oraichain/oraichain-price

AI Oracle background

• The AI Oracle protocol developed by Oraichain allows users to run AI services such as AI price feed, face recognition, or OCR in a decentralized manner.

• Each service contains several data source scripts, which fetch external AI APIs to get the result.

• The Oraichain's validators execute the AI requests, aggregate, and return the results, which are already verified by test cases.

• Third-party test case providers are responsible for the verification process and are rewarded accordingly.

• Users can select any aggregated result provided by any validator for their services.

AI price feed integration process

• To continuously and automatically request and collect tokens' prices using Oraichain, it is recommended to use Cosmosjs powered by Oraichain to easily interact with the network.

• There are three steps to finally retrieve the tokens' prices:

1. Prepare the .env file

The .env file contains several key-value pairs below:

MNEMONIC=
LCD_URL=http://lcd.orai.io
CHAIN_ID=Oraichain
REQUEST_FEES=0
TX_FEES=0
AIORACLE_ADDR=orai16w8yw68wqfgdxvfn3hhspmj6452p3rvzjgqvx3
INTERVAL=240000
QUERY_SLEEP_INTERVAL=10000
QUERY_COUNT=20

where:

• MNEMONIC is your Oraichain mnemonic used to create AI price feed requests. Without this field, you will not be able to create an AI price feed request.

• LCD_URL is the Oraichain network's team endpoint used to query network data or broadcast transactions.

• CHAIN_ID is Oraichain chain id.

• REQUEST_FEES is the fees (in uorai) you need to pay per AI price feed request. Such fees will go directly toward the validators' and providers' wallets. At the moment, the validators charge no fees when creating an AI price feed request, so it is not necessary to specify the value of this key. Example when adding a value: REQUEST_FEES=10, where 10 is 10uorai = 10 * 10^-6 ORAI.

• TX_FEES is the transaction fees (in uorai) of your requests. At the moment, all mainnet transactions cost no fees,

• AIORACLE_ADDR is the AI price feed Oracle smart contract address.

• INTERVAL is the interval (in ms) for creating new AI price feed requests. Reducing this value will create more AI price feed requests within a period

• QUERY_SLEEP_INTERVAL is the sleep interval (in ms) when waiting for the results of an AI price feed request. Decreasing this value will reduce the time waiting for the results to come.

• QUERY_COUNT is the number of times you query the results of an AI price feed request. Decreasing this value will also reduce the time waiting for the price feed's results.

• COUNT is the validator count. The higher the value, the larger the number of validators needed to execute your request.

Using these pairs, you'll be able to manage your request time as well as the tokens you are willing to pay per request.

2. Create a new AI request and collect its request ID

Creating a new AI price feed request requires a certain number of validators involving in the process. At the moment, you can choose which validator to execute your request:

const validators = await init();
const validatorList = randomValidators(validators, process.env.COUNT ? parseInt(process.env.COUNT) : 2);

where the init() function will initialize the cosmosjs object required to broadcast transactions and fetch the current list of Oraichain validators.

const init = async () => {
  global.cosmos = new Cosmos(process.env.LCD_URL, process.env.CHAIN_ID);
  cosmos.setBech32MainPrefix('orai');

  const validators = await cosmos.get(`/cosmos/staking/v1beta1/validators`);
  let validatorsActive = [];
  for (let val of validators.validators) {
    if (!val.jailed && val.status === "BOND_STATUS_BONDED") validatorsActive.push(cosmos.getAddressStr(val.operator_address))
  }
  return validatorsActive;
}

next, create a new AI request:

const aiOracleAddr = process.env.AIORACLE_ADDR;
await setAiRequest(aiOracleAddr, validatorList);

The function will create the price feed message and use your mnemonic to sign the AI price feed request transaction

After creating a new AI request, collect its request ID:

const res = await cosmos.submit(childKey, txBody, 'BROADCAST_MODE_BLOCK', isNaN(process.env.TX_FEES) ? 0 : parseInt(process.env.TX_FEES), 5000000);
console.log(res);
// collect request id and prepare query to get results
let requestId = res.tx_response.logs[0].events[1].attributes[2].value;

3. Query the request's result using the corresponding request ID

await queryRequestResult(parseInt(requestId), 0);

The function will run in interval trying to fetch the results of the provided request id. It will throw an error after sufficient retries. Behind the scene, it creates a query message and invokes a GET request to the Oraichain network.

let queryMsg = JSON.stringify({
    oracle_query: {
      msg: {
        get_request: {
          request_id: requestId
        }
      }
    }
  })

let { data } = await cosmos.get(`/wasm/v1beta1/contract/${process.env.AIORACLE_ADDR}/smart/${Buffer.from(queryMsg).toString('base64')}`);

when the status of the request is true, it means that the request has finally collected all results from the list of validators. The results are also ready to be extracted and used for your services.

if (data.status) {
  console.log("result reports: ", data.reports);
  let aggregatedResult = "";
  // collect the first report
  for (let report of data.reports) {
    if (report.status) {
      aggregatedResult = Buffer.from(report.aggregated_result, 'base64').toString('ascii');
      break;
    }
  }
  // if all reports have false status => something wrong, must throw error
  if (aggregatedResult === "") throw "Cannot collect any valid aggregated result from the validators";
  requestOutput.write(`timestamp: ${new Date()}\nPricefeed aggregated result of request id ${requestId}: ${util.format(aggregatedResult)}\n\n`);
  return;
}

Demo

Type: NODE_ENV=production yarn start to start running the script. The request's results will be written into the output.log file, while all errors are put in the debug.log file.

With this provided template nodejs script, you can extend and add more functionalities for your features such as a server with a database storing all the price feed results.