Carrot is a web3 protocol trying to make incentivization easier and more capital efficient.

Carrot DefiLlama answerer

This package implements a daemon in Rust that handles the lifecycle of DefiLlama oracles created in Carrot.

Getting started

The package is developed using Rust, so in order to do anything it's necessary to install the Rust toolchain on your machine.

In order to simply run the answerer in a sort of "dev" environment locally you need a configuration file stored wherever you want. The configuration file's format is the same as the .config.example.yaml file at the root of the repo. Take that file, copy paste it, and rename it to .config.yaml, changing the values that you want. By default, the data regarding contract addresses and deployment blocks should be correct, so you'd only need to change the other configs.

Important

The program will automatically create the database as specified in the Postgres connection string in the config (key db_connection_string), but it will do so only if the specified user and password combination are correct, so double check the Docker Compose configuration to set them right.

A note on the IPFS API and Postgres connection. For convenience the repo provides a Docker Compose configuration to quickly spin up a local Kubo (Go IPFS client) and Postgres instance exposing the IPFS API at port 5001 and the Postgres service at port 5432 of the host machine. If you decide to use the provided configuration, simply run docker compose up to bootstrap the IPFS node and the Postgres instance locally. The IPFS API endpoint to be used provided you used this solution would be http://127.0.0.1:5001, while the Postgres connections tring would be postgresql://user:password@127.0.0.1:5432/defillama-answerer. These are the default values of the provided .config.example.yaml.

Once the .config.yaml file is ready to be used and you've optionally bootstrapped the IPFS node and Postgres instances through Docker Compose, and assuming the file is named exactly .config.yaml and placed at the root of this repo, you can just compile and start the daemon by running:

CONFIG_PATH="./.config.yaml" cargo run

If the config was set up correctly, at this point you should see the daemon running smoothly.

An additional env variable LOG_LEVEL can be set to regulate which logs will be shown. Its value can be one of trace, debug, info, warn or error. It's also possible to hide logs from ethers by setting LOG_LEVEL to none,defillama-answerer=info

Testing with a local template playground

In most cases this program will be tested with a local Carrot template playground. A few tips can be given for a good developer experience in this case. In particular:

• Remember to start the playground before starting the answerer, as the answerer needs a local RPC URL, and the playground bootstraps a local node as part of its starting procedure.
• Remember to update your local file with the correct local RPC URL depending on the forked network on which the template is being set on.
• In cases where a lot of past blocks must be indexed from the local RPC, the local node's performance will be inevitably degraded, in some occasions even rendering it unusable in the playground when simulating transactions. In this case find out the current block number that the local node is pointing at (also printed in the program's startup logs) and set it in the checkpoints table in the database so that the program will detect the forked network to be completely synced. Having a forked chain with id CHAIN_ID, a block number BLOCK_NUMBER and the default connection string postgresql://user:password@127.0.0.1:5432/defillama-answerer you can easily update the checkpoint in the database by running psql postgresql://user:password@127.0.0.1:5432/defillama-answerer in your terminal followed by a INSERT INTO checkpoints (chain_id, block_number) VALUES (<CHAIN_ID>, <BLOCK_NUMBER>) ON CONFLICT (chain_id) DO UPDATE SET block_number = <BLOCK_NUMBER>; (remember the ending semicolon) in the Postgres prompt.
• It is also possible to run the answerer in dev mode setting the dev_mode configuration property to true in the .config.yaml file. By doing this, the indexing of previous blocks will be disabled, while the index will remain enabled for any future blocks.
• Local nodes such as Ganache work by default in "automining" mode, meaning that no new block is produced unless a transaction is processed or unless manually triggered. This is a problem because the answerer reacts on new block events in order to process active oracles, so it might seem in certain cases that the answerer has stopped processing events while in reality it's just waiting for a new block to come. Luckily, most local nodes (such as Ganache and Anvil) support the evm_mine RPC method, so in order to trigger a new block on a local node running on a certain port PORT you can simply execute curl -X POST --data '{ "method": "evm_mine", "params": [] }' http://localhost:<PORT> in your terminal.

Building a release binary

Building a release (i.e. optimized) binary is simple, just run:

cargo build --release

By default, the binary is placed under /target/release/defillama-answerer at the root of the monorepo. You can run this program as a standalone binary provided you have set the env variables described in the getting started section.

Building a Docker image

Building a Docker image is simple, just run:

docker build .

from the root of the repo