.collapse_zeroth should be a positive value in case that velax is a decreasing sequence, i.e.,

chan = abs(np.diff(velax).mean())

Otherwise, the zeroth moment values (M0) and their uncertainties (dM0) will have opposite signs.

installation woes

would be useful to add an installation note, since imgcube isn't something available in an obvious way. Not via pip, but
git clone https://github.com/richteague/imgcube
inside of the notebooks directory make it work.

Had to comment out the plt.style.use('paper') My anaconda didn't have it. Now looks awfully small, what's the trick to get it working with 'paper' option?

Calculate optimal velocity resolution

Given an analytical line profile and an achieved SNR determine the optimal velocity resolution for calculating the line centroid. Should be useful for proposals. @ryanaloomis

AssertionError: Unexpected shape of `moments`.

Hi,
I'm attempting to make a moment 9 cube fits but I get the error above. Here's my current usage of bettermoments:

bettermoments cube.fits -method ninth -clip 4 -firstchannel 5 -lastchannel 64

Out:

Loading up data...
Loading up user-defined mask...
Defining channel-based mask...
Smoothing the data...
Estimating noise in the data...
Estimated RMS: 9.82e-04.
Calculating threshold-based mask...
Masking the data...
Calculating maps...
Saving maps...
Traceback (most recent call last):
  File "/home/bnorfolk/anaconda3/envs/py38/bin/bettermoments", line 8, in <module>
    sys.exit(main())
  File "/home/bnorfolk/anaconda3/envs/py38/lib/python3.8/site-packages/bettermoments/collapse_cube.py", line 439, in main
    save_to_FITS(moments=moments,
  File "/home/bnorfolk/anaconda3/envs/py38/lib/python3.8/site-packages/bettermoments/io.py", line 258, in save_to_FITS
    assert moments.ndim == 3, "Unexpected shape of `moments`."
AssertionError: Unexpected shape of `moments`.

Coordinate Definition Changes

Header is hardcoded to only return ICRS coordinates, despite not applying any conversion.

Error without outname option in command line

Dear Rich,

I installed the latest version and run in command-line without -outname option, but I got an error saying AttributeError: 'str' object has no attribute 'copy' at line 264 in io.py. Maybe path.copy() should be simply path ?

13 too many words...

So close @dfm!

command line option 'outname' is ignored

After installing the newest version 1.6.1, it looks like the -outname option is ignored. In other words, even when specifying -outname, the output moment 0 maps' names are based on the input cubes' names. Furthermore, they are saved in the same directory as the input cube, so it seems no longer possible to specify the location where the output maps should be saved.

Return model cubes

If creating moment maps using analytical fits, have an option to additionally return a restored model cube.

Default clip

Just following on from our conversation - I think it would be more intuitive to have the default clip set at zero sigma (i.e. the call when no "-clip" is set at all to revert to "-clip 0" rather than "-clip 5" as is), or have it explicitly written out that it's using 5 sigma somewhere...

fits header key 'date-obs' not preserved

It looks like the 'date-obs' keyword in the headers of the moment 0 maps are set to some default value (1858/11/17/00:00:00). Would it be possible to pass along the 'date-obs' from the header of the input cube to the header of the output moment 0 maps?
It wonder where this 1858/11/17/00:00:00 date comes from. Maybe this is astropy's fits.writeto with output_verify='silentfix'? However, I tried output_verify='exception', and it still put that default date into the header without raising an exception.

No bash aliasing with Python2.7

Can't use the defined entry_points in setup.py as for Python 3.

Consider and probably implement the Courteau method for flux conservation

Appendix C of: https://arxiv.org/abs/astro-ph/9709201

Tutorial notebooks need updating

As title.

Submit!

I'm happy for this to go in.
Will you submit to AAS and ArXiv?

Push figures to repo

@richteague: can you add the figures to the git repo so that we can make the paper from a checkout?