[AERIE-1906] Resolve SerializedValue LGTM errors

[pcrosemurgy]

• Tickets addressed: AERIE-1906
• Review: By commit
• Merge strategy: Merge (no squash)

Description

From #220.

Resolves LGTM SerializedValue errors reported here.

Verification

All tests passing.

Documentation

None.

Future work

Resolve LGTM warnings (see https://lgtm.com/projects/g/NASA-AMMOS/aerie/alerts/?mode=tree).

[Twisol]

Sorry... GitHub notifications never actually tell me a PR is just a draft. 🤦‍♂️

[pcrosemurgy]

Sorry... GitHub notifications never actually tell me a PR is just a draft. 🤦‍♂️

No worries! All solid feedback :) Thanks

[pcrosemurgy]

Opened this PR prematurely. Am investigating the e2e-tests failure.

[skovati]

LGTM! (heh) Smart move refactoring SerializedValue into an interface.

Should we create a separate ticket to capture the string escape work or were you planning on implementing that in this PR?

[adrienmaillard]

Looks good, thank you.

[pcrosemurgy]

Should we create a separate ticket to capture the string escape work or were you planning on implementing that in this PR?

Good point! I'll check in with Adrien on this, the preexisting TODO notes about string escaping should probably be removed in place of a ticket.

[dyst5422]

Why are we doing sanitization for GQL instead of using variables? That's one of the great value propositions of the GraphQL standard - built in sanitization.

[pcrosemurgy]

Sure! I'll change the wording of the ticket.

[dyst5422]

I don't think my comment is scoped to this ticket as the string formatting predates this ticket. Specifically looking at lines like

aerie/scheduler-server/src/main/java/gov/nasa/jpl/aerie/scheduler/server/services/GraphQLMerlinService.java

Lines 137 to 145 in 88d926a

	"query getPlanMetadata { "
	+ "plan_by_pk( id: %s ) { "
	+ " id revision start_time duration "
	+ " mission_model { "
	+ " id name version "
	+ " uploaded_file { name } "
	+ " } "
	+ " simulations(limit:1, order_by:{revision:desc} ) { arguments }"
	+ "} }"

where we were already doing string templating instead of using variables

[pcrosemurgy]

Yup! But your point 100% still stands, we should use variables. https://jira.jpl.nasa.gov/browse/AERIE-1949 should address this.