Don't add undefined values to objects when parsing JSON

[benchristel]

Summary:

Previously, when parsing typed objects from JSON, we would add
[key]: undefined for any fields that were specified in the parser schema but
were absent from the JSON. This was a difference from the behavior of
JSON.parse() that could cause bugs in some cases (e.g. if a parsed
object was spread into another object, it could overwrite existing
values with undefined when that was not intended).

Thus, to better match the behavior of JSON.parse(), we now omit
undefined values from the parsed object if they are not present on the
original object.

This PR was inspired by test failures on https://github.com/Khan/webapp/pull/28551.
I'm making this change out of an abundance of caution.

Issue: LEMS-2774

Test plan:

yarn test

[github-actions]

npm Snapshot: Published

Good news!! We've packaged up the latest commit from this PR (a37b5c0) and published it to npm. You
can install it using the tag PR2127.

Example:

yarn add @khanacademy/perseus@PR2127

If you are working in Khan Academy's webapp, you can run:

./dev/tools/bump_perseus_version.sh -t PR2127

[github-actions]

Size Change: +20 B (0%)

Total Size: 1.47 MB

Filename	Size	Change
packages/perseus/dist/es/index.js	395 kB	+20 B (+0.01%)

ℹ️ View Unchanged

Filename	Size
packages/kas/dist/es/index.js	39 kB
packages/keypad-context/dist/es/index.js	760 B
packages/kmath/dist/es/index.js	86.8 kB
packages/math-input/dist/es/index.js	77.6 kB
packages/math-input/dist/es/strings.js	1.79 kB
packages/perseus-core/dist/es/index.js	27.1 kB
packages/perseus-editor/dist/es/index.js	689 kB
packages/perseus-linter/dist/es/index.js	22.2 kB
packages/perseus-score/dist/es/index.js	113 kB
packages/perseus/dist/es/strings.js	5.1 kB
packages/pure-markdown/dist/es/index.js	3.66 kB
packages/simple-markdown/dist/es/index.js	12.5 kB

_{compressed-size-action}