Spike: Move opportunities to react-query

[gomesalexandre]

Overview

We currently have THORChain LP opportunities as react-queries.
These have been implemented greenfield and not abstracted the same way other opportunities are.

This PR is twofold:

1. Move away from RTK to react-query for all opportunities, to benefit from the advanced caching, refetching, invalidation etc that react-query provides
2. Make the THORChain LP opportunities fit the opportunities shape. This means that the opportunities slice will have to support:

• the notion of asymmetrical or symmetrical deposits
• the notion of automagical network rebalancing, i.e deposit asymmetrical, but the pool becomes symmetrical under the hood

Things to consider in this spike to validate whether this is viable:

• Can we keep selection performant with reselect? Since things aren't stored in the redux store when using react-query, derivation is typically done using the select method from queries, but nothing blocks us from using reselect here - instead of selecting from the store, it will select from the return data of queries. We could potentially abstract react-query selectors away as their own domain, and use different selectors for a same query, depending on our needs
• Can we/should we persist react-query data? According to https://tanstack.com/query/v4/docs/framework/react/plugins/persistQueryClient, this should be possible, though there may be implications here

References and additional details

• opportunitiesSlice
• react-query THORChain LP queries

Acceptance Criteria

• We validate we can make the opportunities work without regressions using react-query
• The current THORChain LP data now works with the opportunities abstraction

Need By Date

N/A

Screenshots/Mockups

No response

Estimated effort

No response