Update comparison to pyregion

[joleroi]

This PR adresses #45

It adds a script to compare speed and completeness of the DS9 region parsing to pyregion. It does not check correctness. This must be done manually for each region type, I guess though.

The result confirms what we already know. The regions package is far away from providing a complete DS9 support, but if it works it's much faster than pyregion.

          filename          time_regions time_pyregion compl_pyregion compl_regions
--------------------------- ------------ ------------- -------------- -------------
        ds9.linear.wcsi.reg       -1.000         0.735          91.4%          0.0%
     ds9.mosaic.fk4.hms.reg       -1.000         0.692          91.4%          0.0%
              ds9.color.reg        0.005         0.220         100.0%        100.0%
          ds9.fk5.strip.reg       -1.000        -1.000           0.0%          0.0%
                ds9.fk5.reg       -1.000         0.443          91.7%          0.0%
     ds9.physical.strip.reg       -1.000        -1.000           0.0%          0.0%
   ds9.physical.windows.reg       -1.000         0.407          91.4%          0.0%
            ds9.fk5.hms.reg       -1.000         0.441          91.7%          0.0%
           ds9.icrs.hms.reg       -1.000         0.441          91.7%          0.0%
         ds9.icrs.strip.reg       -1.000        -1.000           0.0%          0.0%
       ds9.color.spaces.reg        0.005         0.219         100.0%        100.0%
ds9.mosaic.ecliptic.hms.reg       -1.000         0.690          91.4%          0.0%
         ds9.linear.wcs.reg       -1.000         0.680          91.4%          0.0%
    ds9.mosaic.ecliptic.reg       -1.000         0.691          91.4%          0.0%
        ds9.linear.wcsc.reg       -1.000         0.714          91.4%          0.0%
      ds9.fk4.hms.strip.reg       -1.000        -1.000           0.0%          0.0%
    ds9.mosaic.galactic.reg       -1.000         0.691          91.4%          0.0%
        ds9.linear.wcsd.reg       -1.000         0.720          91.4%          0.0%
     physical_reference.reg        0.001         0.208         100.0%        100.0%
                ds9.fk4.reg       -1.000         0.439          91.7%          0.0%
        ds9.image.strip.reg       -1.000        -1.000           0.0%          0.0%
     ds9.icrs.hms.strip.reg       -1.000        -1.000           0.0%          0.0%
              ds9.image.reg       -1.000         0.412          91.7%          0.0%
     ds9.mosaic.fk5.hms.reg       -1.000         0.691          91.4%          0.0%
       ds9.mosaic.image.reg       -1.000         0.681          91.4%          0.0%
           ds9.physical.reg       -1.000         0.413          91.7%          0.0%
           ds9.ecliptic.reg       -1.000         0.443          91.7%          0.0%
     ds9.ecliptic.strip.reg       -1.000        -1.000           0.0%          0.0%
               ds9.fits.reg       -1.000         0.242         100.0%          0.0%
               ds9.icrs.reg       -1.000         0.442          91.7%          0.0%
         ds9.mosaic.fk4.reg       -1.000         0.692          91.4%          0.0%
    ds9.mosaic.icrs.hms.reg       -1.000         0.694          91.4%          0.0%
       ds9.ecliptic.hms.reg       -1.000         0.440          91.7%          0.0%
            ds9.fk4.hms.reg       -1.000         0.441          91.7%          0.0%
ds9.mosaic.galactic.hms.reg       -1.000         0.688          91.4%          0.0%
        ds9.mosaic.icrs.reg       -1.000         0.693          91.4%          0.0%
            ds9.comment.reg       -1.000        -1.000           0.0%          0.0%
          fk5_reference.reg        0.009         0.836         100.0%        100.0%
         ds9.mosaic.fk5.reg       -1.000         2.767          91.4%          0.0%
          ds9.fk4.strip.reg       -1.000        -1.000           0.0%          0.0%
     ds9.galactic.strip.reg       -1.000        -1.000           0.0%          0.0%
           ds9.galactic.reg       -1.000         0.442          91.7%          0.0%
          ds9.composite.reg       -1.000         0.441          91.7%          0.0%
        ds9.linear.wcsa.reg       -1.000         2.373          91.4%          0.0%
        ds9.linear.wcsp.reg       -1.000         2.176          91.4%          0.0%
      ds9.fk5.hms.strip.reg       -1.000        -1.000           0.0%          0.0%
       ds9.galactic.hms.reg       -1.000         0.586          91.7%          0.0%
     galactic_reference.reg        0.002         0.827         100.0%        100.0%
 ds9.galactic.hms.strip.reg       -1.000        -1.000           0.0%          0.0%
    ds9.mosaic.physical.reg       -1.000         1.550          91.4%          0.0%
 ds9.ecliptic.hms.strip.reg       -1.000        -1.000           0.0%          0.0%

[cdeil]

@joleroi - Thanks!

I think it might be nice to have the result table (with some comments above or below summarising the table) in the docs (in some non-prominent page) to show status? What is the goal here, have 100% for all the files? Is that possible? Why doesn't pyregion achieve that?

Or if you don't think the table belongs in the docs, just put the results table somewhere in the repo so that one can see it and link to it. (Github renders .cvs files nicely as tables now, but also markdown or rst work).

[keflavich]

should be n_regions += line.count("(") since there are some example files that have ;-separated regions instead of newline-separated.

[joleroi]

@keflavich
Thanks for the comment
@cdeil
I added the csv file to the repo for further discussion

ok to merge?