Reorganizing repo

[danvk]

General principles:

• Keep data and code separate. Data lives in data, code lives in oldnyc.
• Original sources of data go in data/originals. Derived data lives in data. This line can be a bit fuzzy.
• Code lives under coherent subdirectories in oldnyc:
  • ingest: code used to produce images.ndjson
  • crop: finding and extracting photos and text from larger images
  • geocode: locating items
  • ocr: transcribing text from the ocrbacks
  • feedback: processing user feedback
  • site: generating the OldNYC site in oldnyc.github.io
  • analysis: code that is not part of the pipeline
• One-off analysis scripts and notebooks all live under oldnyc/analysis.

Apparently __init__.py is no longer a thing. Also one does not name a file types.py. pyright and the e2e tests were extremely helpful in putting this together.

Working spreadsheet here.