Firefly Routing API

This repository contains routing API for the Firefly V3 protocol.

It deploys an API to AWS that uses @firefly-dex/smart-order-router to search for the most efficient way to swap token A for token B.

Development

To develop on the Routing API you must have an AWS account where you can deploy your API for testing.

Deploying the API

The best way to develop and test the API is to deploy your own instance to AWS.

1. Install and configure AWS CLI and AWS CDK V1.
2. Create .env file in the root directory of the project with :

   THROTTLE_PER_FIVE_MINS = '' # Optional
   JSON_RPC_PROVIDER_{CHAIN ID} = { RPC Provider}
   # RPC Providers must be set for the following CHAIN IDs:
   # MAINNET = 1
   # ROPSTEN = 3
   # RINKEBY = 4
   # GOERLI = 5
   # KOVAN = 42
   # OPTIMISM = 10
   # OPTIMISTIC_KOVAN = 69
   # ARBITRUM_ONE = 42161
   # ARBITRUM_RINKEBY = 421611
   # POLYGON = 137
   # POLYGON_MUMBAI = 80001
   # BNB = 56
   TENDERLY_USER = '' # For enabling Tenderly simulations
   TENDERLY_PROJECT = '' # For enabling Tenderly simulations
   TENDERLY_ACCESS_KEY = '' # For enabling Tenderly simulations

3. Install and build the package

   npm install && npm run build

4. To deploy the API run:

   cdk deploy RoutingAPIStack

   This will deploy to the default account your AWS CLI is configured for. Once complete it will output something like:

   RoutingAPIStack.Url = https://...

   You can then try it out:

   curl --request GET '<INSERT_YOUR_URL_HERE>/quote?tokenInAddress=0xC02aaA39b223FE8D0A0e5C4F27eAD9083C756Cc2&tokenInChainId=1&tokenOutAddress=0x1f9840a85d5af5bf1d1762f925bdaddc4201f984&tokenOutChainId=1&amount=100&type=exactIn'

Tenderly Simulation

1. To get a more accurate estimate of the transaction's gas cost, request a tenderly simulation along with the swap. This is done by setting the optional query param "simulateFromAddress". For example:

curl --request GET '<INSERT_YOUR_URL_HERE>/quote?tokenInAddress=<0x...>&simulateFromAddress=<FROM_ADDRESS>&...'

2. Tenderly simulates the transaction and returns to us the simulated gasLimit as 'gasUseEstimate'. We use this gasLimit to update all our gas estimate heuristics. In the response body, the

{'gasUseEstimate':string, 'gasUseEstimateQuote':string, 'quoteGasAdjusted':string, and 'gasUseEstimateUSD':string}

fields will be updated/calculated using tenderly gasLimit estimate. These fields are already present even without Tenderly simulation, however in that case they are simply heuristics. The Tenderly gas estimates will be more accurate.

3. If the simulation fails, there will be one more field present in the response body: 'simulationError'. If this field is set and it is set to true, that means the Tenderly Simulation failed. The

{'gasUseEstimate':string, 'gasUseEstimateQuote':string, 'quoteGasAdjusted':string, and 'gasUseEstimateUSD':string}

fields will still be included, however they will be heuristics rather then Tenderly estimates. These heuristic values are not reliable for sending transactions on chain.

Testing

Unit Tests

Unit tests are invoked by running npm run test:unit in the root directory. A 'watch' mode is also supported by running npm run test:unit:watch.

Integration Tests

Integration tests run against a local DynamoDB node deployed using dynamodb-local. Note that JDK 8 is a dependency of this package. Invoke the integration tests by running npm run test:integ in the root directory.

End-to-end Tests

The end-to-end tests fetch quotes from your deployed API, then execute the swaps on a Hardhat mainnet fork.

1. First deploy your test API using the intructions above. Then update your .env file with the URL of the API, and the RPC URL of an archive node:

   UNISWAP_ROUTING_API='...'
   ARCHIVE_NODE_RPC='...'

2. Run the tests with:

   npm run test:e2e