Revised file-upload logic.

[bschroeter]

Due to the API changes from ME.org, the file upload logic needed changing.

Summary:

• API endpoint changed.
• API response changed to immediately return a file_id
  • This negates the need to perform a file_status, but does still require waiting some interval for the file to actually move to the object store.
• The entire client and CLI have been updated and CLI modified with testing.
• Multiple files can be passed to the client on the command line with glob expansion

Originally, we thought that we could upload everything in a directory at once, however, the receiving server has a packet limit, which we have had raised to 200mb. This will allow large (but singular) files to be transferred up to this size. Smaller files can feasibly be sent all at once.

[bschroeter]

@SeanBryan51, New commit added with the suggested fixes. I think we're in a good spot now, it's much cleaner.

[bschroeter]

I've added some comments regarding the upload_files function. The design of upload_files is bugging me a bit. Would it make more sense to have a separate function for uploading in memory file objects? This way I don't think you would need a temporary _files variable. Sorry I probably should have raised this in my first review.

A singular function is implemented. It would be basically a heap of replicated code to handle objects in a separate function. The use case on objects is slim, mostly for testing and direct connection via Python. 99% of the usage will be on the command line with a path that gets opened.