RedisSearch! An option worth considering for a search engine

Problem

Your system needs to help users find the data they want. For example, a movie website allows users to search for movies by name, author, actor, etc.

The first thing that comes to mind is using SQL's LIKE. It is a good and fast approach to problem solving. But overall it's not really good. We want more than that. Such as full-text search support, search query suggestions, etc., it is time to implement a search engine system.

There are many famous engine support: ElasticSearch, Solr, ... And there's a new guy who's also quite strong.

That is RedisSearch. even though it just appeared recently but he is good choice to consider by those things:

simple syntax, easy to approach.
does not require a heavy infrastructure to maintain (like ElasticSearch for example)
the search speed is quite fast
especially if you are using Redis for other features like Cache, Queue, ... Then implementing more search using Redis is great

Let's start exploring RedisSearch

RedisSearch ?

For ease of understanding, RedisSearch is an extension module of Redis. It's support full-text search, suggest keywords, so on.

verbose info

Flow implement RedisSearch

Above is simple schema of RedisSearch integration. There are a few points to note:

It is related to the fields you need to search in the future. As the name of the video, actors,...
Need to build pusher and querier:
pusher: supports pushing data from the current database to Redis.
querier: current API support that interacts with RedisSearch

Practice

For example, there is a movie website. And need to integrate the search feature.

Setup RedisSearch
For quickly, we will use docker to build environment
```
docker run -p 6377:6379 redislabs/redisearch:latest
```

Create index document

// createIndexDocument.ts
const createVideoIndexDocument = async () => {
  const redis = await useRedisClient();
  await redis.ft.create(movieIndex, {
    name: {
      type: SchemaFieldTypes.TEXT,
      SORTABLE: true,
    },
    description: {
      type: SchemaFieldTypes.TEXT,
      SORTABLE: true,
    },
    duration: {
      type: SchemaFieldTypes.NUMERIC,
      SORTABLE: true,
    },
  });
};
createVideoIndexDocument();

ts-node createIndexDocument.ts # the script will generate video index document

Import sample data
To make your life easier, I have created a json file which contains over 300 movies
[ { "name": "[Star Talk] Which Penang hill is a must visit for Singapore leading male idol Li Nanxing?", "description": "[star talk] which penang hill is a must visit for singapore leading male idol li nanxing?", "duration": 716 }, { "name": "[Lu Lu Land] Jackson Wang Sends JJ Lin Birthday Greetings for His Birthday", "description": "[lu lu land] jackson wang sends jj lin birthday greetings for his birthday", "duration": 983 }, { "name": "[Test Me If You Can] Great Gadgets for Mopping and Sweeping! | EP 2/8", "description": "[test me if you can] great gadgets for mopping and sweeping! | ep 2/8", "duration": 202 }, { "name": "[Ah Boys No Limit] Challenge starts! Team PVP mode on! | EP 41/45", "description": "[ah boys no limit] challenge starts! team pvp mode on! | ep 41/45", "duration": 296 }, { "name": "[Ah Boys No Limit] ABNL Cheerleading captain is here! The prophecy of Hafiz came true? | EP 42/45", "description": "[ah boys no limit] abnl cheerleading captain is here! the prophecy of hafiz came true? | ep 42/45", "duration": 676 }, ... ]
Run the script to import above info info into Redis
```
// pusher.ts
import { createClient } from "redis";
import fs from "fs";
import path from "path";

const videos = JSON.parse(
  fs.readFileSync(path.resolve(__dirname, "./data/videos.json"), "utf-8")
);

const importMoviesData = async function () {
  const client = createClient({
    url: "redis://localhost:6380",
  });
  console.info("connecting redis");
  await client.connect();
  console.info("inserted item");

  let idx = 0;
  for (const item of videos) {
    await client.hSet("video:" + idx, item);
    idx++;
  }
};
importMoviesData();
```
ts-node pusher.ts

Implement RedisSearch

// search.ts
const fullTextSearch = async function () {
  const client = await redisClient();

  const readline = rl.createInterface({
    input: process.stdin,
    output: process.stdout,
  });
  while (true) {
    const obtainAnswerResult = () => {
      return new Promise((res) => {
        readline.question("input your text: ", (t) => {
          res(t);
        });
      });
    };

    const text = await obtainAnswerResult();
    console.info("searching for", text);
    const rs = await client.ft.search(movieIndex, `*${text}*`);
    console.info(rs.documents);
  }
};
fullTextSearch();

ts-node search.ts

Result

image-01-2

Advanced Features

There are still many features on RedisSearch. You can explore here

Thank for reading the post.