PHP Database Web API

Version: 0.7.0

Github: https://github.com/marcocesarato/Database-Web-API

Author: Marco Cesarato

Description

Dynamically generate RESTful APIs from the contents of a database table. Provides JSON, XML, and HTML. Supports most popular databases.

What problem this solves

Creating an API to access information within existing database tables is laborious task, when done as a bespoke task. This is often dealt with by exporting the contents of the database as CSV files, and providing downloads of them as a “good enough” solution.

How this solves it

Database Web API acts as a filter, sitting between a database and the browser, allowing users to interact with that database as if it was a native API. The column names function as the key names. This obviates the need for custom code for each database layer.

When Alternative PHP Cache (APC) is installed, parsed data is stored within APC, which accelerates its functionality substantially. While APC is not required, it is recommended highly.

Wiki

Go to this link for go to the wiki of the platform and have a better informations about the usage of this platform and have more examples:

https://github.com/marcocesarato/Database-Web-API/wiki

Databases supported

• 4D
• CUBRID
• Firebird/Interbase
• IBM
• Informix
• MS SQL Server
• MySQL
• ODBC and DB2
• Oracle
• PostgreSQL
• SQLite

Requirements

• PHP
• Database
• APC (optional)

Installation

• Set the configuration on config.php. Follow the below example to register a new dataset. Tip: It's best to provide read-only database credentials here if you want read only.
• (Optional) If you want enable an authentication system you must compile on the config.php the constant __API_AUTH__ as on the example below.
• (Optional) Document the API. For this you can use auto-documentation using file docs.php
• (Optional) Use Hooks for manage permissions (as can_read, can_write, can_edit, can_delete)

Configuration

Edit config.php to include a single instance of the following for each dataset (including as many instances as you have datasets):

Define API Name

define("__API_NAME__", "Database Web API"); // API Name

Define datasets

Dataset configurations:

Example complete with explanation

define("__API_DATASETS__", serialize(array(
	'dataset' => array(
		'default' => true, // Default dataset
		'api' => true, // Accessible from API request url (ex. is false if you have a different database for auth users)
		'name' => 'database_name', // Database name
		'username' => 'user', // root is default
		'password' => 'passwd', // root is default
		'ttl' => 1, // Cache time to live. Disable cache (1 second only)
		'server' => 'localhost',  // localhost default
		'port' => 5432, // 3306 is default
		'type' => 'pgsql', // mysql is default
		'table_list' => array( // Tables's whitelist (Allow only the tables in this list, if empty allow all)
			'users'
		),
		'table_blacklist' => array( // Tables's blacklist
			'passwords'
		),
		'table_docs' => array(
			/** @example
			'table' => array(
				"column" => array(
				"description" => "Column description",
					"example" => "1",
				),
			),
			*/
		), // For Autodocoumentation, url ex. /dataset/docs/table.html
		'column_list' => array( // Columns's whitelist (Allow only the columns in this list, if empty allow all)
			'users' => array(
				'username',
				'name',
				'surname'
			)
		),
		'column_blacklist' => array( // Columns's blacklist
			'users' => array(
				'password',
			)
		),
	),
)));

Note: All fields of __API_DATASETS__ (except the name of database) are optional and will default to the above.

Example

Here is a dataset example for a MySQL database named “inspections,” accessed with a MySQL user named “website” and a password of “s3cr3tpa55w0rd,” with MySQL running on the same server as the website, with the standard port of 3306. All tables may be accessed by Database to API except for “cache” and “passwords,” and among the accessible tables, the “password_hint” column may not be accessed via Database to API. All of this is registered to create an API named “facility-inspections”.

array( 
	'default' => true,
	'name' => 'inspections',
	'username' => 'website',
	'password' => 's3cr3tpa55w0rd',
	'server' => 'localhost',
	'ttl' => 1,
	'port' => 3306,
	'type' => 'mysql',
	'table_docs' => array(),
	'table_list' => array(),
	'table_blacklist' => array('cache', 'passwords'),
	'column_blacklist' => array('password_hint'),
	'column_list' => array(),
);

Retrieving the contents of the table history within this dataset as JSON would be accomplished with a request for /facility-inspections/history.json. Note that it is the name of the dataset (facility-inspections) and not the name of the database (inspections) that is specified in the URL.

For a SQLite database, simply provide the path to the database in name.

For an Oracle database, you can either specify a service defined in tsnames.ora (e.g. dept_spending) or you can define an Oracle Instant Client connection string (e.g., //localhost:1521/dept_spending).

(Optional) Authentication system

Auth configuration

Users configuration

define("__API_AUTH__",  serialize(array( // Set null for disable authentication
	'sqlite' => false, // Enabled save token on SQLite file
	'sqlite_database' => 'api_token', // SQLite filename (only with sqlite = true)
	'api_database' => 'dataset', // Authentication database
	'api_table' => 'api_authentications', // API token table name
	'users' => array(
		'database' => 'dataset', // Database where users are stored
		'table' => 'users', // Table where users are stored
		'columns' => array(
			'id' => 'user_id', // Id column name
			'username' => 'user_name', // Username column name
			'password' => 'password', // Password column name
			'admin' => array('is_admin' => 1) // Admin bypass condition. With this condition true API bypass all black/whitelists and permissions. Set NULL for disable
		),
		'search' => array('user_id', 'email', 'username'), // Search user by these fields
		'check' => array('active' => 1) // Some validation checks. In this case if the column 'active' with value '1'. Set NULL for disable
	),
)));

API Structure

Format availables:

• JSON

• XML

• HTML

Generic URL format for all kind of request:

Standard

• Fetch all: /[database]/[table].[format]
• Fetch all with limit: /[database]/[limit]/[table].[format]
• Fetch: /[database]/[table]/[ID].[format]
• Fetch search by column: /[database]/[table]/[column]/[value].[format]
• Documentation: /[database]/docs/[table].[format]

Authentication

Authentication needed for browse the database.

The authentication permit to managed the privilege of the users (read, write, modify, delete)

• Authentication: /auth.[format]

Set the header Auth-Account with the username/user id and Auth-Password with the encrypted password like this:

Request example:

GET /auth.json HTTP/1.1
Host: localhost
Auth-Account: marco.cesarato
Auth-Password: md5password

Response example:

[{"token": "b279fb1d0708ed81e7a194e0c5d928b6"}]

Example of token usage on GET, POST, PUT, PATCH and DELETE requests:

Set the header Access-Token with the token values received from auth request like this:

GET /database/users.json` HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

Check Authentication

Check authentication check is needed for verify if a token is valid.

• Check authentication: /auth/check.[format]

Request example:

GET /auth/check.json HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

Response example:

{

	"user": {
		"id": "1",
		"role_id": "",
		"is_admin": true
	},
	"response": {
		"status": 200,
		"message": "OK"
	}

}

GET Request

Retrieve data from dataset

• Fetch all: /[database]/[table].[format]

• Fetch all with limit: /[database]/[limit]/[table].[format]

• Fetch: /[database]/[table]/[ID].[format]

• Fetch search by column: /[database]/[table]/[column]/[value].[format]

• Fetch all joining table:

  join[table] = array(
    'on' => <column_id>,		   // Column of the table joined
    'value' => <value>,			// Column of main table or value
    'method' => (left|inner|right) // Optional
  )

  Example with value:

  join[users]['on'] = id
  join[users]['value'] = 1
  join[users]['method'] = 'INNER'

  Example with column:

  join[users]['on'] = id			// Column of the table joined
  join[users]['value'] = user_id	// Column of the main table (no users)
  join[users]['method'] = 'INNER'

• Additional parameters

ex: /[database]/[table]/[column]/[value].[format]?order_by=[column]&direction=[direction]

Examples of GET requests:

GET /dataset/users.json HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

GET /dataset/10/users.json HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

GET /dataset/users/1.json HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

GET /dataset/users/is_active/1.json?order_by=username&direction=desc HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

Advanced search

Note: These examples are valid only for GET and PUT requests

Search single value

where[column]			   = 1	  // column = 1
where[column][=]		   = 1	  // column = 1
where[column][!]		   = 1    // column != 1
where[column][>]		   = 1	  // column > 1
where[column][<]		   = 1	  // column < 1
where[column][>=]		   = 1	  // column >= 1
where[column][<=]		   = 1	  // column <= 1
where[column][%]		   = "%1" // column LIKE "%1"

Search multiple values

where[column]			   = array(1,5,7)	  // IN (...) (IN can be equal to an OR)
where[column][=]		   = array(1,5,7)	  // IN (...) 
where[column][!]		   = array(1,5,7)	  // NOT IN (...)
where[column][>]		   = array(1,2)	      // column > 1 OR column > 2
where[column][<]		   = array(1,2)	      // column < 1 OR column < 2
where[column][>=]		   = array(1,2)	      // column >= 1 OR column >= 2
where[column][<=]		   = array(1,2)	      // column <= 1 OR column <= 2
where[column][%]		   = array("%1","%2") // column LIKE "%1" OR column LIKE "%2"

Specify column's table

where['table.column'][=] = array(1,5,7)

Compare between two different table columns

where['table_a.column_a'] = 'table_b.column_b'

Compare between different columns of main table

where['column_a'] = 'table_a.column_b'
// OR
where['table_a.column_a'] = 'table_a.column_b'
	
// WRONG
where['column_a'] = 'column_b'

Additional parameters

• order_by: column_name

  Can be and array or a string

  order_by = 'username, name, surname'
  // OR
  order_by = array('username', 'name', 'surname')

  for more specific order direction

  order_by['users.username'] = 'DESC'

  for cast a specific type

  order_by['users.username::varchar'] = 'DESC'
  order_by['users.id::int'] = 'DESC'

• direction: ASC or DESC (default ASC)

• limit: max elements to retrieve

ex: /[database]/[tabel]/[colomn]/[value].[format]?order_by=[column]&direction=[direction]

Documentation

PS: Work only with pgsql and mysql database type at the moment

For get auto-documentation of a database table:

• Documentation index URL format : /[database]/docs/index.[format]
• Documentation URL format: /[database]/docs/[table].[format]

For have a separated file where document your database you can use /docs.php

POST Request

Insert data

Single insert:

• Select the table on URL: /[database]/[table].[format]
• Insert parameter: insert[<column>] = <value>

Multiple insert:

• Select dataset on URL: /[database].[format]
• Insert parameter: insert[<table>][<column>] = <value>

Note: At the moment you can add only one row for table

Examples of POST requests:

Single insert:

POST /dataset/users.json HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

insert[username]=Marco&insert[email][email protected]&insert[password]=3vwjehvdfjhefejjvw&insert[is_active]=1

Multiple insert:

POST /dataset.json HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

insert[users][username]=Marco&insert[users][email][email protected]&insert[users][password]=3vwjehvdfjhefejjvw&insert[users][is_active]=1

POST /dataset.json HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

insert[users][0][username]=Marco&insert[users][0][email][email protected]&insert[users][0][password]=3vwjehvdfjhefejjvw&insert[users][0][is_active]=1&insert[users][1][username]=Brad&insert[users][1][email][email protected]&insert[users][1][password]=erwerwerffweeqewrf&insert[users][1][is_active]=1

PATCH/PUT Request

Update data

Single update:

• Select the row on URL: /[database]/[table]/[id].[format]
• Update parameter: update[<column>] = <value>

Multiple update:

• Select the dataset on URL: /[database].[format]
• Update parameter: update[<table>][values][<column>] = <value>
• Multiple update parameter conditions: update[<table>][where][<column>] = <value>

Note: At the moment you can update only one row for table

Examples of PUT Requests:

Single Update:

PUT /dataset/users/1.json HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

update['username']=Marco&update['email'][email protected]&update['password']=3vwjehvdfjhefejjvw&update['is_active']=1

Multi-table Update:

PUT /dataset.json HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

update[users][values][username]=Marco&update[users][values][email][email protected]&update[users][where][id]=1&update[cities][values][name]=Padova&update[cities][where][id]=1

Multiple Update:

PUT /dataset.json HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

update[users][][values][username]=Marco&update[users][][values][email][email protected]&update[users][][where][id]=1&update[cities][][values][name]=Padova&update[cities][][where][id]=1&update[cities][][values][name]=Milano&update[cities][][where][id]=2

DELETE Request

Delete data

• Select the row on table: /[database]/[table]/[id].[format]

Examples of DELETE Requests:

DELETE /dataset/users/1.json HTTP/1.1
Host: localhost
Access-Token: b279fb1d0708ed81e7a194e0c5d928b6

Hooks

For write hooks you can use plugins/custom folder or edit manually the examples on plugins/actions.php or plugins/filters.php

Remember to name file like this pattern: [FILENAME].hooks.php or it will not be included automatically (else you can include it manually)

Tips

You can use this code for have a database instance and the current user authenticated row:

$user = Auth::getUser(); // User row
$db = API::getDatabase('dataset'); // You can specify dataset. Return PDO Object

Hooks list

https://github.com/marcocesarato/Database-Web-API/wiki/3.2)-Hooks:-List

Most important hooks

Hooks detail

• Filter: sql_restriction

  Options of $permission:

  case 'READ':
  case 'EDIT':
  case 'DELETE':

  Return

  // All denied
  $sql = "'1' = '0'";
  // All allowed
  $sql = "'1' = '1'";

  Examples:

  // Only Created
  $sql = 'created_by = '.$user['id'];
  // Only Team
  $sql = 'created_by IN ('.implode(',',$teams_ids).')';

Clients

PHP Client

Filename: apiclient.class.php

Class name: APIClient

Credits

https://github.com/project-open-data/db-to-api

https://github.com/voku/php-hooks