Files
threshold-ui/addons/api/addon/services/scope.js
T

152 lines
3.9 KiB
JavaScript

import Service from '@ember/service';
import { tracked } from '@glimmer/tracking';
import { computed } from '@ember/object';
import { inject as service } from '@ember/service';
/**
* This module includes two parts:
*
* - `Scope` class: a data structure representing an org/project tuple
* - `ScopeService`: manages the scope and exposes it to the application
*/
/**
* Represents a _scope_, which is defined as an organization/project tuple.
* This class is intended as a simple data structure and is used by the
* scope service (see below).
* Immutable.
*/
class Scope {
/**
* @type {?object}
* @private
*/
#org = null;
/**
* @type {?object}
* @private
*/
#project = null;
/**
* Initializes an instance of `Scope` with optional org and/or project.
* @param {?object} org
* @param {?object} project
*/
constructor(org = null, project = null) {
if (org) this.#org = org;
if (project) this.#project = project;
}
/**
* Org for this scope. Immutable.
* @type {?object}
*/
get org() {
return this.#org;
}
/**
* Project for this scope. If org is null, so is project.
* @type {?object}
*/
get project() {
return this.org ? this.#project : null;
}
/**
* Returns the JSON serialization of this scope. This is necessary because
* scopes may be persisted in storage with their JSON representation.
* @return {?object}
*/
toJSON() {
const { org, project } = this;
return { org, project };
}
}
export { Scope };
/**
* The scope service manages and persists the current scope. To interact with
* this service, get and set `service.org` or `service.project` directly, rather
* than using the tracked `scope` member, which is internal.
*
* A scope is valid only if the user has access to the organization and project
* specified in the scope, and they exist. However, validating against
* available organizations and projects is outside the scope (tee hee) of this
* service because it is used by the adapter mixin to generate API paths, which
* could lead to problems were this service to attempt to fetch data.
*
* For now, should an invalid scope arise, it is left to the user to refresh,
* since they are likely to encounter 404s. This is an acceptable tradeoff for
* an initial release because invalid scopes are considered uncommon.
*/
export default class ScopeService extends Service {
/**
* The storage service is used to persist the current scope.
* @type {StorageService}
*/
@service storage;
/**
* The currently selected scope. This is an internal member only.
* @type {?Scope}
*/
@tracked scope = new Scope();
/**
* When setting org, `service.org = {...}`, a new scope is instantiated,
* since scopes are immutable. Also persists scope on set.
* @type {?object}
*/
@computed('scope.org')
get org() {
return this.scope ? this.scope.org : null;
}
set org(org) {
this.scope = new Scope(org, this.project);
this.saveScope(this.scope);
}
/**
* When setting project, `service.project = {...}`, a new scope is
* instantiated, since scopes are immutable. Also persists scope on set.
* @type {?object}
*/
@computed('scope.project')
get project() {
return this.scope ? this.scope.project : null;
}
set project(project) {
this.scope = new Scope(this.org, project);
this.saveScope(this.scope);
}
/**
* Saves the current scope into storage.
*/
saveScope(scope) {
this.storage.setItem('scope', scope);
}
/**
* Retrieves a saved scope from storage, if any.
* @return {?Scope}
*/
fetchScope() {
const json = this.storage.getItem('scope');
if (json) return new Scope(json.org, json.project);
}
/**
* Initializes the service's current scope from storage, if any.
*/
init() {
const savedScope = this.fetchScope();
if (savedScope) this.scope = savedScope;
super.init(...arguments);
}
}