The Data API extension provides developers with the power to rapidly develop APIs against their Preside data models. With a few simple object annotations and optional i18n resource keys, developers are able to provide full CRUD APIs to the data model with beautifully rendered API documentation.
Note: Version 3.0.0 requires Preside version 10.11.0 and above.
The default root URI of the API is /api/data/v1/
(see Namespaces and multiple APIs, further below). OpenAPI v3 specification can be browsed at /api/data/v1/docs/spec/
and HTML documentation based on the spec can be found at /api/data/v1/docs/html/
(or /api/data/v1/docs/swagger/
, if you prefer).
As of v3.3.0, two feature flags are available that bring backward compatibility from previous versions. New default behaviour is to return null
values for empty strings and numbers. Previously, the extension would return an empty string, ""
. To get the old behaviour back, you may set the following features to false
:
settings.features.dataApiUseNullForNumerics.enabled = false;
settings.features.dataApiUseNullForStrings.enabled = false;
The bare minimum configuration to enable an object for the API is to set @dataApiEnabled true
:
/**
* @dataApiEnabled true
*
*/
component {
// ...
}
Additional optional annotation options at the object level are:
dataApiEntityName
: Alternative entity name to use in API urls, e.g. instead ofcrm_contact
, you may wish to usecontact
.dataApiCategory
: For the HTML documentation. Allows sub-groups of entities. Especially useful for large APIs.dataApiQueueEnabled
: Whether or not the change queue is enabled for this objectdataApiQueue
: Specific queue name for this objectdataApiQueueDeleteDetail
: Whether or not the deleted record detail is available in the change queue data for this objectdataApiSortOrder
: Sort order for paginated results. Default is date last modified ascending.dataApiSavedFilters
: Comma-separated list of saved filters to apply to all requests to this object (e.g. only return active records)dataApiIgnoreDefaultFilters
: Comma-separated list of ignore default filters to apply to all requests to this objectdataApiVerbs
: Supported REST HTTP Verbs. If not supplied, all verbs and operations are supported (i.e. GET, POST, PUT and DELETE)dataApiFields
: Fields to return in GET requests (defaults to all non-excluded fields)dataApiUpsertFields
: Fields to accept in POST/PUT request (defaults todataApiFields
)dataApiExcludeFields
: Fields to exclude from GET API callsdataApiUpsertExcludeFields
: Fields not to accept in POST/PUT requests (defaults todataApiExcludeFields
)dataApiFilterFields
: Fields to allow as simple filters for paginated GET requests (defaults to foreign keys, boolean and enum fields)dataApiAllowIdInsert
: Whether or not to allow the ID field to be set during a POST operation to create a new recorddataApiQueueRelevantFields
: Fields that are relevant to trigger a record update to be queued. If not used, then any data change will be queued. If used then only the specified fields will be examined and in case of atomic-changes for the queue only the relevant field changes will be included in the queue item. Inserts and deletes are always queued, only during updates this annotation is evaluated.
Object properties support the following optional annotations:
dataApiAlias
: An alternative name to use for the property in the API. i.e. instead ofcontact_status
you may wish to just usestatus
through the API.dataApiRenderer
: A non-default renderer to use when returning data through the API. See custom renderers, below.dataApiDerivative
: For image assets, specify the derivative to be used when rendering URLs.dataApiType
: For the documentation. The data type of the property, e.g. 'string'.dataApiFormat
: For the documentation. The format for the property, e.g. 'Email address'.dataApiEnabled
: Whether or not this property should be included in the APIdataApiUpsertEnabled
: Whether or not this property should be included in the POST/PUT operations.dataApiRenderEmptyValues
: Whether to force always calling a custom render when the value being processed is empty or null
If you specify a non-default renderer for an object property, it will be rendered using Preside's content rendering system. For example, the following property definition specifies a myCustomRenderer
renderer:
property name="my_prop" dataApiAlias="myProp" dataApiRenderer="myCustomRenderer";
By default for performance reasons, empty/null values are skipped, to force always calling the custom renderer, set the following property dataApiRenderEmptyValues=true
.
To implement this, you will need a corresponding viewlet at renderers.content.myCustomRenderer.dataapi
or renderers.content.myCustomRenderer.default
(a renderer context of dataapi
will be used and the system will fallback to the default renderer if that context is not implemented).
For example:
// /handlers/renderers/content/MyCustomRenderer.cfc
component {
private string function dataApi( event, rc, prc, args={} ){
var value = args.data ?: "";
return renderView( view="/renderers/content/myCustomRenderer/dataApi", args={ value=value } );
}
}
All of the labelling and text in the generated documentation can be found at /i18n/dataapi.properties
and you should refer to that when customizing the default text. A bare minimum override might look like:
api.title=My Application's API
api.description=This is my awesome application's API and here is some general information about it.\n\
\n\
Team awesome xxx.
api.version=v2.0
api.favicon=data:image/x-icon;base64,iVBOR...
If you annotate your objects with a @dataApiCategory
property, your categories can be documented with:
category.my_category.name=My Category
category.my_category.description=Markdown enabled description of my category.
category.my_category.sort.order=20
The following set of .properties
file keys can be added per object in your own dataapi.properties
files to customize the documentation per entity/field:
# OBJECT LEVEL:
dataapi:entity.my_entity.name=My entities
dataapi:entity.my_entity.name.singular=My entity
dataapi:entity.my_entity.description=Description of my entity (or API description for this section of the docs)
dataapi:entity.my_entity.sort.order=10
dataapi:operation.my_entity.get.description=Description for the paginated GET operation for your entity
dataapi:operation.my_entity.get.200.description=Description for the successful (200) response documentation for paginated GET requests for your entity
dataapi:operation.my_entity.get.by.id.description=Description for the GET /{recordId}/ operation for your entity
dataapi:operation.my_entity.get.by.id.200.description=Description for the successful (200) response documentation for GET /{recordId}/ requests for your entity
dataapi:operation.my_entity.get.by.id.404.description=Description for the not found (404) response documentation for GET /{recordId}/ requests for your entity
dataapi:operation.my_entity.put.description=Description for the PUT / (batch update) operation for your entity
dataapi:operation.my_entity.put.body.description=Description for the http json body for the PUT / (batch update) operation for your entity
dataapi:operation.my_entity.put.200.description=Description for the successful (200) response documentation for PUT / requests for your entity
dataapi:operation.my_entity.put.422.description=Description for the validation failed (422) response documentation for PUT / requests for your entity
dataapi:operation.my_entity.put.by.id.description=Description for the PUT /{recordid}/ (single record update) operation for your entity
dataapi:operation.my_entity.put.by.id.body.description=Description for the http json body for the PUT /{recordid}/ (single record update) operation for your entity
dataapi:operation.my_entity.put.by.id.200.description=Description for the successful (200) response documentation for PUT /{recordid}/ requests for your entity
dataapi:operation.my_entity.put.by.id.422.description=Description for the validation failed (422) response documentation for PUT /{recordid}/ requests for your entity
dataapi:operation.my_entity.put.by.id.404.description=Description for the record not found (404) response documentation for PUT /{recordid}/ requests for your entity
dataapi:operation.my_entity.post.description=Description of the POST / (batch insert) operation for your entity
dataapi:operation.my_entity.post.body.description=Description for the http json body for the POST / (batch insert) operation for your entity
dataapi:operation.my_entity.post.200.description=Description for the successful (200) response documentation for POST / requests for your entity
dataapi:operation.my_entity.post.422.description=Description for the validation failed (422) response documentation for POST / requests for your entity
dataapi:operation.my_entity.delete.description=Description of the DELETE /{recordid}/ operation for your entity
dataapi:operation.my_entity.delete.200.description=Description of the successful (200) response documentation for DELETE /{recordId}/ operations for your entity
# Field level
dataapi:operation.my_entity.get.params.fields.my_field.description=Description for *filter* field
dataapi:entity.my_entity.field.my_field.description=Description of field
The following interception points are used to allow you to more deeply customize the integration.
Fired when generating OpenApi v3 specification for the API. A spec
struct will be present in the interceptorArgs
that you can use to augment the specification.
Fired before selecting data through the API. Receives the following keys in the interceptData
:
selectDataArgs
: Arguments that will be passed to theselectData()
callentity
: Name of the entity being operated on
Fired after selecting data through the API. Receives the following keys in the interceptData
:
selectDataArgs
: Arguments that were passed to theselectData()
callentity
: Name of the entity being operated ondata
: Rendered and prepared data that will be returned to the API caller
Fires before inserting data through the API. Receives the following keys in the interceptData
:
insertDataArgs
: Arguments that will be passed to theinsertData()
callentity
: Name of the entity being operated onrecord
: The data that will be inserted (struct)
Fires after inserting data through the API. Receives the following keys in the interceptData
:
insertDataArgs
: Arguments that were passed to theinsertData()
callentity
: Name of the entity being operated onrecord
: The data that will be inserted (struct)newId
: Newly created record ID
Fires before updating data through the API. Receives the following keys in the interceptData
:
updateDataArgs
: Arguments that will be passed to theupdateData()
callentity
: Name of the entity being operated onrecordId
: ID of the record to be updateddata
: The data that will be inserted (struct)
Fires after updating data through the API. Receives the following keys in the interceptData
:
updateDataArgs
: Arguments that were passed to theupdateData()
callentity
: Name of the entity being operated onrecordId
: ID of the record to be updateddata
: The data that will be inserted (struct)
Fires before deleting data through the API. Receives the following keys in the interceptData
:
deleteDataArgs
: Arguments that will be passed to thedeleteData()
callentity
: Name of the entity being operated onrecordId
: ID of the record to be deleted
Fires after deleting data through the API. Receives the following keys in the interceptData
:
deleteDataArgs
: Arguments that were passed to thedeleteData()
callentity
: Name of the entity being operated onrecordId
: ID of the record to be deleted
By default, they system enables a queue feature: settings.features.dataApiQueue
. When enabled, API users can be subscribed, through the admin UI, to listen for data changes to all, or a number, of entities in the system.
As of v3.0.0, you have the ability to configure multiple queue endpoints per API (see also, Namespaces and multiple APIs, below). This can be used to group entities into queues so that multiple different services can process the queues independently. The available settings per queue are:
pageSize
: Number of records returned with each call to queue. Default is 1atomicChanges
: Whether or not the queue should return atomic changes. If true, each item in the queue will contain only the fields that have changed. If false, default, each item in the queue will contain the latest state of the record. Default is false.returnTotalRecords
: Whether or not to return X-Total-Records header with the total queue size when getting items from the queue (default istrue
)
Queues are defined in the Preside rest API configuration in Config.cfc
. To add a queue definition to the default API:
settings.rest.apis[ "/data/v1" ].dataApiQueueEnabled = true;
settings.rest.apis[ "/data/v1" ].dataApiQueues = {
default = { pageSize=100, atomicChanges=true } // the default queue
, highvolume = { pageSize=1000, atomicChanges=false, returnTotalRecords=false }
};
Annotate your preside object CFC with @dataApiQueueEnabled
(default is true) and @dataApiQueue
(default is default) properties. For example:
/**
* @dataApiQueueEnabled false
*/
or
/**
* @dataApiQueue highvolume
*/
To prevent data api queue to be created when you insert data into an object, you can set skipDataApiQueue = true
as argument in the insertData
function. For example:
$getPresideObject( "example_object" ).insertData(
data = { example_field_name = "example_value" }
, skipDataApiQueue = true
);
Introduced in v2.0.0
By default, the API is exposed at /api/data/v1/
. However, there will be occasions when you want to expose your data in different ways for different purposes. Or, if you are writing an extension, you may want to namespace your API so that it does not clash with any existing default API implementation. You can also use this feature to host multiple versions of your API concurrently.
With just a small amount of configuration, you can use all of the Data API's functionality in a separate, namespaced instance. First, configure the endpoints in your Config.cfc
:
settings.rest.apis[ "/myGroovyApi/v1" ] = {
authProvider = "dataApi"
, description = "REST API to expose data with an alternate structure"
, dataApiNamespace = "myGroovyApi"
, dataApiQueueEnabled = true
, dataApiDefaults = { allowIdInsert=true, verbs="GET" }
, dataApiQueues = {
default = { pageSize=100, atomicChanges=true }
}
};
settings.rest.apis[ "/myGroovyApi/v1/docs" ] = {
description = "Documentation for myGroovyApi REST API (no authentication required)"
, dataApiNamespace = "myGroovyApi"
, dataApiDocs = true
};
A few things to note here:
- The key within
settings.rest.apis
(e.g./myGroovyApi/v1
) is the base URI for the API. This will have/api
prepended to it in the full URL. dataApiNamespace
is the namespace for the alternate API, and will be used when configuring objects. This will usually be the first part of the URI, but does not need to be.dataApiDocs
marks that this is the endpoint for the Swagger documentation. This whole endpoint could be omitted if you do not require the automatic document generation.- Use
authProvider
to mark an endpoint as needing authentication. If you omit this, the resource will not require authentication. In the API Manager in Preside, you can allow users to have access to individual APIs - so a user could have access to/api/myGroovyApi/v1
but not to the default/api/data/v1
, if you wish.
You can annotate your objects using exactly the same annotations as described above, but with :{namespace}
appended. For example:
/**
* @dataApiEnabled true
* @dataApiEnabled:myGroovyApi true
* @dataApiVerbs:myGroovyApi GET
*
*/
component {
property name="label" dataApiAlias:myGroovyApi="some_other_label";
}
This would enable this object both for the default /data/v1
API, and for your custom /myGroovyApi/v1
API. However, for myGroovyApi
the object would only allow GET
access, and the label field would be called some_other_label
instead of label
.
You could also use this to specify an alternate renderer for a property, e.g. dataApiRenderer:myGroovyApi="alternateRenderer"
.
Note that namespaces do not inherit any annotations from the default API. Any annotations must be made explicitly with the :{namespace}
suffix.
You can use all the same label and description customisations in your /i18n/dataapi.properties
file as above; simply prefix the key name with {namespace}.
. For example:
myGroovyApi.api.title=My Second API
myGroovyApi.api.description=This is an alternate API to my application
myGroovyApi.entity.my_entity.name=My alternate entity
Unlike annotations, i18n properties will cascade up to the defaults. So if you do not make any customisations, you will see all the default text.
All the same interceptors will run when actions are taken in a namespaced API. However, the interception point name will have _{namespace}
as a suffix. Again, there is no fallback to the default interceptor, so any interceptors will need to be explicitly defined for your namespace. For example:
public void function preDataApiSelectData( event, interceptData ) {
// action to take for preDataApiSelectData on the default API
}
public void function preDataApiSelectData_myGroovyApi( event, interceptData ) {
// action to take for preDataApiSelectData on the myGroovyApi API
}
Provided that your APIs use the 'dataApi' auth provider (default), you will be able to manage user access to the APIs through the admin interface.
- Navigate to: System > API Manager
- Find the API you want to manage & click 'Data API authentication'
- Click Add user
- Enter a user name (can lookup existing users also) and configure access to individual entities, their verbs and optionally queue access