feat: add testing for plo endpoints

[TarikGul]

Added:

• Rococo Registry and metadata
• TypeFactory class
• Tests for the following methods
  • ParasService.crowdloansInfo
  • ParasService.crowdloans
  • ParasService.leaseInfo
  • ParasService.leasesCurrent
  • ParasService.paras
  • ParasService.auctionsCurrent

[emostov]

Overall it looks good, but some minor grumbles

[emostov]

auctions::AuctionInfo is a Option<(LeasePeriodOf<T>, T::BlockNumber)> so Option<[beginEnd, leasePeriodIndex]>. I don't see the Option part here. In the service we have it modelled as Option<Vec<AbstractInt>> - should wrap this in an Option

[TarikGul]

Yea so this is actually wrapped in an Option inside of auctionInfoAt, but there were two methods using vectorAuction's so I didn't wrap it in an Option in the outer scope. But you raised a good point with a previous comment typeFactory.optionOf(null) so I fixed it all to be more on point, and now the variable are all inside of auctionInfoAt and don't need to be in the outer scope anymore.

Inside of TypeFactory we now have a nullOf, and that allows you too create a Null primitive type and pass it into Option. Good catch and great addition.

[emostov]

Should this be longer? like a length of 7 (Or whatever the SLOT_RANGE_COUNT is)? The rest of the values can just be Option::None, but I am concerned there could be behavior that relies on the length of auctions::Winning.

It is defined as [Option<(<T as frame_system::Config>::AccountId, ParaId, BalanceOf<T>)>; SlotRange::SLOT_RANGE_COUNT] where SLOT_RANGE_COUNT is the length it will always be.

[TarikGul]

So I couldnt find where it says the minimum SLOT_RANGE_COUNT is, but I know it's suppose to be a maximum of 10 per the older code about a week ago.

/// Total number of possible sub ranges of slots.
pub const SLOT_RANGE_COUNT: usize = 10;

This was the comment that they had for slot range.

Now the codes changed a bit which includes a generate_slot_range macro.

/// * A constant SLOT_RANGE_COUNT will count the total number of enum variants.

[emostov]

Lets hardcode it to 10 (I think we covered the rest in offline discussion.)

[TarikGul]

@emostov I restructured the tests so that they all reflect 'user contract' tests. They all use sanitizeNumbers as well as toMatchObject to mock an actual user response. I also added a typeFactory.nullOf for Null primitives that can be added to Option. Everything is now looking good, the only last piece is the Slot range constant that is a little confusing. Whats I guess a little confusing for me is the macro itself generate_slot_range!.

[emostov]

LGTM % nits

[emostov]

Suggested change

	it('Should return the correct null values when auctionInfo is not of type Option', async () => {
	it('Should return the correct null values when auctionInfo is `None`', async () => {

It should be of type Option: Option::None. In JSON though that gets converted into null.

[emostov]

The sanitizeNumber specs are exhaustive for this stuff. See

substrate-api-sidecar/src/sanitize/sanitizeNumbers.spec.ts

Lines 572 to 575 in ae74ef7

	it('converts None to null', () => {
	const none = kusamaRegistry.createType('Option<Text>', null);
	expect(sanitizeNumbers(none)).toBe(null);
	});

[TarikGul]

ahh gotcha, yea that makes sense. Empty Options are returned as null with sanitize numbers :)!

[emostov]

Can you just pass in null here instead of creating this more complicated polkadot-js null thing?

[TarikGul]

Well no is the simplest answer. The way its implemented to in polkadot/api is to create a emptyOption method. Which is what I just added. It's definitely better than adding the nullOf method as you mentioned.

public emptyOption = <T extends Codec>(typeName: keyof InterfaceTypes): Option<T> => 
        new Option<T>(this.#registry, typeName);

And then we just add the Null primitive as the typeName

[emostov]

Lets hardcode it to 10 (I think we covered the rest in offline discussion.)