Add LazyQuery Hook

[ZielonyBuszmen]

I added LazyQuery Hook binding, because I cannot find it in this repository.

[mbirkegaard]

Hi @ZielonyBuszmen,

Thank you for doing this! The lack of bindings to useLazyQuery has been a real weak spot of this package.

There are a few things that don't work with the useLazyQuery hook (see the docs here).

The signature of useLazyQuery on the ts side is

function useLazyQuery<TData = any, TVariables = OperationVariables>(
  query: DocumentNode,
  options?: LazyQueryHookOptions<TData, TVariables>,
): [
  (options?: QueryLazyOptions<TVariables>) => void,
  QueryResult<TData, TVariables>
] {}

• LazyQueryHookOptions<TData, TVariables> doesn't accept the skip parameter so neither should options here.
• The QueryResult<TData, TVariables> returned by useLazyQuery has a called field. That field should be exposed in the bindings and used to set an appropriate value for the simple variant when it's not true (e.g. NotCalled)
• The fetchData function on the ts side has this signature (options?: QueryLazyOptions<TVariables>) => void where

export interface QueryLazyOptions<TVariables> {
    variables?: TVariables;
    context?: Context;
}

so beware that the options passed to useLazyQuery are not the same as those passed to the fetchData function.

Then there's a few things to do for consistency with the rest of reason-apollo-hooks.

• This hook should also accept context. I don't think that it's particularly widely used, but the query and mutation hooks have it.
• useLazyQuery should be available from the top-level module (same as how ApolloHooks.re has definitions for useQuery, useMutation and useSubscription)

Regarding using useMemo to wrap fetchData, I think the intention is that it is callable multiple times even with the same variables (akin to refetching) so I don't think useMemo is appropriate here (and the memo comparison [|variables|] will only check referential equality anyway).

[mbirkegaard]

Just took a look at ApolloHooksMutation which wraps mutate in useMemo. I think that's a mistake for similar reasons as I mention above.