hook which adds fuzzy search for mongodb through $search in find query

Downloads in past


2.0.13 years ago7 years agoMinified + gzip package size for feathers-mongodb-fuzzy-search in KB


npm version Build Status
Add $search to mongodb service.find, update, patch and remove queries. Full-text search on documents with stemming as well as pattern matching on individual fields.
For full-text search, be sure to index your text fields, as this plugin uses mongodb $text.
For field pattern matching, mongodb $regex is used.


npm install feathers-mongodb-fuzzy-search


const service = require('feathers-mongodb')
const search = require('feathers-mongodb-fuzzy-search')

app.use('/messages', service({
  Model: db.collection('messages'),
  whitelist: ['$text', '$search'], // fields used by feathers-mongodb-fuzzy-search

// add search hook
// may also use service.hooks to apply it on individual services only
  before: {
    all: [
      search(), // full text search on text indexes
      search({  // regex search on given fields
        fields: ['firstName', 'lastName']

// create a text index on title property, for full-text search
// you may add multiple fields to the text index
// see the mongodb documentation for more on $text
const messages = app.service('messages')

// If you're using MongoDB database adapter:
messages.Model.createIndex({ title: 'text' })
// or if you're using Mongoose database adapter:
// messages.Model.index({ title: 'text' })

// find documents with title containing 'cat'
// will find titles including 'cat', 'cats', etc. thanks to mongodb stemming
// note: you can only use await inside async functions
let catDocuments = await messages.find({ query: { $search: 'cat' } })

// find users with first name containing a 's' and last name containing 'art'
let userDocuments = await app.service('users').find({
  query: {
    firstName: { $search: 's' },
    lastName: { $search: 'art' }

Complete example here.

REST usage

Full text search for qwerty with mongodb $text:
curl http://localhost:3030/messages?$search=qwerty

Search for qwerty on field firstName with mongodb $regex:
curl http://localhost:3030/users?firstName[$search]=qwerty


Full-text search

As default " in $search is removed and $search is padded with ". E.g. some " text becomes "some text". If you want to disable this behaviour and leverage the full MongoDB $text API, you can disable escaping like this:
  before: {
    find: search({ escape: false })

RegExp field search

Remember to allow $regex in the service:
app.use('/messages', service({
  Model: db.collection('messages'),
  whitelist: ['$regex'], // field used by feathers-mongodb-fuzzy-search

If not, you will get the error BadRequest: Invalid query parameter $regex on requests.
The options object given to search(options) supports the following:
  • fields: Array of field names to allow searching in.
  • excludedFields: Array of field names that can't be searched. If given, any field not in array can be searched.
  • fieldsNotEscaped: Array of fields to be excluded from RegExp escape. As default any field not given are escaped to avoid RegExp denial of service attacks.

  before: {
    find: search({
      // make all fields but 'fullName' are searchable
      excludedFields: ['fullName'],
      // do not escape RegExp special characters for the field 'firstName'
      fieldsNotEscaped: ['firstName']

MongoDB options

You can pass MongoDB options for $text, like $language, $caseSensitive and $diacriticSensitive with your query. E.g. If you'd like to disable stemming add $language: 'none' to your query parameters:
  query: {
    $search: 'cats',
    $language: 'none'

NOTE: Remeber to allow the fields in your service:
app.use('/messages', service({
  Model: db.collection('messages'),
  // fields used by feathers-mongodb-fuzzy-search
  whitelist: ['$text', '$search', '$caseSensitive', '$language', '$diacriticSensitive'],

Additional information

This package is tested with MongoDB version 3.2. You will probably run into problems using older versions of MongoDB, for example version 2.4 does not support $text search.
See mongodb documentation for more details about $text.
See mongodb documentation for more details about $regex.


npm test  # runs mocha


MIT © 2017 Arve Seljebu / Luc Claustres