Refactor IO Directory Processing and Compression Logic

[Daniil-Golikov]

• First, the directory processing has been completely redone to replace the outdated and unmaintainable code used for old directory processing. Now, it simply invokes process_single_file.

• Second, the logic has been changed so that it no longer creates a new folder named "foldername-(de-)compressed". Instead, (de-)compressed files are now placed in the same folder.

• Third, I've reverted to the previous logic; if a file is unsupported (for example, raw), the code shouldn't stop. Instead, the file should be skipped.

Note: All the changes were discussed in meeting with Carlos.

P.S. The integration tests are really great. I've modified the directory tests due to the change in logic.

P.S. As discussed with Carlos, if the changes are deemed not beneficial enough and/or if the trade-offs(such as returning Options instead of just results from the readers) are deemed significant enough, I'll close this PR and eliminate all the directory processing

[rukai]

I'll defer to carlos on the design of the directory handling functionality, but code wise nice cleanup!

[cjrolo]

Good cleanup, let's give this a try!