yoganandgopalan/slim-skeleton

Boilerplate for building a REST API with Slim PHP micro-framework With Limit, Filtering, Full text search, Sorting

v1.0.0 2018-05-28 20:07 UTC

This package is not auto-updated.

Last update: 2024-03-24 10:48:45 UTC


README

This is an app skeleton for the Slim PHP Micro-Framework to get started quickly building a REST API, with CRUD, Sorting, Pagination &filtering

Features

Installation

$ composer create-project -s dev yoganandgopalan/slim-skeleton [app-name]

Download vendor packages

$ composer update

Features

Create database tables

$ php bin/console db

Create users

$ php bin/console user:create

Use --admin option to set the user as admin

Dump routes

Execute the following command at the project root to print all routes in your terminal

$ php bin/console routes

Use --markdown or -m option to display routes in markdown format

$ php bin/console routes -m > API.md

If you're using Oh My Zsh, you can install the symfony2 plugin, which provides an alias and autocompletion:

# Without Symfony2 plugin
$ php bin/console db

# With Symfony2 plugin
$ sf db

Note

You might want to replace the authentication part with a real OAuth implementation

Sort, Filter, Limits

Custom identification columns

If you pass an array as the second parameter to parseSingle, there now have to be column/value pairs. This allows us to pass multiple conditions like:

SortFilter::parseSingle($request, $books, array('id_origin' => 'Random Bookstore Ltd', 'id' => 1337));

URL parsing

Url parsing currently supports:

  • Limit the fields
  • Filtering
  • Full text search
  • Sorting
  • Define limit and offset

There are two kind of api resources supported, a single object and a collection of objects.

Single object

If you handle a GET request on a resource representing a single object like for example /api/books/1, use the parseSingle method.

parseSingle($request, $queryBuilder, $identification, [$queryParams]):

  • $request: Request from controller
  • $queryBuilder: Query builder object, Eloquent model or Eloquent relation
  • $identification: An integer used in the id column or an array column/value pair(s) (array('isbn' => '1234')) used as a unique identifier of the object.
  • $queryParams: An array containing the query parameters. If not defined, the original GET parameters are used.
SortFilter::parseSingle($request, $book, 1);

Collection of objects

If you handle a GET request on a resource representing multiple objects like for example /api/books, use the parseMultiple method.

parseMultiple($request, $queryBuilder, $fullTextSearchColumns, [$queryParams]):

  • $request: Request from controller
  • $queryBuilder: Query builder object, Eloquent model or Eloquent relation
  • $fullTextSearchColumns: An array which defines the columns used for full text search.
  • $queryParams: An array containing the query parameters. If not defined, the original GET parameters are used.
SortFilter::parseMultiple($request, $book, array('title', 'isbn', 'description'));

Result

Both parseSingle and parseMultiple return a Result object with the following methods available:

getBuilder(): Returns the original $queryBuilder with all the functions applied to it.

getResult(): Returns the result object returned by Laravel's get() or first() functions.

getResultOrFail(): Returns the result object returned by Laravel's get() function if you expect multiple objects or firstOrFail() if you expect a single object.

getResponse($resultOrFail = false): Returns a Laravel Response object including body, headers and HTTP status code. If $resultOrFail is true, the getResultOrFail() method will be used internally instead of getResult().

cleanup($cleanup): If true, the resulting array will get cleaned up from unintentionally added relations. Such relations can get automatically added if they are accessed as properties in model accessors. The global default for the cleanup can be defined using the config option cleanup_relations which defaults to false.

SortFilter::parseSingle($books, 42)->cleanup(true)->getResponse();

Filtering

Every query parameter, except the predefined functions _fields, _sort, _limit, _offset and _q, is interpreted as a filter. Be sure to remove additional parameters not meant for filtering before passing them to parseMultiple.

/api/books?title=The Lord of the Rings

All the filters are combined with an AND operator.

/api/books?title-lk=The Lord*&created_at-min=2014-03-14 12:55:02

The above example would result in the following SQL where:

WHERE `title` LIKE "The Lord%" AND `created_at` >= "2014-03-14 12:55:02"

Its also possible to use multiple values for one filter. Multiple values are separated by a pipe |. Multiple values are combined with OR except when there is a -not suffix, then they are combined with AND. For example all the books with the id 5 or 6:

/api/books?id=5|6

Or all the books except the ones with id 5 or 6:

/api/books?id-not=5|6

The same could be achieved using the -in suffix:

/api/books?id-in=5,6

Respectively the not-in suffix:

/api/books?id-not-in=5,6
Suffixes
Suffix Operator Meaning
-lk LIKE Same as the SQL LIKE operator
-not-lk NOT LIKE Same as the SQL NOT LIKE operator
-in IN Same as the SQL IN operator
-not-in NOT IN Same as the SQL NOT IN operator
-min >= Greater than or equal to
-max <= Smaller than or equal to
-st < Smaller than
-gt > Greater than
-not != Not equal to

Sorting

Two ways of sorting, ascending and descending. Every column which should be sorted descending always starts with a -.

/api/books?_sort=-title,created_at

Fulltext search

Two implementations of full text search are supported. You can choose which one to use by changing the fulltext option in the config file to either default or native.

Note: When using an empty _q param the search will always return an empty result.

Limited custom implementation (default)

A given text is split into keywords which then are searched in the database. Whenever one of the keyword exists, the corresponding row is included in the result set.

/api/books?_q=The Lord of the Rings

The above example returns every row that contains one of the keywords The, Lord, of, the, Rings in one of its columns. The columns to consider in full text search are passed to parseMultiple.

Native MySQL implementation

If your MySQL version supports fulltext search for the engine you use you can use this advanced search in the api handler.
Just change the fulltext config option to native and make sure that there is a proper fulltext index on the columns you pass to parseMultiple.

Each result will also contain a _score column which allows you to sort the results according to how well they match with the search terms. E.g.

/api/books?_q=The Lord of the Rings&_sort=-_score

You can adjust the name of this column by modifying the fulltext_score_column setting in the config file.

Limit the result set

To define the maximum amount of datasets in the result, use _limit.

/api/books?_limit=50

To define the offset of the datasets in the result, use _offset.

/api/books?_offset=20&_limit=50

Be aware that in order to use offset you always have to specify a limit too. MySQL throws an error for offset definition without a limit.