Feature request: rate limit middleware

[joscha]

This package should provide a rate limit middleware that respects the API rate limits

• throttle requests based on the actual free API requests that are sent with each response:

     "x-ratelimit-limit-org": "unlimited",
     "x-ratelimit-limit-org-remaining": "unlimited",
     "x-ratelimit-limit-org-reset": "2143843",
     "x-ratelimit-limit-user": "900",
     "x-ratelimit-limit-user-remaining": "863",
     "x-ratelimit-limit-user-reset": "26",

* automatic retry/requeue requests that failed because of a `429` response.

[joscha]

With a combination of V1 api and V2 api, the best workaround is currently:

import { Affinity } from "@planet-a/affinity-node/v1";
import {
  createConfiguration,
  helpers,
  PersonsApi,
} from "@planet-a/affinity-node/v2";
import assert from "assert";
import pThrottle from "p-throttle";

const {
  paginator: { paginated },
} = helpers;

const config = createConfiguration({
  authMethods: {
    bearerAuth: {
      tokenProvider: {
        getToken: async () => {
          assert(process.env.AFFINITY_API_KEY);
          return process.env.AFFINITY_API_KEY;
        },
      },
    },
  },
});

assert(process.env.AFFINITY_API_KEY);
const affinity = new Affinity(process.env.AFFINITY_API_KEY);
const {
  rate: {
    api_key_per_minute: { limit },
  },
} = await affinity.rateLimit.get();

const throttle = pThrottle({
  limit,
  interval: 60 * 1000,
});

const personsApi = new PersonsApi(config);
const personFn = personsApi.getV2Persons.bind(personsApi);
const throttledPersonFn = throttle(personFn);
const personIterator = paginated(throttledPersonFn);