Tasks and Roles

iHRIS utiliza un mecanismo de seguridad basado en Roles y Tareas para limitar el acceso a las diversas partes del sistema. A un usuario se le asigna un rol y un rol es un conjunto de tareas que ese rol puede realizar.

Este artículo describe como se definen los roles y tareas en Datos Magic y utilizados por el sistema iHRIS.

Roles

Un rol es una colección de tareas que pueden asignarse a la cuenta de un usuario.

• los nombres de los roles se definen como los secundarios de ruta de magic data /I2CE/roles/names
• un rol $role tiene un display definido en /I2CE/roles/names/$role/display_name
• los roles pueden heredar tareas de otro al agregarlo como el valor de un nodo secundario de /I2CE/roles/names/$role/trickle_up
  Por ejemplo, en iHRIS Manage en el /I2CE/roles/names/hr_staff tenemos:
  • display_name => HR Staff
  • trickle_up => Array
    • 0 => admin
    • 1 => hr_manager

dice que el rol de hr_staff aparece como 'HR Staff' y que un hr_manager o admin tiene todas las tareas que tiene un hr_staff.

Tareas

Una tarea puede ser tanto una colección de sub-tareas que tiene esta tarea y la descripción de alguna acción cuyo permiso se puede verificar. La información de las tareas se guarda en datos magic bajo /I2CE/tasks/. Para crear una tarea se crea un nodo secundario de tipo escalar de /I2CE/tasks/task_description. El nombre del nodo es el nombre utilizado para referenciar la tarea. El valor del nodo es la descripción de la tarea que se muestra en la página de Manejo de Tareas y Roles. Por ejemplo, el nodo de magic data /I2CE/tasks/task_description puede verse así:

• custom_report_admin => Permite la administración del Sistema de Informes Personalizados
• custom_reports_can_access => Permite acceso mínimo al Sistema de Informes Personalizados
• custom_reports_delete => Permite eliminación de datos acerca de los informes personalizados
• custom_reports_can_access_relationships => Permite acceso a las Relaciones de Informes Personalizados

Puede definir la sub-tarea de una tarea $task al especificar /I2CE/tasks/task_trickle_down_task. Por ejemplo, el nodo de datos magic /I2CE/tasks/task_trickle_down/custom_reports_admin puede verse así:

• 0 => custom_reports_can_access
• 1 => custom_reports_delete_reports

lo que dice que la tarea de 'custom_report_admin' tiene todas las tareas y derechos definidos por 'custom_reports_can_access' y 'custom_reports_delete_reports.'

Las tareas que se asignan a un rol $role son los valores de los secundarios bajo /I2CE/tasks/role_trickle_down/$role

Un usuario con el rol 'admin' tiene todas las tareas.

Uses of Tasks and Roles

The tasks and roles are used in several places:

• The main I2CE_Page class checks for basic permission for the page.
• Several pages perform checks for specific roles and tasks in their action() method.
• Just before displaying the HTML the I2CE_Template, class verifies that all tasks, roles and permissions are satisfied on each node.

Task and Role Administration

For deployment across many computers, tasks and roles should be set up in an appropriate module configuration file.

For setting tasks and roles dynamically, the module 'tasks-roles' provides the page named 'roles' and the page named 'tasks' that allows creating new roles and tasks as well as defining the permission inheritance.

Permissions and the Permission Parser

The permission parser allows logical expressions to combine severals types permissions, such as task, roles, into a permission string.

We can assign tasks, roles and permissions to DOM nodes by:

• Setting the attribute role.
  If the values is X, this results in the permission string role(X) which is passed to the permission parser
• Setting the attribute task.
  If the values is X, this results in the permission string task(X) which is passed to the permission parser
• Setting the attribute permission.

If the node fails any of the role, task or permission checks it will remove the node

Permission Types: task and role

The task and role type permissions are formed by surrounding a role name with role() or a task name with task(). For example, you can create the following permission string:

(task(can_edit_database_list_facility_type) & task(can_edit_database_list_fav_color) || role(admin)

By default, tasks and roles are 'OR'ed together so the following are all the same:

• task(can_edit_database_list_facility_type) or task(can_edit_database_list_fav_color)
• task(can_edit_database_list_facility_type) | task(can_edit_database_list_fav_color)
• task(can_edit_database_list_facility_type) task(can_edit_database_list_fav_color)
• task(can_edit_database_list_facility_type,can_edit_database_list_fav_color)
• task(can_edit_database_list_facility_type can_edit_database_list_fav_color)
• task(can_edit_database_list_facility_type|can_edit_database_list_fav_color)

Permission Type: module

Any public function of a module class can be called by the permission parser. For example, suppose that the module 'my_module' has a method 'my_method()' then we can use as the permission string with arguments:

module('my_module','my_method', [arg1], ... , [argN])

which would results in the call:

$module->my_method($arg1,..,$argN)

where $module is the instance of the module class for the module 'my_module.'

Permission Type: form

The 'forms' module adds in the form type. The permission string with arguments:

form('form_name', 'form_method', [arg1] , .., [argN])

results in the call:

$form->form_method($arg1,..,$argN)

where $form is the result of getting the form by the name of 'form_name' via template data for node (if there was any) the permission string was assigned to.

Arguments

A permission type (such as role, task, form or module) in a permission string behaves essentially like a function. Suppose that we have the general shape for a piece of a permission string:

type([arg1],[arg2],...,[argN])

Then this results in the method call:

$permissionParsrer->hasPermission_$type($node,$args)

where $node is the DOMNode the permission string was called on and $args is the array($arg1,..$argN). The permission parser turns [argM] into $argM according to the following rules:

• if [argM] starts with a $ then it refers to template data and the following rules apply:
  • The string has the form $abcd. The value of $argM becomes the template display data with name 'abcd.'
  • The string has the form ${WXYZ}abcd. The value of $argM becomes the template data with category 'WXYZ' and with name 'abcd.'
• <NODE> becomes the instance of DOMNode (if any) that the permission string was called on
• <TEMPLATE> becomes the instance of I2CE_Template (if any) that the permission parser was called on
• <USER> becomes the instance of I2CE_User that is this session
• if [argM] starts with a single quote ' then it is a string until the next non-escaped ' is found
• if [argM] starts with a double quote " then is is a string until the next non-escaped " is found.
  In addition the following substitution rules apply:
  • any substring starting with $ and consisting of alpha-numeric characters, - or _ is interpreted as template display data to be substituted
    For example "my name is $name" becomes "my name is Joe" if the template data named 'name' and with type DISPLAY is "Joe"
  • any substring starting with {$ is read until an enclosing } is found. The string between the ${ and } is the name of DISPLAY template data which is then substituted.
  • To prevent the above, { and $ may be escaped with a \
• any other string of alpha-numeric characters (and a few permitted punctuation marks) is interpreted as a string

Arguments may be separated by a comma a space or a |.

New Types

A module can add in a fuzzy method of the form hasPermision_$type to the I2CE_PermissionParser class to enable a new permission type. For example the 'forms' module does this by adding in a new permission type 'form.'