Download the PHP package k-samuel/faceted-search without Composer

On this page you can find all versions of the php package k-samuel/faceted-search. It is possible to download/install these versions without Composer. Possible dependencies are resolved automatically.

FAQ

After the download, you have to make one include require_once('vendor/autoload.php');. After that you have to import the classes with use statements.

Example:
If you use only one package a project is not needed. But if you use more then one package, without a project it is not possible to import the classes with use statements.

In general, it is recommended to use always a project to download your libraries. In an application normally there is more than one library needed.
Some PHP packages are not free to download and because of that hosted in private repositories. In this case some credentials are needed to access such packages. Please use the auth.json textarea to insert credentials, if a package is coming from a private repository. You can look here for more information.

  • Some hosting areas are not accessible by a terminal or SSH. Then it is not possible to use Composer.
  • To use Composer is sometimes complicated. Especially for beginners.
  • Composer needs much resources. Sometimes they are not available on a simple webspace.
  • If you are using private repositories you don't need to share your credentials. You can set up everything on our site and then you provide a simple download link to your team member.
  • Simplify your Composer build process. Use our own command line tool to download the vendor folder as binary. This makes your build process faster and you don't need to expose your credentials for private repositories.
Please rate this library. Is it a good library?

Example code of k-samuel/faceted-search

Informations about the package faceted-search

PHP Version Total Downloads Build and Test Codacy Badge Codacy Badge

PHP Faceted search library 3.x

Simplified and fast faceted search without using any additional servers such as ElasticSearch, etc.

It can easily process up to 500,000 items with 10 properties. Create individual indices for product groups or categories and you won't need to scale or use more complex tools for a long time. The software is more effective when operates together with Roadrunner, Swoole, etc.

In addition to faceted filters, it supports exclusive filters. The software is optimized for uncompromising performance.

2.x version

Install

composer require k-samuel/faceted-search

Aggregates

The main advantage of the library is the fast and simplified building of aggregates.

Aggregates in Simple Terms

Imagine that a user has chosen several filters in the interface. We need to update the interface so that only filters compliant with the user’s choice (overlapping product properties) are represented in a list of available filters. We also have to display a number of available products hidden behind each filter.

Every time a user selects a new parameter, аt the code level, determine how many options are available based on the user's selection and display a new list of filters in the interface

This is simple enough, even if products have different structure of properties.

Notes

We recommend to create an individual index for each product category or type and include in such index only fields of concern.

Use your database to store frequently changing fields, such as price, quantity, etc. The faceted search should be used for preliminary data filtering.

Try to reduce the number of records processed. Try to index only products that, for example, are held in stock to exclude processing data on unavailable products.

Performance tests

Testing on a set of products with ten attributes, searching with filters on three fields.

v3.1.0 Bench PHP 8.2.10 + JIT + opcache (no xdebug extension)

ArrayIndex

Items count Memory Query Aggregate Aggregate & Count Sort by field Results Found
10,000 ~3Mb ~0.0003 s. ~0.0006 s. ~0.001 s. ~0.0002 s. 907
50,000 ~20Mb ~0.001 s. ~0.002 s. ~0.006 s. ~0.0005 s. 4550
100,000 ~40Mb ~0.002 s. ~0.006 s. ~0.013 s. ~0.001 s. 8817
300,000 ~95Mb ~0.006 s. ~0.016 s. ~0.036 s ~0.002 s. 26891
1,000,000 ~329Mb ~0.024 s. ~0.053 s. ~0.134 s. ~0.009 s. 90520
1,000,000 UB ~324Mb ~0.046 s. ~0.078 s. ~0.159 s. ~0.015 s. 179856

FixedArrayIndex

Items count Memory Query Aggregate Aggregate & Count Sort by field Results Found
10,000 ~2Mb ~0.0005 s. ~0.0009 s. ~0.002 s. ~0.0003 s. 907
50,000 ~12Mb ~0.002 s. ~0.003 s. ~0.012 s. ~0.0007 s. 4550
100,000 ~23Mb ~0.006 s. ~0.010 s. ~0.029 s. ~0.001 s. 8817
300,000 ~70Mb ~0.012 s. ~0.022 s. ~0.072 s. ~0.003 s. 26891
1,000,000 ~233Mb ~0.045 s. ~0.070 s. ~0.257 s. ~0.012 s. 90520
1,000,000 UB ~233Mb ~0.068 s. ~0.101 s. ~0.290 s. ~0.017 s. 179856

(Apple M2 macOS 14.0)

Benchmark of a library experimental port at Golang https://github.com/k-samuel/go-faceted-search

Bench v0.3.3 go 1.21.1 darwin/arm64 with parallel aggregates.

Items count Memory Query Aggregate & Count Sort by field Results Found
10,000 ~7Mb ~0.0002 s. ~0.002 s. ~0.0001 s. 907
50,000 ~14Mb ~0.001 s. ~0.017 s. ~0.001 s. 4550
100,000 ~21Mb ~0.002 s. ~0.037 s. ~0.001 s. 8817
300,000 ~47Mb ~0.008 s. ~0.107 s. ~0.004 s. 26891
1,000,000 ~140Mb ~0.031 s. ~0.363 s. ~0.015 s. 90520
1,000,000 UB ~138Mb ~0.059 s. ~0.899 s. ~0.028 s. 179856

(Apple M2 macOS 14.0)

The internal structure of index arrangement in versions on PHP and Golang will be different starting from experimental port ver. 0.0.3 due to peculiarities of the Hash Map internal structure in these languages. In Go, we had to stop using Hash Map to make data storage in slices more effective, which initially allowed us to match PHP version performance.

In PHP, array (hashMap) is more effective for the current task due to using DoubleLinkedList and HashMap key packing.

Go has more effective methods of reduction of the size of slices without copying data (used for list deduplication). This allows to find overlapping using sorted slices.

Further comparison makes little sense because of different algorithms.

Examples

Create an index using console/crontab etc.

Using in application

Indexers

If there are too many values for a certain field in your data, you may use Range Indexer to accelerate RangeFilter operation. For example, searching by price ranges of products. Prices can be divided into intervals with a required increment.

Please, remember that RangeFilter is a rather slow solution, and it’s better to avoid facets with high value variability.

Sorting values inside the range is only possible during the process of index creation, since this aspect is lost in case with real value. Thus, when using RangeIndexer you shouldn’t add individual values to a ready index. As a way to solve this problem, library adds new values to the end of the range and sorts them only between themselves (sorts new values and adds them to the end).

This specific feature makes sense only when you use results sorting by the field that is a range indexed using RangeIndexer.

RangeListIndexer allows creating your own ranges without using an increment as in case with RangeIndexer

Also, you can create your own indexers with range detection method

FixedArrayIndex

FixedArrayIndex is slower than ArrayIndex, but it uses much less RAM. FixedArrayIndex data are compatible with ArrayIndex.

Filter. Self-Filtration Features

When building aggregates, self-filtering of properties is disabled. This allows the user selecting a different value of the same field for filtering (switch the selection) with filter by a certain value of such field being on.

Example. The user wants to find a phone with 32Gb RAM, ticks this checkbox from the provided list (16, 32, 64). If self-filtering is on, then other options will disappear from the user interface. Only 32 Gb value will remain as it will be filtered on the basis of the user’s choice. In this case the user won’t be able to change his/her choice to 64 Gb or 16 Gb.

When building aggregates, field values are used to limit the list of available options of other fields. For example: Filter by “size” field value is used to limit the list of “brand” field results.

Everything depends on your library use scenario. Library was initially designed to simplify the user UI building. If you use library for technical analysis or statistics, enabling self-filtering will help you get expected results.

For all filters:

For individual filter:

More Examples

Tested but discarded concepts

Bitmap

Bloom Filter

Creating multiple indexes for different operations

Q&A

All versions of faceted-search with dependencies

PHP Build Version
Package Version
Requires php Version >=7.4.0
Composer command for our command line client (download client) This client runs in each environment. You don't need a specific PHP version etc. The first 20 API calls are free. Standard composer command

The package k-samuel/faceted-search contains the following files

Loading the files please wait ....