Process needs to: 1. collate: - m/s models-generated fields - nara catalog fields - aperitiif endpoints - ocr text 2. transform above data and export to stable JSON documents per schemas 3. lint/validate docs to schemas Contingent on [#12](https://github.com/migrants-and-the-state/reading-room/issues/12) and #42
