Readme + docs improvements

[ngreenwald]

Is your feature request related to a problem? Please describe.
Some improvements to documentation that I think would be helpful to make it more readable/understandable

1. Add a small example dataset. This could be a subset of the ark example dataset, or if it's easier to just stick the files in the repo and not deal with huggingface, that works too. Main thing would be to have 10 or so FOVs in the cell table, the corresponding compartment masks, and the metadata file. We should pick FOVs that represent the various ways FOVs can be linked together in the metadata file (i.e. multiple FOVs from same timepoint, multiple timepoints per patient, a patient with only a single image, etc) so people can see how that's set up.
2. Provide an explanation for expected formatting of cell table and metadata file. This will describe the columns in the example files, but also explain which columns need to be present, what the naming requirements are, the relationships between the columns.
3. Provide more guidance anytime the user has to change column names. For example, what should go in the markers list? What should go in the cell_data_columns list? Basically make this code look like the code in the Feature Generation section, so that the stuff that needs to be changed is obvious, and everyting else that can just be plugged in and run at once is together in one area.
4. Add statements in readme for what is on the roadmap. i.e. in compartment section, say that the option to generate custom masks, as well as take pre-existing masks and assign cells to them, is in the works.
5. We should summarize the types of features that get generated by spacecat, rough description of what they mean, and which ones are still being ported over.

If you're not sure about the wording/verbiage for any of these, feel free to just add a placeholder section in readme once we open this PR and I can fill in.