Feat: Store swaps

[bvotteler]

Resolves #106

Store swap information:

• From / To:
  • account
  • asset
  • amount
  • amountHuman
• Fees paid
• Timestamp

---

Todo before ready for review:

• [x] Run against Kintsugi mainnet and check if results look plausible
• ❌ (attempted, too slow) Attempt to run against Kintsugi testnet - if performance allows it
• [x] Regenerate db migration files after last fixes/changes have been made

[bvotteler]

Sample query:

fragment AmountFields on PooledAmount {
  amount
  amountHuman
  token {
    ... on NativeToken {
      __typename
      token
    }
    ... on ForeignAsset {
      __typename
      asset
    }
    ... on StableLpToken {
      __typename
      poolId
    }
    ... on LendToken {
      __typename
      lendTokenId
    }
  }
}

query SwapQuery {
  swaps(limit: 10) {
    id
    timestamp
    fromAccount
    toAccount
    feeRate
    from {
      ...AmountFields
    }
    to {
      ...AmountFields
    }
  }
}

Sample result:

      {
        "id": "swap_(KSM,KBTC)_0002681725-000012-2842a",
        "timestamp": "2023-03-23T03:01:36.212000Z",
        "fromAccount": "a3dkY3yWweLMkUjbSSNC2NawWegH3NZkWpYjsnBr2srQsp9LZ",
        "toAccount": "a3dkY3yWweLMkUjbSSNC2NawWegH3NZkWpYjsnBr2srQsp9LZ",
        "feeRate": "0.0015",
        "from": {
          "amount": "21370527482",
          "amountHuman": "0.021370527482",
          "token": {
            "__typename": "NativeToken",
            "token": "KSM"
          }
        },
        "to": {
          "amount": "2538",
          "amountHuman": "0.00002538",
          "token": {
            "__typename": "NativeToken",
            "token": "KBTC"
          }
        }
      }

It looks a bit awkward to have the from/to account ids outside of the fields from and to which contain PooledAmount data (ie. no account in there, because it's also used for cumulative volumes).

I was in two minds whether we should introduce a new type including the account ids, or keep it as is. Slightly leaning towards keeping it so we can better reuse fragments when structuring queries. Having said that, I am more than happy to be challenged on that.