# Xpresser Meilisearch Plugin

# STAGE: (RFC)

This plugin will:

Add Meilisearch to your xpresser project.
Add cli commands to install/start Meilisearch.
Provide Search Models.

# Setup

npm install @xpresser/meilisearch
#OR
yarn add @xpresser/meilisearch

# Register Plugin

Add the following to your plugins.json file:

{
    "npm://@xpresser/meilisearch": true
}

# Add CLI Commands

Add the following to your use-xjs-cli.json file:

Note: If you don't have a use-xjs-cli.json file see xjs-cli#init-file first.

{
  "extensions": [
    "npm://@xpresser/meilisearch"
  ] 
}

# Configuration

Import plugin configuration:

npx xjs import meilisearch configs

This will import the meilisearch configuration file.

# Usage

# Install Meilisearch

npx xjs mei:install
# OR
node your-boot-file.js cli mei:install

The command above will install Meilisearch to your project in the 'storage/meilisearch/bin' directory by default. To change the bin directory, see pathToBinary in the configuration file.

# Start Meilisearch

npx xjs mei:start
# OR
node your-boot-file.js cli mei:start

Note: The command above will only start Meilisearch in the foreground.

# Start Meilisearch in Background

To run Meilisearch in the background, There are two ways:

The simple xpresser way (Using Pm2)
Running Meilisearch as a service.

# Using Pm2

Pm2 (opens new window) is a production process manager for Node.js. To start meilisearch using pm2, all we have todo is call your xpresser bootFile with the necessary arguments:

pm2 start your-boot-file.js -- cli mei:start

The command above is same as running npx xjs mei:start but this time will be handled by pm2.

# Run Meilisearch as a Service

# Search Model

A search model is a class that extends the MeiliSearchModel class.

# Define Search Model

import { defineSearchModel } from "@xpresser/meilisearch";
import { Index } from "meilisearch";

export const CustomSearchModel = defineSearchModel({
    // The name of the model.
    name: "Custom",

    // (Optional) The name of the model's table/index.
    // If not provided, the lower cased name of the model will be used.
    index: "custom",

    /**
     * (Optional) Modify/Customize index
     * @param index
     */
    async init(index: Index): Promise<void> {
        // do something with the index
        // maybe set the index's settings
    },

    /**
     * Provide the data to be indexed.
     */
    async data<T>(index: Index): Promise<T[]> {
        // return an array of objects to be indexed
        return [] as T[];
    }
});

# Sync Data

To sync the data you provided with Meilisearch, you can use the syncData method.

 const [lenghtOfQueuedData, QueryResult] = await UsersSearchModel.syncData();

The first data is the length of the queued data. The second data is the query result.

This can be called anywhere in your application and will sync the data.

Note: Your meilisearch instance must be running.

# Interact with Meilisearch Index

To interact with the raw Meilisearch index API, you can use the index property. For example:

await  UsersSearchModel.index.updateSettings({
    "searchableAttributes": ["name", "email"]
});

# Meilisearch Client

To interact with the Meilisearch client directly, you should use the useMeilisearch helper method.

import { useMeilisearch } from '@xpresser/meilisearch';

const client = useMeilisearch();

OR with your xpresser instance. If no instance is provided, the default instance will be used.

import { $ } from '../some-xpresser-instance';  

const client = useMeilisearch($);

# Sync with Cron

Meilisearch handles most tasks asynchronously so the need to sync data with cron is rare. only if you want to be timed and precise about when to sync data.

To sync with cron, simply create a job and call the syncData method.

import JobHelper from "xpresser/src/Console/JobHelper";
import { UsersSearchModel } from "search/models/folder";

export = {
    async handler(args: string[], job: JobHelper): Promise<any> {
        try {
            // sync data
            const [length] = await UsersSearchModel.syncData();
            // log the length of the data synced
            job.$.logSuccess(`Synced ${length} users`);
        } catch (e) {
            console.error(e);
        }

        // End the job
        job.end();
    }
};