Support ARCSAT observatory images and add console script

[stevenstetzler]

I ran into some issues using koffi out of the box with images from the ARCSAT observatory. I looked into them and implemented some bug fixes and functionality changes so that koffi works with these images and in a workflow suitable for me.

The bug fixes are a few typos in the code and when specifying fits headers.

The changes include:

1. querying for the MPC list of observatories and checking if the OBSERVAT header matches an MPC obs-code before using it, falling back to observatory longitude and altitude.
2. Using the DATE-OBS and EXPTIME headers to compute the observation time (as observation start + half the exposure time) in case DATE-AVG or the mjd_key are not available. (ARCSAT sets the JD header instead).
3. Returning no objects if there are none in the field (interpreting the JPL query results)
4. Creating a console script so that koffi can be used as an executable on the command line with reasonable printing, like so:

$ koffi ./decam/calexp_DECam_VR_VR_DECam_c0007_6300_0_2600_0_1031745_S29_steven_bad_mask_B1a_20210909_step1_20230422T062824Z.fits --query JPL
    Object Name             RA                 Dec                  x                  y         
                           deg                 deg                                               
------------------- ------------------ ------------------- ------------------- ------------------
191962 (2005 UJ483)  351.4052083333333  -4.090527777777777 -508.68617785163644 2011.7077708054658
290566 (2005 UH115) 351.37641666666667 -4.1730833333333335   626.3775356202042 1618.8031201806855
 308715 (2006 GT36) 351.46904166666667  -4.106777777777777  -286.3405338958438  2881.786867739447
577486 (2013 EX118)  351.3792083333333  -4.170583333333334   592.0096278015878 1656.7991577724874
        (2015 BW18)  351.4807083333333  -4.060861111111111  -919.8912384145148  3046.138852885603

x and y are the pixel locations of the objects and are valuable for followup. Note that koffi may want to filter out objects that land outside of the bounds of the CCD, which three of these do with negative pixel values. Machine readable formats can be printed using e.g. --format csv.

[maxwest-uw]

these changes are great!! and I'll add an issue to filter out negative pixel values (or at least add the option to do so, I think colin said that there might be some cases in active asteroid science where ppl want objects that don't fall in the image because there might be a coma visible)