TODO:

• Add all the options for PDF export (landscape, size)
• Add tests for the new endpoints
• Should PDF also return an iterable-stream?
• How do we feel about csv returning an iterable-stream?
• How do we feel about calling 'data' > 'csv' on the model?
• Clean up error strings (They're find/replace right now)
• Do we want to rename image to png?
• I called 'type' for the pdf size 'output_size' but open to better names. I don't like type

/cc @RussTheAerialist