Region handling: some valid ds9 regions are not supported

[keflavich]

I encountered an issue where a region file formatted like this:

icrs
point(17:25:26.34, -28:30:40.54) # point=x

is not recognized as a region file by CARTA, but it is a valid ds9 region file.

[kswang1029]

@keflavich weird. I cannot reproduce the issue. Could you send me the region file? The file is not in dos format, for sure?

[keflavich]

that's exactly the region file; I saved it as test.reg and could not open it. It is not DOS format, though an earlier test version was - is this a problem?

[kswang1029]

We had an issue due to the dos format. Now fixed in dev #658
Not sure if that’s related or this is a new one🤔

[kswang1029]

Do you have the header line starting with #?

[keflavich]

no, there is no header line

[kswang1029]

please add

# Region file format: DS9 CARTA 1.4

as the 1st line, otherwise it is hard to tell the region file is in ds9 format or crtf.

[keflavich]

I know that's a workaround, but that is not a requirement for a file to be a valid ds9 region file. The user should be able to tell CARTA what type of region file it is and try to read it. This is important because there are potentially enormous numbers of ds9-compatible .reg files around, and if CARTA can only read a subset of them, it results in creation of a new format standard.

[veggiesaurus]

I know that's a workaround, but that is not a requirement for a file to be a valid ds9 region file. The user should be able to tell CARTA what type of region file it is and try to read it. This is important because there are potentially enormous numbers of ds9-compatible .reg files around, and if CARTA can only read a subset of them, it results in creation of a new format standard.

I agree. CRTF requires #CRTF in the first line, so that narrows things down, and we can attempt to parse things as DS9 if they're missing that header line

[keflavich]

iirc, ds9 provided a large suite of test files, but I can't find it. In astropy-regions, we have a bunch too, though all of these presently have headers - this issue motivates adding a headerless test case:
https://github.com/astropy/regions/tree/main/regions/io/ds9/tests/data

[keflavich]

ah, the .strip.reg files are headerless

[o-smirnov]

Experiencing a similar problem with DS9 line regions:

Region file: im2di_clustered.tessel.reg.txt