@noriller/easy-filter

v2.0.7

Published

2 years ago

EasyFilter is a lightweight, zero dependencies, minimal setup, intuitive and powerful filter for all your filter needs.

Downloads

0High
0Medium
0Low

noriller

filter filter-lists query query-parser javascript match matching-engine search search-engine sort

EasyFilter

🎈 Welcome to EasyFilter! 👋

EasyFilter is a lightweight ☁️, just one dependency 🚢, minimal setup 😮, intuitive 😃 and powerful 💪 filter for all your filter needs.

It's as easy as this:

  const filter = EasyFilter(sourceArray)
  const filteredResult = filter.search('your query')

The one dependency is EasyFilterParser that was part of this package before. 😉

Get Started

Use your choice of package manager

npm install @noriller/easy-filter

yarn add @noriller/easy-filter

Then import it with the syntax of your choice

import EasyFilter from '@noriller/easy-filter'

const EasyFilter = require('@noriller/easy-filter')

Finally, to actually use it

  const filter = EasyFilter(sourceArray)
  const filteredResult = filter.search('your query')

That's it! 🧙‍♂️

Check out the section EasyFilter Operators to see all that you can pass to the filter, the real ✨magic✨ is there!

✨ Magic like turning this:
  `search for something "this between quotes" and then here:"you search for this"`
✨
  Into something that works for single values, quoted values and even values nested inside keys. AND MORE!
✨

Really? That's it?

Ok. If you need more options, here's the full setup you can do using all options available:

const filter = EasyFilter(sourceArray, {
    filterOptions: {
      dateFormat: 'DD-MM-YYYY',
      normalize: true,
      indexing: true,
      limit: 10,
    },
    tagAliases: {
      tag: ['tag1', 'tag2', 'tag3'],
    }
  })

const filteredResult = filter.search('your query')

It's still that simple. 👨‍💻

All the options will be explained in EasyFilter Options. And most of them you can pass in the search 🔎 string when you need.

Inspirations and motivation

In corporate scenarios, sometimes we have too much information 😵. We make pages with endless columns and if we need to filter that data, we either use something generic like Object.keys(object).join(' ').includes('string') or we have to make a custom search... for. each. table. 😫

Meanwhile I saw awesome (and probably custom solutions) in things we use everyday.

Check out the ones I was aiming for 🌟:

Github
Stackoverflow
Gmail/Google Search

In the latter, users can use their UI to create their queries while powerusers can just type that and much more.

I too needed to provide a way to users to filter the data and ended up settling at a simpler version of this project. Mostly because of all the solutions I was able to find were neither user or developer friendly. 😢

(Also a little rant: search, filter, matcher and the like are a nightmare to search for... too many hits and too little relevant results 😕)

This is what I'm trying to offer here: a powerful engine to make your queries. 😎👍

Then it's up to you to offer a UI for what makes sense for your data. And it's still intuitive for common users and powerful for powerusers.

Why should you use

Unless searching and filtering data IS your business... each table, each data source you need to filter based on any criteria could be another source of frustration. 👨‍🏭

While sometimes this can be easy as dataSource.filter(x => x.attribute === 'something'), usually it's more complicated than that.

And when you have to factor what the users might want at any given time, it can become a nightmare. 😫

And if you're seeing yourself here... then you probably should use it.

When should you use, or better yet, when you shouldn't use?

If something like Object.keys(object).join(' ').includes('string') or dataSource.filter(x => x.attribute === 'something') is enough for your need, you probably wouldn't want to bother using it.
EasyFilter is not a fuzzy filter (at least not yet, who knows? 🤔), so if you're expecting clumsy typers searching your data... they might have a hard time.
EasyFilter also don't care about upper or lowercase and/or ordering of the words in your queries (again: at least not yet), so if that's important... I'm sorry? 🙇‍♂️
This one is a little relative: if you need maximum performance... well, test it out if it work for you! The more options you use and the more your data objects branches out on objects upon objects... the more time it needs to traverse everything. (And even then, I believe you can use it for prototyping and/or as a crutch while you make your own custom filter. Don't worry, I understand. 😉)

The trade off is clear: we give you a powerful engine that will return the data following what you're searched for, but it comes at a cost.

For your everyday use, you're probably fine and your users will love it. 😎

(As a side note, I would love to know how EasyFilter fare against any other solution you might try. ❤😍❤)

EasyFilter Operators

Most of this should be intuitive for most users... that's what I was aiming for after all. 🧐

OR query

Any word or operators are, primarily and lastly, treated as OR queries.

To be a match, the data needs only to match any of them to be returned.

OR Example:

filter.search('word1 word2 tag:value "quoted value"')

word1, word2, tag:value and "quoted value" each become separated entities and a match of any one of those will return.

AND query

Anything inside quotes (either double " or single ') will be treated as one entity.

EasyFilter relies heavily on recursion and this one entity will be split into multiple entities, following those entities rules.

To be a match, the data must get a match from each subquery inside quotes.

AND Remarks

An AND query can contain: OR, TAG and even nested AND queries.

In case of nested TAG and AND queries, the nested quote must not match the parent quote.

AND Example:

filter.search('"quoted value tag:value"')

quoted, value and tag:value first are an AND query and will match only if all of the subqueries match.

In this case, both quoted and value become OR queries and tag:value becomes TAG query.

TAG query

TAG here is equivalent to any key of a Javascript object.

TAG, like AND queries, return subqueries where they target a slice of the object, namely the key.

To be a match, the data must get a match from the subquery after the TAG.

TAG Remarks

A TAG query can contain: OR, AND and then NULL and RANGE/DATE_RANGE queries.

TAG doesn't support nested TAG queries. (As of now.)

Types of TAG queries with Examples:

TAG - Simple

filter.search('tag:value')

Just the TAG followed by a colon and the value.

value in this example will become an OR query.

TAG - OR

filter.search('tag:(value1 value2 value3)')

By using brackets, you can have an OR query with multiple values at once.

TAG - AND

filter.search('tag:"value1 value2 value3"')

By using quotes (single/double), you can have an AND query.

TAG - Null Values

filter.search('tag:null tag:nil tag:none tag:nothing')

By passing, alone, any of the words: NULL, NIL, NONE or NOTHING as the value of the TAG, it will match only if the key doesn't exist in the object (or if the key is null or undefined).

tag:(nothing), in contrast, will match only if the key contains the string "nothing".

TAG - Chaining Tags

filter.search('tag.subTag.thirdTag:value')

You can chain tags together using a . (full stop/period).

This would be equivalent to nested TAGs. (Nested tags aren't supported.)

TAG - Arrays

filter.search('tag.0:value tag2.*.subTag:value')

While the most common use case for EasyFilter would be, indeed, common objects (think JSON Objects with key/value pairs), arrays are supported.

The main difference is that the key they use are numerical and ordered.

tag.0:value will search inside the "tag" key, then in the element "0" for the "value".

tag2.*.subTag:value will search inside the "tag2" key, then in ALL elements for the "subTag" and then for the "value".

So, if the position in the array matters, use the index. Otherwise, use * (asterisk) to search all elements in the array.

TAG - RANGE

filter.search('tag:range(0,5)')

By passing, alone, the operator RANGE() you can pass one or two arguments that will filter based on the numbers.

RANGE can only be used inside TAG and with number values.

The first argument is the lower bound (-Infinity) and the second argument is the upper bound (Infinity).

Passing only one argument sets only the lower bound. To set only the upper bound, pass it empty: RANGE(,5).

TAG - DATE_RANGE

filter.search('tag:dateRange(2020-05-01, 2021-09-05)')

By passing, alone, the operator DATERANGE() you can pass one or two arguments that will filter based on the dates.

DATERANGE can only be used inside TAG and with date values.

The first argument is the lower bound (0000-01-01) and the second argument is the upper bound (9999-01-01).

Passing only one argument sets only the lower bound. To set only the upper bound, pass it empty: DATERANGE(,2021-09-05).

More on accepted Date Formats in Date Format (Query), but you can use all the common formats like DD/MM/YYYY, MM/DD/YYYY and YYYY/MM/DD as long as you pass it as an OPTION. If no Date Format is provided, the Javascript default implementation of new Date('your date string') will be used.

NOT query

By nesting any and multiple queries inside the syntax NOT() you can invert those and it will NOT return anything that matches.

If there's a match in the NOT query, it won't return even if there's a match in other queries.

If your query contains only NOT queries, it will return everything that don't have a match.

When combining with other queries, NOT queries will filter out matches from those.

NOT Remarks

A NOT query can contain: OR, AND and TAG queries.

All NOT are parsed at the same level, nesting it inside other queries will just remove them from the query.

NOT Example:

filter.search('not("quoted value tag:value")')

Any element with quoted, value and tag:value will not be returned.

EasyFilter Options

There's three types of options:

Those that can be passed any time:
Those that can only be passed in the setup:
- Filter Options
  - Date Format (Setup)
- Tag Aliases
Those that can only be passed with the query:
- Date Format (Query)

OPTION keyword

Using the syntax OPTION() or OPTIONS() you can pass the following options inside your search string.

The OPTION keyword is parsed first, it will be just removed if nested in other queries and anything else inside will be either parsed as an option or ignored.

DATEFORMAT (Query)

When passed as an OPTION, DateFormat will be used to parse the dates used in DATE_RANGE.

This way your users can use their locale date format in their query.

When using DATE_RANGE, if no DateFormat is passed as an option the Javascript default implementation of new Date('your date string') will be used.

The formats can be: YYYY-MM-DD, DD-MM-YYYY and MM-DD-YYYY while the separators can be: -, ., , and /.

DateFormat Example

filter.search('tag:dateRange(30-12-2020,30-12-2022) option(dateFormat:DD.MM.YYYY)')

NORMALIZE

When the NORMALIZE option is used, EasyFilter will discard/ignore every and all diacritics when making comparisons. It's FALSE by default.

This means that with NORMALIZE: Crème brûlée is equal to Creme brulee.

EasyFilter uses the string.normalize('NFD') javascript API to decompose the strings and then remove all Combining Diacritical Marks.

NORMALIZE uses a boolean flag, and when used in OPTIONS alone like option(normalize) it will assume the TRUE value, but you can explicitly use: normalize:true.

You can also use normalize:false to disable a setup default normalization for a specific query.

INDEXING

When the INDEXING option is used, EasyFilter will use a numerical ranking system and if there's a match, it will return a copy of the element with an additional key _EasyFilterIndex that will have a "relevance score" that you can use to sort the results. It's FALSE by default.

INDEXING uses a boolean flag, and when used in OPTIONS alone like option(index) or option(indexing) it will assume the TRUE value, but you can explicitly use: index:true.

You can also use normalize:false to disable a setup default indexing for a specific query.

How it indexing works in EasyFilter

EasyFilter gives a number based on the types of queries and multiplies the results returned by nested queries.

OR = 1
AND = 3
TAG = 5
RANGE/DATE_RANGE/NULL = 10

This way, using more generic queries, those with the more specific parameters will have a higher indexing value.

LIMIT

When the LIMIT option is used, EasyFilter will return only the LIMIT number of results. It's Zero/FALSE by default.

LIMIT needs a number value, when used in OPTIONS you need to also pass a number value: option(limit:1).

You can also use limit:0 to disable a setup default limit for a specific query.

Setup Options

In the setup you may pass:

Filter Options

The following options works the same way as if passing in the query:

By passing it in the setup, they will be used in every search.

DATEFORMAT (Setup)

When passed in the Setup, DateFormat will be used to parse the dates in your Source Array.

If no DateFormat is passed in the setup, the Javascript default implementation of new Date('your date string') will be used.

If that default implementation wouldn't work with your source array, then provide a DateFormat.

The formats can be: YYYY-MM-DD, DD-MM-YYYY and MM-DD-YYYY while the separators can be: -, ., , and / (you can use the provided typing).

Tag Aliases

Pass TAG Aliases in the setup to expose to users more friendly (or broader) terms that they can call your data using TAG.

Tag Aliases should be a dictionary with key/value pairs where the key is what your users can use and the value is a array of strings that will refer to your actual data.

Our data sources might not always be the most user friendly, or something important might be nested where users Tag Aliases couldn't possibly know. This is where you use Tag Aliases.

Aliases Examples

const filter = EasyFilter(sourceArray, {
    tagAliases: {
      // if you want more friendly aliases
      data: ['DT_0001X420'],
      name: ['nm_first', 'nm_last'],
      // if the important data is nested
      age: ['person.info.age'],
      // if your users expect to find everything related to a word
      address: ['address', 'city', 'country', 'province', 'zip_code'],
      // and you have no idea which words they will search for
      // just create multiple aliases with the same tags
      city: ['address', 'city', 'country', 'province', 'zip_code'],
      country: ['address', 'city', 'country', 'province', 'zip_code'],
      province: ['address', 'city', 'country', 'province', 'zip_code'],
      zip: ['address', 'city', 'country', 'province', 'zip_code'],
      location: ['address', 'city', 'country', 'province', 'zip_code'],
      where: ['address', 'city', 'country', 'province', 'zip_code'],
      position: ['address', 'city', 'country', 'province', 'zip_code'],
    }
  })

What's next?

Here's something you can expect in the future:

New objective: EasyFilter will now become a trilogy! They will all share the same parser, so you will be able to filter values already buffered and values in your databases all the same, simple way.
- This, which filter values in objects.
- EasyFilter-SQL - That will create SQL queries.
- EasyFilter-Mongo - That will create Mongo queries.
(TBD) Streams support ⚡: be it from an API or from a data source, data usually comes as a stream. If the Javascript Engine can handle streaming fine, why wait for it to buffer everything?

There's a problem or it could be better

Either if you're encountered a problem: 😢 or if you're have an idea to make it better: 🤩

Feel free to contribute, to open issues, bug reports or just to say hello! 🤜🤛

In case of bugs or errors, if possible, send an example of your data, of the query you're using and what you've expected.

Since it supports any kind of data or queries... who knows what can happen?

I do have one remark to say: It will probably err on the side of caution... maybe returning more than you might have expected and when using the negation query, filtering more than intended, so keep that in mind.