Use virtual file system when checking CRC

[alex-schulster]

Hi, I noticed the CRC file which verifies the downloaded file checksum does not use the file system provided in the user config.

This causes an issue for a file system which changes the root location of CFDP files (I created a file systems which handles all cfdp files in /tmp/cfdp/ to avoid polluting my executable dir for example).

I therefore edited the crc file to make use of this vfs.

All the tests have been run and passed successfully:

❯ python -m pytest
================================ test session starts =================================
platform linux -- Python 3.10.12, pytest-8.0.0, pluggy-1.4.0
rootdir: /home/alex/code/python/cfdppy/module
plugins: pyfakefs-5.3.5
collected 77 items

tests/test_checksum.py .                                                       [  1%]
tests/test_dest_handler_acked.py ...............                               [ 20%]
tests/test_dest_handler_naked.py ..............                                [ 38%]
tests/test_filestore.py ........                                               [ 49%]
tests/test_lost_seg_tracker.py .............                                   [ 66%]
tests/test_request.py ...                                                      [ 70%]
tests/test_src_handler_acked.py ........                                       [ 80%]
tests/test_src_handler_nak_closure.py .......                                  [ 89%]
tests/test_src_handler_nak_no_closure.py ........                              [100%]

================================= 77 passed in 1.31s =================================

Also minor adjusments in other files:

• Authorize HostFilestore to create parent dirs when creating file
• Replace absolute imports with relative imports
• Typing consistency between HostFilestore and VirtualFilestore

[robamu]

Thanks for this PR. Some questions I have:

1. Is there a special advantage for using relative imports instead of absolute ones?
2. I am not fully sure whether the automatic creation is the best solution. I'd like this interface to behave closely to a UNIX filesystem. For example, when I use touch /tmp/test/hello.txt and test does not exist in tmp, the command will fail. I think it might be a better idea to inform the user that the filesystem structure might not be the way they expect it to be.

To be honest, practically it might also be a good idea to handle the CRC inside the VFS via a dedicated interface function. The VFS knows the most efficient way to read/write a file, so it would also know the most efficient way to calculate the CRC of a file. In C++/Rust, the CRC calculation being performed by the VFS was the only practical approach. I might consider moving the CRC calculation into the VFS in the future..

[robamu]

I have made the changes mentioned in this PR: #11

The CRC calculation is now part of the VFS abstraction and I also fixed a bug where the CRC algorithm was not used for empty files.