Sanity.io + Serverless + Algolia

handler.js

const algoliasearch = require('algoliasearch');
const request = require('request');

const ndjson = require('ndjson');
const {bindNodeCallback} = require('rxjs');
const {streamToRx} = require('rxjs-stream');
const {bufferCount, map, mergeMap, toArray, tap} = require('rxjs/operators');

// Algolia configuration
const algoliaApp = process.env.ALGOLIA_APP_ID;
const algoliaIndex = process.env.ALGOLIA_INDEX;
// Sanity configuration
const projectId = process.env.SANITY_PROJECT_ID;
const dataset = process.env.SANITY_DATASET;
const sanityExportURL = `https://${projectId}.api.sanity.io/v1/data/export/${dataset}`;

module.exports.indexContent = function indexContent(event, context, cb) {
  // Initiate an Algolia client
  const client = algoliasearch(algoliaApp, process.env.ALGOLIA_TOKEN);
  // Initiate the Algolia index
  const index = client.initIndex(algoliaIndex);

  // bind the update function to use it as an observable
  const partialUpdateObjects = bindNodeCallback((...args) => index.saveObjects(...args));

  streamToRx(
    request(sanityExportURL).pipe(ndjson())
  ).pipe(
    /*
     * Pick and prepare fields you want to index,
     * here we reduce structured text to plain text
     */
    map(function sanityToAlgolia(doc) {
      return {
        objectID: doc._id,
        body: blocksToText(doc.body || []),
        blurb: blocksToText(doc.blurb || []),
        title: doc.title,
        name: doc.name,
        slug: doc.slug,
      };
    }),
    // buffer batches in chunks of 100
    bufferCount(100),
    // 👇uncomment to console.log objects for debugging
    // tap(console.log),
    // submit actions, one batch at a time
    mergeMap(docs => partialUpdateObjects(docs), 1),
    // collect all batches and emit when the stream is complete
    toArray()
  )
    .subscribe(batchResults => {
      const totalLength = batchResults.reduce((count, batchResult) => count + batchResult.objectIDs.length, 0);
      cb(null, `Updated ${totalLength} documents in ${batchResults.length} batches`);
    }, cb);
};

const defaults = {nonTextBehavior: 'remove'};

function blocksToText(blocks, opts = {}) {
  const options = Object.assign({}, defaults, opts)
  return blocks
    .map(block => {
      if (block._type !== 'block' || !block.children) {
        return options.nonTextBehavior === 'remove' ? '' : `[${block._type} block]`;
      }

      return block.children.map(child => child.text).join('');
    })
    .join('\n\n');
}

serverless.yml

service: algolia-sync

provider:
  name: aws
  runtime: nodejs8.10

plugins:
  - serverless-offline
  - serverless-offline-scheduler

functions:
  indexContent:
    handler: handler.indexContent
    events:
      - schedule: rate(1 day)
    environment:
      ALGOLIA_APP_ID: 'your_app_id_here'
      ALGOLIA_TOKEN: 'your_token_here'
      ALGOLIA_INDEX: 'your_index_name_here'
      SANITY_PROJECT_ID: 'your_sanity_project_id_here'
      SANITY_DATASET: 'your_sanity_dataset_here'