Clean up apd json

[antislice]

• Whole file: neaten up unicode codepoints, fix eligibility/audience field naming, add "notes" attribute, correctly parse basic urls
• First ~3000 lines: Double check data with original docx file, make manual edits & data rearranging
• Added a couple of tiny ruby scripts that helped with the data fixing, along with a shorter json file for testing

Diffing this version of open_ref.json and the previous one could show some places to improve the data importing. Some specific issues: grabbing URLs from the descriptions, variations in data structuring. The "Notes" field in the doc is usually what we want to match to the "how_to_apply" attribute.

[antislice]

Hey @aprilsteed @graysonwright, let me know if there's anything you'd like me to add to this.

[jbarbuto]

Thanks for your great work on this! @graysonwright gave verbal green light to merge.