This is a quick workaround for shark wildcard expansion picking up the hidden .shark files pyshark is generating. Not something I think we should do longer term, but it's blocking me right now, so this is a quick work around.
Ideally the pyshark metadata would be on xattr in ZFS, but I had to push it to dot files as shark-publish will lose those. I think longer term I need to have shark-publish take shark xattr data and publish it.
Another option (not necessarily mutually exclusive) is to have a targeted wildcard where we say ".geojson" rather than "" - I think that could be useful, but is the wrong solution here as the user won't except to need to consider the .shark files.
This is a quick workaround for shark wildcard expansion picking up the hidden .shark files pyshark is generating. Not something I think we should do longer term, but it's blocking me right now, so this is a quick work around.
Ideally the pyshark metadata would be on xattr in ZFS, but I had to push it to dot files as shark-publish will lose those. I think longer term I need to have shark-publish take shark xattr data and publish it.
Another option (not necessarily mutually exclusive) is to have a targeted wildcard where we say ".geojson" rather than "" - I think that could be useful, but is the wrong solution here as the user won't except to need to consider the .shark files.