Mlabs: Airdrop + database input file generation

[samuelWilliams99]

Following the Sendings module created to verify Sendings onchain, this task is to generate the input files for the airdrop and database from the PrivateSale type defined below:

type Amount = Natural

-- Invariants
-- Σ percentages = 100%
-- Description : Represent Vesting Tranches (Time Sequential and contiguous)
newtype Tranches = Tranches (NonEmpty Tranche)

data Tranche 
  = Tranche
    { percentage :: Integer -- out of 10,000
    , duration :: Integer -- number of slots
    }

data PrivateSale
  = PrivateSale
    { start :: POSIXTime
    , tranches :: Tranches
    , assetClass :: AssetClass
    , investors :: NonEmpty PrivateInvestor
    }

data PrivateInvestor
  = PrivateInvestor
    { address :: Address
    , allocation :: Amount
    }

This task is the following:

• Create a CLI command that takes a filepath
• Decode the file at the path into a PrivateSale (using derived JSON instances), throwing a reasonable error on failure to parse (see Sendings)
• Verify the commented requirement on Tranches (Sum of Tranche %'s must be 100% (10,000))
• Generate the list of (NativeScript, Integer) pairs for how much of PrivateSale.assetClass to send to each script.
  • The Scripts are generated like so - For each investor, split the allocation according to the tranches, with unlock time starting at posixTimeToSlot PrivateSale.start + PrivateSale.tranches[0].duration, then adding the duration of each subsequent Tranche to the unlock time.
  • Use the NativeScript type defined in the database spec:

    data NativeScript = NativeScript 
      { pkh :: String 
      , unlockTime :: Integer 
      }

  • Example: Say we have tranches: [(50%, 100 slots), (25%, 100 slots), (25%, 200 slots)], with a start time equilalent to slot 1000. For a user with 10000 tokens, we would send 5000 to a script that unlocks at slot 1100, 2500 to one that unlocks at 1200, and 2500 to one that unlocks at 1400.
• Using the above, generate the following files:
  • Airdrop input file, using the datatype and json instances defined here: https://github.com/smart-chain-fr/tokenomia/blob/main/src/Tokenomia/TokenDistribution/Distribution.hs#L47
  • Database input file, as a json file containing Map Address [LockedFund], defined in the database spec.
  • The above file paths may be fixed, or inputs to the script (but with defaults), depending on time.

Create your branch and PR from mlabs/private-sale-staging

[Luis-omega]

Should I be worried about the sequence of PrivateInvestors ?

For example, if we have :

[PrivateInvestor 2 23 , PrivateInvestor 3 14 , PrivateInvestor 2 99]

With the described approach the PrivateInvestor with address 2 would pay its part for every currency of it in the list, is this intended or should I reject such a list? If this is intended, should I merge all the occurrences or just leave them as is (this would impact the Airdrop table output)?

[samuelWilliams99]

Let's merge, probably as a first pass to have them unique by address. You could even convert it to a Map Address Amount to be correct at a type level.