feat: track unused fields

[JoviDeCroock]

Resolves #143

Summary

When we are dealing with co-located fragments and operations it can be hard to determine when a field isn't used, especially as your component grows. This tries to prevent that by for queries and fragments implementing basic tracking of a result obtained from either of the following ways

• const data = useFragment() || const data = await client.query()
• const [result] = useQuery() --> urql
• const { data } = useQuery() --> Apollo-client
• const { data: { pokemon } } = useQuery();
• const { name } = useFragment()
• const [{ data }] = useQuery()

I still need to add tests for this whole ordeal and clean up the code as we can probably simplify this by combining both the crawling steps

When people are missing a fragment in a sub-component like

const Q = graphql(`query { pokemons { id name fleeRate } }`)

const [result] = useQUery(Q)

return result.data.pokemons.map(x => <Pokemon key={x.id} pokemon={x} />

Then it will show name and fleeRate as unused as we expect this to be part of the fragment exported by Pokemon.

Notes

We don't track mutations/subscriptions due to how they might be getting used in the normalised cache updaters/...

---

We don't resolve fragments as we make the assumption that those come from external components so we won't tell you that a fragment/... is unused

[changeset-bot]

🦋 Changeset detected

Latest commit: a07d830

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@0no-co/graphqlsp	Minor

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[TheMightyPenguin]

question: why do these start at 52k? (just curious)

[JoviDeCroock]

No particular reason, hoping for no clashes 😅 if tsserver ever reliably allows us to call into the code-fixes we could use these to i.e. add the missing fragment import, remove the unused fields, replace the deprecated field if we find a use instead, ...