Add loggers to schema

[saskiad]

Loggers that describe data format headers. Bruno and Josh to provide examples

[jtyoung84]

Probably should be an issue in aind-metadata-mapper

[bruno-f-cruz]

This might require a longer discussion to try to solve this issue in a more holistic way. Regardless, I will try to start this discussion with some notes:

The case for flat binary files

People tend to gravitate towards the same formats in neuroscience. For instance, ephys and ophys data is often saved as flat binary files, where data can be readily decoded by having access to some logging "metadata". For this specific case, a user would need to know:

Layout: [ColumnMajor, RowMajor]
StructFormat: U16,U16,F32,F64....
If one assumes the file is homogenous (i.e. all values in a row have the same value), the metadata could even be:
Layout: [ColumnMajor, RowMajor]
DataFormat: U16
NumberOfChannels: 16

CSVs

CSVs are another great example of how people tend to log their data. They have a defined standard, and although far from efficient, do offer some human-readability to the user.
A metadata file for such a file could simply contain:
Delimiter: ','
Header: [ColA, ColB, ColC]

Harp binary files

The harp protocol is a well-defined binary protocol, with recommend logging practices. Similary to the binary file, a few metadata fields could suffice:
HarpBoard: [Behavior, AnalogInput, Lickometer, Olfactometer...]
Suffix: The suffix append to the file generated from each register

Json Files

This one is a bit more obvious, but saving a deserialized object that follows a pre-existing schema is super powerful, especially if one allows arrays of objects in a single file

SchemaSource: url to schema
IsMultiObject: bool

---

These are just some examples. Obviously, each of these logger objects should inherit from a base object that should have some core properties:

RootFolder
StreamName
FileName
FileExtension