@lidofinance/rpc

Fetch extension to use it with RPC endpoints.

Installation

yarn add @lidofinance/rpc

Getting started

You can just use it as normal fetch, there are just a few types restrictions, and it will set right HTTP method for you.

fetchRPC(url, { body: req.body })

Examples

Caching

Here is an example how you can extend fetchRpc by adding caching

import { FetchRpc, fetchRpc } from '@lidofinance/rpc'
import { Cache } from 'memory-cache'

export type CachedFetchRPC = FetchRpc<{ cacheKey: string }>

// Just a sample how to use extend fetchRPC
export const cachedFetchRpcFactory = (timeout: number): CachedFetchRPC => {
  const cache = new Cache<string, Response>()

  return async (url, init, { cacheKey }) => {
    const cachedValue = cache.get(cacheKey)
    if (cachedValue != null) {
      return cachedValue
    }
    const response = await fetchRpc(url, init)
    cache.put(cacheKey, response, timeout)
    return response
  }
}

iterateUrls fallback callback

iterateUrls(urls, callback, onError?) iterates provider URLs until one succeeds. The optional onError handler fires when an attempt fails — either the callback threw, or it returned a non-ok Response. The second argument carries context so you can log structured fallback transitions.

import { iterateUrls } from '@lidofinance/rpc'

await iterateUrls(
  providers,
  (url) => fetch(url),
  (error, { failedUrl, nextUrl, attempt, response }) => {
    if (nextUrl) {
      logger.warn(`RPC ${failedUrl} failed (attempt ${attempt}), switching to ${nextUrl}`)
    } else {
      logger.error(`All RPC providers failed, last one was ${failedUrl}`)
    }
  },
)

Handlers with a single-argument signature ((error) => ...) keep working — the extra context argument is simply ignored.