Name the usage a 51Did was created for, and take the raw byte surface off FodId

[jwrosewell]

Two changes to the same surface, in the order they were made.

1. Name the usage a 51Did was created for

FodId gains usage (None, NonMarketing, Standard, Personalized) and usageFromConsent, reading bits 0 to 2 and bit 3 of the flags byte. The usage bits are cumulative, so the accessor answers the highest usage granted rather than leaving callers to mask bits and misread every marketing identifier as non-marketing. Same accessor, same names, in all six 51Did packages. Found by the Trusted Server work, which must work only through this package.

2. Take the raw byte surface off FodId

This is issue 193, raised on all six packages so the surface stays the same in every language. Once every bit has a name, the raw byte and the offsets are the remaining way to get it wrong, so they go.

• The flags getter is removed. The byte stays as the private _flags, which the typed getters read and which the tests that pin the byte now read, because those tests exist to prove every bit survives the read and the typed getters cannot see bits 4 and 5.
• The hash getter, HASH_OFFSET and HASH_LENGTH are deleted.
• The nine layout constants move to a new internal/layout.js, which index.js does not export. package.json and remote_package.json gain an exports map so . and ./package.json are the only reachable subpaths, since without one the internal module would still have been reachable by path from outside. No package in this repository had an exports map before, and nothing here imports the 51Did package by name, so nothing is narrowed that was being used.
• The dateMinutes getter is removed. In this port date was already the same integer, being minutes since 2020 as the OWID JavaScript library represents it, so dateMinutes was a duplicate. The unsigned reading it forced moves into date, which keeps the behaviour for a date past the signed range.
• A new test asserts that no removed name is reachable on the class or on the package entry point.

The open question, decided

The issue left the raw date minutes as a candidate rather than a definite. It goes, in all five packages that have it, and Node is the case where it was a duplicate of date rather than a raw encoding beside a typed one.

Regenerated type declarations

Rebuilding types/ for this package caught three things that had been changed in source and never regenerated: Usage was not exported and had no declaration file, usage and usageFromConsent were missing from the FodId declaration, and the client declaration lacked the MISCONFIGURED and INVALID_DATE outcomes. Those are fixed here. Only this package was rebuilt, so the stale declarations in the other four are left for their own change.

The specification

The byte structure is now specified once for all six languages, on specifications 25, at Identifier layout with the surface each package exposes at Package surface. The class documentation and the README link there instead of repeating the offset table, and the README gains a table of every removed name and its replacement. Those links resolve once that pull request merges.

Verification

npx jest fiftyone.pipeline.did --testPathIgnorePatterns ".*integration.*"
Test Suites: 3 passed, 3 total
Tests:       144 passed, 144 total

npx eslint . --ext .js
106 problems (0 errors, 106 warnings)

npx tsc -b fiftyone.pipeline.did --force
(no output, exit 0)

The warning count was 112 before this change and is 106 after, so none is new. The integration test needs a resource key and a live cloud, so it was not run. The offline example runs and prints the usage.

Closes #193

Produced with AI assistance and needs human review.