Simplify File field model

[Arithmomaniac]

Currently, file fields expose RSAPI's FileMetadata and FileValue classes directly. These contain far more information than we need.

We should map to/from a custom object with the following three fields:

• byte[] data
• string FileName (can be inferred if FilePath is supplied)
• string FilePath (left blank on server reads)

[Arithmomaniac]

I've given more though to this, and think we should have two different types of File DTOs: one for reading/writing to a file pointer, and one for reading/writing to a byte array. Depending on which one was defined in the model, the results on a read operation would either be written to a temp file (whose path would be available on the object), or into the array.

As discussed in #106, we could also add a hidden field that would store the MD5 hash of the file from the last read/write operation. We could use this hash and compare it to the current value to determine whether the file needs to be updated on an update operation.

abstract class FileDto
{
	internal string LastOperationMD5 { get; set; }
}

class PersistedFile : FileDto
{
	public string FilePath { get; }
	public InMemoryFile StoreInMemory();
}

class InMemoryFile : FileDto
{
	public byte[] ByteArray { get; set; }
	public PersistedFile WriteToFile(string filePath);
}

[mrobustelli]

For now download if part of model...much like single choice field. If do not want to deal with file, do not add to model.