Raise exceptions instead of calling sys.exit in decommutator

[joyvelasquez]

Closes #45

Replaces sys.exit calls in the decommutator with raised exceptions that describe what went wrong and which file was involved. sys.exit() with no argument exits with code 0, so failures looked like successes to anything running these scripts. main.py is unchanged.

Each except: that led to an exit becomes except Exception as exc: and uses raise ... from exc, so the original error (for example, FileNotFoundError) stays in the traceback.

Progress

ccsds_reader_pipeline.py

• file2bytestr: input file cannot be read → RuntimeError

swp_spc_l02l1.py (main)

• Log file cannot be opened → RuntimeError
• Leap second kernel not found or not loaded → RuntimeError
• SCLK kernel not found or not loaded → RuntimeError
• L1 CDF exists and overwrite is not set → FileExistsError
• Could not check for or delete an existing L1 CDF → RuntimeError

This is not the final form: more specific exception classes can be added later, as noted in the issue.

[joyvelasquez]

@namurphy Ready for review. All sys.exit calls in ccsds_reader_pipeline.py and swp_spc_l02l1.py now raise exceptions with messages that name the file involved.

[namurphy]

Nice. This code follows the best practice of defining long exception messages prior to raising an exception with them. When this is not done, exceptions tracebacks get horrible-er.

[namurphy]

🎗️ After switching to using logging (#50), we can replace the ***ERROR*** messages with a logger.error("...") so that users can have finer control over the verbosity that gets reported. That would be better in a later pull request, though.