concourseopen / defi-pulse-adapters Goto Github PK

The validate script fails due to the category of some projects not matching the categories[] expected by the test.

There's also the misspelled 'categoy' but i'll open a pull for that.


./testing/validate/export.js

  const categories = [
    'derivatives',
    'dexes',
    'lending',
    'payments',
    'assets'
  ]
...
it('categoy matches one of the defined options', () => {
        chai.expect(project.category).to.be.oneOf(categories);
});
...

• connext uses 'Lending'
• curve uses 'DEXes'
• dydx uses 'Lending'
• kyber uses 'DEXes'

The confusion is being created due to the _template project having a comment mentioning the incorrect values.

./projects/template/index.js

module.exports = {
...
 category: 'assets'  // allowed values as shown on DefiPulse: 'Derivatives', 'DEXes', 'Lending', 'Payments', 'Assets'
...
  }

Description

Having a package-lock.json and a yarn.lock file at the same time is not recommended:

We highly recommend you to delete the package-lock.json file if you decide to use yarn in order to avoid future confusion and possible consistency issues.

Solution

Delete either the package-lock.json or the yarn.lock file.

Tests fail at npm run test due to run.js's: output = (await sdk.api.util.unwrap({balances: output, block: point.block})).output;

async function Run(runFunction, project, timeUnit = 'day', timeOffset = 0) {
    try {
      let timestamp;

      if(typeof timeUnit == 'number') {
        timestamp = timeUnit;
      } else {
        timestamp = moment().utcOffset(0).startOf(timeUnit).add(timeOffset, timeUnit).unix();
      }

      let point = await sdk.api.util.lookupBlock(timestamp);

      await sdk.api.util.resetEthCallCount();
      let output = await project[runFunction](point.timestamp, point.block);
      console.log('RUN before', output) // <----- this returns the normal expected data shape: { [tokenAddress: string]: BalanceAsString }
      if(runFunction == 'tvl') {
        output = (await sdk.api.util.toSymbols(output)).output;
        output = (await sdk.api.util.unwrap({balances: output, block: point.block})).output;
      }
      console.log('RUN after', output)
      let ethCallCount = await sdk.api.util.getEthCallCount();

      return {
        ...ethCallCount,
        ...point,
        output
      }

    } catch(error) {
      console.log(error);
    }
  }

the console log console.log('RUN after', output) returns:

RUN after {
  ETH: { balance: '2220.846437', id: 1027 },
  USDT: { balance: '605155.984481', id: 825 },
  TUSD: { balance: '10006.044458', id: 2563 },
  USDC: { balance: '3101071.305708', id: 3408 },
  PAX: { balance: '5093.068189', id: 3330 },
  GUSD: { balance: '142.37', id: 3306 },
...
}

which breaks inside test-run.js here:

_.each(projectRun.output, (value, symbol) => {
            let balance = BigNumber(value).toNumber();   // <----- "value" here is the object above e.g { balance: '2220', id: 1000 } 
            console.log(symbol, balance, value)
...

Obviously this fails as BigNumber({ balance: '2220.846437', id: 1027 }) breaks.
Btw, the log above console.log(RUN before, output) returns data as expected:
{
[tokenAddress: string]: string
}

Any ideas?
ok it's the unwrap function:
output = (await sdk.api.util.unwrap({balances: output, block: point.block})).output;

Which returns the object for each token symbol

No CSVs are generated :(

There seems to be a significant discrepancy between DeFi Pulse and btconethereum.com regarding the number of BTC on Ethereum. DeFi Pulse is consistently reporting a larger amount.

Right now DeFi Pulse reports 109K BTC, of which all are on Ethereum expect for 1.1K BTC on the Lightning Network. That's a total 107.9K BTC on Ethereum. However, btconethereum.com reports 100.8K BTC. Where does the 7.1K BTC discrepancy come from?

Because btconethereum.com pulls its data straight off the smart contracts' totalSupply functions my best guess is that DeFi Pulse has a double-counting bug.

(Apologies if this is the wrong place to submit this issue.)

Here's spreadsheet @wir made which shows how to calculate the TVL without double-counting.

Some discussion here and here.

I'll be looking into this...