Table of Contents | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
|
What is an Automation App?
An automation app in PIPEFORCE is like an app for mobile phones. Such an app groups together resources like scripts, templates, configurations and others to solve a certain business task. Any automation pipeline, form or , workflow etc. is part of exactly one app. Such an app is the logical unit of an automation solution having a unique name.
An automation app can be shared with others on the marketplace and installed, updated or uninstalled using commands.
For each automation app, certain access rules can be specified. Furthermore, it is also possible to use staging and versioning for automation apps. They can be developed online using the workbench or offline using source files and the CLI. You can think of automation apps also like “plug-ins” for PIPEFORCE which can be installed and updated at runtime.
Also see this tutorial to learn how to create an automation app in PIPEFORCE.
Info |
---|
For simplicity reasons, automation apps will sometimes also be called just apps in PIPEFORCE. |
App home folder
Typically, all properties (resources) of an app reside in the property store and having a property path with a prefix like this:
...
This path is called the app home folder.
Qualified app naming
In order to avoid a naming clash with other apps from other users which could probably have the same naming as your app, as best practise, you should give the app always a name which follows the reversed domain name package conventions from the Java package specification, which works like this:
...
Info |
---|
The internet domain names of your app prefixes must be valid ones if you want your apps to become trusted in the marketplace. |
App folders and resources
Inside of an app path, there is a certain folder structure at top app home folder level which defines the main resource types of an app. Depending on in which folders properties are stored inside an app, auto-backend tasks could be fired. The typical folder structure looks like this (the folders could vary, depending on your setup):
Naming convention
Any resource inside an app should follow this naming convention:
...
Code Block |
---|
global/app/tld.domain.myapp/pipeline/001-send-invite-email global/app/tld.domain.myapp/pipeline/A1-upload-invoice global/app/tld.domain.myapp/pipeline/add-to-sap global/app/tld.domain.myapp/pipeline/inform-hr |
/config
This optional folder contains all app specific configuration properties.
/app
The default configuration property in here for an app is the config
app
property which is a JSON document.
/form
This optional folder contains all form configurations for any form of the app, whereas the name of the property is the name of the form. For example:
Code Block |
---|
global/app/tld.domain.myapp/form/create-user
global/app/tld.domain.myapp/form/delete-user |
See here for more details: Forms Framework
/function
This optional folder contains all Python FaaS functions which will be automatically deployed when such a property of this app is stored in the property store in this folder.
For more information see: Python Functions
Info |
---|
Any time the hub service restarts (for example because of rescaling), all function scripts inside this folder will be automatically scanned and re-deployed to the FaaS backend if required. |
/i18n
This optional folder contains all internationalization (i18n) files like message files and other resources.
/install
Status | ||||
---|---|---|---|---|
|
This optional folder can also contain pipeline YAML scripts. These pipelines will be executed only in case the app will be installed using the app.install
command.
This is handy in case you would like to do some preparation on the install phase of an app.
The pipelines will be executed in their natural naming order. So if you would like to make sure a pipeline gets executed the very first, make sure it is at the very top of the naming order by using prefixes like 001-
, 002-
and so on.
Info |
---|
The pipelines in this folder wont be auto-executed on save of properties in the portal or on publish using the CLI or on hub restart. Only the initial |
/list
This optional folder contains all list configuration properties for any list of the app, whereas the name of the property is the name of the list. For example:
Code Block |
---|
global/app/tld.domain.myapp/list/all-users
global/app/tld.domain.myapp/list/all-employees |
See here for more details: List Framework.
/object
This optional folder contains any application model (schema) and its instances (if there are any).
/schema
The schema of an object is stored in a property having this pathThis is automatically created if you create an app via CLI or the online property wizard. It is a JSON document of this basic format:
Code Block | ||
---|---|---|
| ||
{
"title": "My App",
"description": "The description of the app",
"icon": "some_icon_id",
"color": "teal-8",
"hidden": false,
"tags": [
"admin"
],
"permissions": {
"read": [
"CAN_APP_io.pipeforce.myapp",
"ROLE_ADMIN",
"ROLE_DEVELOPER"
],
"write": [
"CAN_APP_io.pipeforce.myapp",
"ROLE_ADMIN",
"ROLE_DEVELOPER"
],
"view": [
"CAN_APP_io.pipeforce.myapp",
"ROLE_ADMIN",
"ROLE_DEVELOPER"
]
}
} |
"title"
The short title of the app. Can optionally be a $uri:i18n key for translations.
"description"
The description of the app. Can optionally be a $uri:i18n key for translations. Can be empty.
"icon"
The id of the Google Material icon to be used for this app. See here for listing of all available icons: https://fonts.google.com/icons . Note: Make sure to use the id of the icon not its name. If no icon is set, the system default icon will be used.
"color"
The color to be used for the app icon. If not set, the default color will be used. Also see here: Colors for more details about supported color codes.
"hidden"
An optional flag. If true
, the app wont be shown in the apps listing in the portal. The default value is false
.
"tags"
A list of optional search keywords. This improves finding this app via UI search.
"permissions"
Defines the required permissions for this app.
read
= A list of ROLES, the logged-in user must be member of (at least in one of it) to be able to read values in this app (for example list properties / data stored in this app).write
= A list of ROLES, the logged-in user must be member of (at least in one of it) to be able to write values in this app (for example properties / data to the property store, submit forms).view
= A list of ROLES, the logged-in user must be member of (at least in one of it) to be able to see the app tiles, forms and lists in the web UI.
/data
In case there are data objects (JSON documents), typically based on a schema, they should be stored inside this path structure:
Code Block |
---|
global/app/tld.domain.myapp/data/<NAME>/<UUID> |
Whereas <NAME>
is the name of the data object.
<UUID>
is the unique id of a single data object.
For example:
Code Block |
---|
global/app/tld.domain.myapp/data/person/fa471958-fdb7-4bf6-a0a3-c5e8c782893e |
Each instance property will contain as value the data of the object instance which typically matches the schema, for example:
Code Block | ||
---|---|---|
| ||
{
"firstName": "Homer",
"lastName": "Simpson",
"age": 48,
"gender": "male"
} |
Info |
---|
Note: In case of an update of an app, the data folder wont be deleted by default. See the options of the |
/form
This optional folder contains all form configurations for any form of the app, whereas the name of the property is the name of the form. For example:
Code Block |
---|
global/app/tld.domain.myapp/form/create-user
global/app/tld.domain.myapp/form/delete-user |
See here for more details: Forms Framework
/function
This optional folder contains all Python FaaS functions which will be automatically deployed when such a property of this app is stored in the property store in this folder.
For more information see: Python Functions
Info |
---|
Any time the hub service restarts (for example because of rescaling), all function scripts inside this folder will be automatically scanned and re-deployed to the FaaS backend if required. |
/i18n
This optional folder contains all internationalization (i18n) files like message files and other resources.
Also see: Internationalization (i18n).
/install
Status | ||||
---|---|---|---|---|
|
This optional folder can also contain pipeline YAML scripts. These pipelines will be executed only in case the app will be installed using the app.install
or app.update
command.
This is handy in case you would like to do some preparation on the install or update phase of an app like migrations for example.
The pipelines will be executed in their natural naming order. So if you would like to make sure a pipeline gets executed the very first, make sure it is at the very top of the naming order by using prefixes like 001-
, 002-
and so on.
Info |
---|
The pipelines in this folder wont be auto-executed on save of properties in the portal or on publish using the CLI or on hub restart. Only the initial |
/list
This optional folder contains all list configuration properties for any list of the app, whereas the name of the property is the name of the list. For example:
Code Block |
---|
global/app/tld.domain.myapp/list/all-users
global/app/tld.domain.myapp/list/all-employees |
See here for more details: List Framework
/object
(deprecated)
Note | ||||||
---|---|---|---|---|---|---|
DEPRECATED since
/data folder instead. |
This optional folder contains any application model (schema) and its instances (if there are any).
/schema
(deprecated)
Note | ||||||
---|---|---|---|---|---|---|
Folder
/schema folder in app root instead. |
The schema of an object is stored in a property having this path:
Code Block |
---|
global/app/tld.domain.myapp/object/<NAME>/<VERSION>/schema |
Whereas <NAME>
is the name of the object.
<VERSION>
is the version of the object schema.
For example:
Code Block |
---|
object/person/v1/schema |
The schema property typically contains as value a JSON schema, which describes this object. For the person object, the schema could, for example, look like this:
Code Block | ||
---|---|---|
| ||
{
"type": "object",
"properties": {
"firstName": {
"type": "string"
},
"lastName": {
"type": "string"
},
"age": {
"type": "number"
},
"gender": {
"type": "string",
"enum": ["male", "female", "neutral"]
}
}
} |
See the JSON schema section for more details: JSON Schema
/instance
(deprecated)
Note | ||||||
---|---|---|---|---|---|---|
DEPRECATED since
/data folder instead. |
In case there are object instances (JSON documents) based on a schema, they should be typically stored inside this path structure:
Code Block |
---|
global/app/tld.domain.myapp/object/<NAME>/<VERSION>/instance/schema<UUID> |
Whereas <NAME>
is the name of the object.
<VERSION>
is the version of the object schema.
<UUID>
is the unique id of a single object.
For example:
Code Block |
---|
global/app/tld.domain.myapp/object/person/v1/schema |
...
/v1/instance/fa471958-fdb7-4bf6-a0a3-c5e8c782893e |
Each instance property will contain as value the data of the object instance which matches the object schema, for example:
Code Block | ||
---|---|---|
| ||
{ "typefirstName": "objectHomer", "propertieslastName": {"Simpson", "firstName": { age": 48, "typegender": "stringmale" }, "lastName": { "type": "string" }, "age": { "type": "number" }, "gender": { "type": "string", "enum": ["male", "female", "neutral"] } } } |
See the JSON schema section for more details: JSON Schema
/instance
...
} |
/pipeline
This optional folder contains all persisted pipeline YAML scripts for the given app. A pipeline can be seen as the business logic or data integration part of an application.
Each property name corresponds with the name of the pipeline and contains as value the pipeline YAML script.
Examples:
Code Block |
---|
global/app/tld.domain.myapp/pipeline/001-send-invite-email
global/app/tld.domain.myapp/pipeline/add-to-sap
global/app/tld.domain.myapp/pipeline/inform-hr |
Such a pipeline YAML script could look like this:
Code Block | ||
---|---|---|
| ||
pipeline:
- mail.send:
to: hr@company.de
subject: "A new employee was addded!" |
Also see Automation Pipeline for more details how to create persisted pipelines.
Info |
---|
Any time the hub service restarts (for example because of rescaling), all pipeline scripts inside this folder (recursively) will be automatically scanned and any trigger command inside each pipeline will be re-registered (for example jobs, message listeners, …). |
/report
This optional folder can contain all report configurations.
See here for more details: Reporting Framework
/resource
This optional folder can contain any resource required inside the app. For example images, CSS files or similar.
/schema
The schema of an data object is stored in a property having this path:
Code Block |
---|
global/app/tld.domain.myapp/objectschema/<NAME>/<VERSION>/instance/<UUID> |
Whereas <NAME>
is the name of the object.
<VERSION>
is the version of the object schema.
<UUID>
is the unique id of a single objectdata object this schema stands for.
For example:
Code Block |
---|
global/app/tld.domain.myapp/objectschema/person/v1/instance/fa471958-fdb7-4bf6-a0a3-c5e8c782893e |
...
The schema property typically contains as value a JSON schema, which describes this object. For the person object, the schema could, for example, look like this:
Code Block | ||
---|---|---|
| ||
{ "type": "object", "properties": { "firstName": { "type": "Homer",string" }, "lastName": "Simpson",{ "age": 48, "gendertype": "male" } |
/pipeline
This optional folder contains all persisted pipeline YAML scripts for the given app. A pipeline can be seen as the business logic or data integration part of an application.
Each property name corresponds with the name of the pipeline and contains as value the pipeline YAML script.
Examples:
Code Block |
---|
global/app/tld.domain.myapp/pipeline/001-send-invite-email
global/app/tld.domain.myapp/pipeline/add-to-sap
global/app/tld.domain.myapp/pipeline/inform-hr |
Such a pipeline YAML script could look like this:
Code Block | ||
---|---|---|
| ||
pipeline: - mail.send:string" }, "age": { "type": "number" }, "gender": { to"type": hr@company.de"string", subject: "A new employee was addded!" |
Also see Command and Pipeline for more details how to create persisted pipelines.
Info |
---|
Any time the hub service restarts (for example because of rescaling), all pipeline scripts inside this folder (recursively) will be automatically scanned and any trigger command inside each pipeline will be re-registered (for example jobs, message listeners, …). |
/report
This optional folder can contain all report configurations.
See here for more details: Reporting Framework .
/resource
This optional folder can contain any resource required inside the app. For example images, CSS files or similar.
...
"enum": ["male", "female", "neutral"]
}
}
} |
See the JSON schema section for more details: JSON Schema
/setup
(deprecated)
Note | ||||||
---|---|---|---|---|---|---|
DEPRECATED since
/install folder instead. |
This optional folder can also contain pipeline YAML scripts. These pipelines will be executed only in case the app will be installed using the app.install
command.
This is handy in case you would like to do some preparation on setup phase of an app.
The pipelines will be executed in their natural naming order. So if you would like to make sure a pipeline gets executed the very first, make sure it is at the very top of the naming order by using prefixes like 001-
, 002-
and so on.
Info |
---|
The pipelines in this folder wont be auto-executed on save of properties in the portal or on publish using the CLI or on hub restart. Only the inital |
/script
(deprecated)
Note | ||||||
---|---|---|---|---|---|---|
DEPRECATED since
/installfunction folder instead. |
This optional folder can also contain pipeline YAML scripts. These pipelines will be executed only in case the app will be installed using the app.install
command.
This is handy in case you would like to do some preparation on setup phase of an app.
The pipelines will be executed in their natural naming order. So if you would like to make sure a pipeline gets executed the very first, make sure it is at the very top of the naming order by using prefixes like 001-
, 002-
and so on.
Info |
---|
The pipelines in this folder wont be auto-executed on save of properties in the portal or on publish using the CLI or on hub restart. Only the inital |
/script
Note | ||||||
---|---|---|---|---|---|---|
DEPRECATED since
/function folder instead. |
...
/system
This folder contains all system related backend data of this specific app. For example tmp data, workflow or form data. It is not meant for manual handling and you should not be required to touch or change content in it. It is fully managed by the platform.
Info |
---|
Note: By default this folder wont be deleted on |
/template
This optional folder should contains all templates like FreeMarker templates an others.
See here for more details: Template Transformation .
/test
This optional folder typically contains pipelines for tests only. Whenever necessary, PIPEFORCE automatically executes the test pipelines inside this folder to make sure the app is working as expected. Therefore you have to make sure that these tests can be executed at any time and are fully reentrant (once a test has been finished it can be executed again as often as necessary).
...
The property contains the test pipeline as value. Such a test pipeline could look like this:
Code Block | ||
---|---|---|
| ||
pipeline: - testassert: assertTrue(false)true: ${someTestExpression} |
See here for more details: App Testing .Automation Apps
/uninstall
Status | ||||
---|---|---|---|---|
|
...
Info |
---|
The pipelines in this folder wont be auto-executed on save of properties in the portal or on publish using the CLI or on hub restart. So triggers in these pipelines wont work. Only a call of |
/update
Status | ||||
---|---|---|---|---|
|
This optional folder can also contain folder is optional and contains all migration pipeline YAML scripts in order to migrate from one app version to another. These pipelines will be executed only in case the app will be updated installed using the app.update
command.This is handy in case you would like to do some update and migration tasks on update.
The pipelines will be executed in their natural naming order. So if you would like to make sure a pipeline gets executed the very first, make sure it is at the very top of the naming order by using prefixes like 001-
, 002-
and so on.
Info |
---|
The pipelines in this folder wont be auto-executed on save of properties in the portal or on publish using the CLI or on hub restart. So triggers in these pipelines wont work. Only a call of |
/workflow
This optional folder contains any BPMN workflow files defining a business process.
...
Code Block |
---|
global/app/tld.domain.myapp/workflow/approve-new-employee |
See here for more details: BPMN Workflows
Developing an automation app
In its simple case you can manage all properties of an app in the property store with the property.*
commands and the CLI using pi publish
or using the online workbench.
...