wy-mongo-es

Mongo-ES

A MongoDB to Elasticsearch connector

从mongo-es fork 过来的项目，由于默认mongdb中的表映射进入es需要每个字段都写对应关系， 我做了一个默认映射全部字段的修改, 在task 定义中加入了一个参数 transAll，映射全部字段，默认为false，

不仅如此，还改动了 scan这个任务的具体拉取数据方式，原本是 new ObjectID(id) 我给直接改成 string了

Installation

npm i -g mongo-es

Usage

Command Line

# normal mode
mongo-es ./config.json

# debug mode, with debug info printed
NODE_ENV=dev mongo-es ./config.json

Programmatically

const fs = require('fs')
const Redis = require('ioredis')
const { Config, Task, run } = require('mongo-es')

const redis = new Redis('localhost')

Task.onSaveCheckpoint((name, checkpoint) => {
  return redis.set(`mongo-es:${name}`, JSON.stringify(checkpoint))
})

// this will overwrite task.from in config file
Task.onLoadCheckpoint((name) => {
  return redis.get(`mongo-es:${name}`).then(JSON.parse)
})

run(new Config(fs.readFileSync('config.json', 'utf8')))

Concepts

Scan phase

scan entire database for existed documents

Tail phase

tail the oplog for documents' create, update or delete

Configuration

Structure:

{
  "controls": {},
  "mongodb": {},
  "elasticsearch": {},
  "tasks": [
    {
      "extract": {},
      "transform": {},
      "load": {}
    }
  ]
}

Detail example

controls

• mongodbReadCapacity - Max docs read per second (default: 10000). (optional)
• elasticsearchBulkInterval - Max bluk interval per request (default: 5000). (optional)
• elasticsearchBulkSize - Max bluk size per request (default: 5000). (optional)
• indexNameSuffix - Index name suffix, for index version control. (optional)

mongodb

• url - The connection URI string, eg: mongodb://user:password@localhost:27017/db?replicaSet=rs0. notice: must use a admin user to access oplog.
• options - Connection settings, see: MongoClient. (optional)

elasticsearch

• options - Elasticsearch Config Options, see: Configuration.
• indices - If set, auto create indices when program start, see: Indeces Create. (optional)

task.from

• phase - scan or tail
• time - tail oplog with query: { ts: { $gte: new Timestamp(0, new Date(time).getTime() / 1000) } }
• id - scan collection with query { _id: { $gte: id }} 改动了一下，id直接使用传入的string 而不是 ObjectID(id)

task.extract

• db - Database name.
• collection - Collection name in database.
• projection - Projection selector, see Projection.

task.transform

• mapping - The field mapping from mongodb's collection to elasticsearch's index.
• parent - The field in mongodb's collection to use as the _parent in elasticsearch's index. (optional)
• transAll - If set, all documents will be indexed

task.load

• index - The name of the index.
• type - The name of the document type.
• body - The request body, see Put Mapping.

License

Mozilla Public License Version 2.0