Update filenames for targets to make them more unique

[geordie666]

Currently, targets (and the associated directory structure) look like:

target/catalogs/dr8/0.39.0/targets/sv1/resolve/bright/sv1-targets-dr8-hp-21.fits

To make the filenames unique but free of extraneous information, make the following updates in io.find_target_files():

• Remove the trailing hyphen between sv1 and targets in the filename (sv1targets instead of sv1-targets).
• Remove the imaging release in the filename completely (targets- instead of targets-dr8).
  • Should probably also add a DR keyword to the header, to ensure that information is recorded somewhere in the file.
• Main Survey filenames should (still) look like targets- without any trailing main.
• Add the observing "layer" (bright/dark) to the filenames.
• Add noresolve to the filenames where needed for the rare cases that we release targets that are not resolved across the "North"/"South" imaging.

Some example results should look like:

target/catalogs/dr8/0.41.0/targets/main/resolve/dark/targets-dark-hp-33.fits
target/catalogs/dr8/0.41.0/targets/main/resolve/bright/targets-bright-hp-33.fits
target/catalogs/dr8/0.41.0/targets/sv1/resolve/dark/sv1targets-dark-hp-33.fits
target/catalogs/dr8/0.41.0/targets/main/noresolve/bright/targets-dark-noresolve-hp-33.fits

[sbailey]

Thanks. Expanding on this point more generally:

Should probably also add a DR keyword to the header, to ensure that information is recorded somewhere in the file

Indeed. As a general principle throughout all our code and files, any metadata that is encoded in the directory structure and filename should also be present in the header keywords. i.e. the directories and filenames are useful for finding the files and keeping ourselves organized, but the contents of the file should also be sufficient to re-derive the associated metadata to recreate the filepath and filename.

[geordie666]

Addressed in #637.