Add top gainers losers to token discovery service

[Bigshmow]

Explanation

We're introducing methods that allow consumers to fetch top-gainers and top-losers from the portfolio api (ultimately via the Moralis service).

I've also consolidated some of the types in this PR so they are reusable and provided a base type inheritance.

References

Changelog

@metamask/package-a

• Added: param types inherit from a ParamsBase

• Added: getTopGainersByChains to the TokenDiscoveryApiService

• Added: getTopLosersByChains to the TokenDiscoveryApiService

• Changed: - BREAKING: Renamed TokenTrendingResponseItem name to MoralisTokenResponseItem

  • Consumer will get ts compilation errors after upgrading until they update their imports

Checklist

• I've updated the test suite for new or updated code as appropriate
• I've updated documentation (JSDoc, Markdown, etc.) for new or updated code as appropriate
• I've highlighted breaking changes using the "BREAKING" category above as appropriate
• I've prepared draft pull requests for clients and consumer packages to resolve any breaking changes

[GuiBibeau]

Just leaving a few proposals. Code looks great 👍

[GuiBibeau]

Do we need to validate the response in any way or not?

[Bigshmow]

I agree with the validation but how do you feel about revisiting this in portfolio api first since it's closest to the server, and bad shapes would never even reach this point?
(this gives more credence to refactoring the error handling in the core repo you mention elsewhere)

[GuiBibeau]

Nitpick: this seems unrelated to a vanilla Response object and looks to be only the tokenItem. Adding Response into the name confuses me a bit and could lead people to think this includes status and other properties of an HTTP Response

https://developer.mozilla.org/en-US/docs/Web/API/Response

[Bigshmow]

Recommend dropping the ResponseItem portion all together to simply be MoralisToken? I like that too.

[mcmire]

What are your thoughts on adding more test coverage? When the thresholds are at 100, it makes it really easy to know what's untested because you can just look at the output of yarn test. When it's below 100, however, you end up having to guess or open the coverage report and manually search, which ends up being annoying.

[Bigshmow]

I saw that a majority of packages here have their coverage set to 100. I think a more pragmatic 60-70% strikes a good balance between confidence and effort. Would you be willing to compromise on a level under 100 that still gives you confidence or is the 100% a dogma in your bones!