Pipelines

What Is a Pipeline?

A pipeline is an extendable list of subsequently executed functions called steps. Your app can call a pipeline by name through a pipeline request. By default, you provide a pipeline step by an extension.

Pipeline Example

For example, you app displays a catalog of products, and users can navigate to a certain product within your app. To navigate to the product, you create a getProduct pipeline. The getProduct pipeline includes a product description, a price, and a picture that displays on the product page of your app. You can create separate pipeline steps for different parts of the product, such as getting the product description, calculating the price, and adding a picture to the product. You create all of these steps through a catalog extension.

Every pipeline is extendable, meaning you can add more steps to it. For example, you can extend the getProduct pipeline with a bonus point count for the product. The new step does not need to be a catalog extension. For example, the bonus point count can be in a bonus points extension instead of the catalog extension.

When you send a pipeline request to this pipeline, the result is a product page with bonus points attached to it.

Structure

Pipelines are specified in a .json file and are a part of an extension. The pipeline consists of pipeline metadata, an input/output definition, and a list of steps.

Metadata

The metadata includes the following information:

• A pipeline id, which is the pipeline name and filename
• A flag indicating whether the pipeline is callable from the outside. If false, you can only call the pipeline by another pipeline in the same environment.

Input/output Definitions

The input definitions specify pipeline requirements to generate the defined output. These specifications must be met by the request and by the steps generating the output; otherwise, the pipeline request fails. Refer to the Pipeline Reference.

Step List

A pipeline is a list of steps. These steps must consume all provided pipeline input to generate the specified output.

Regular and trusted pipelines

You can run a pipeline within a trusted environment or the regular environment.

Pipelines and steps running in a trusted environment have access to certain steps that regular pipelines can not access. You use a pipeline within a trusted environment for sensitive data and actions such as payments or user login.

Pipelines can only use steps that are in their respective environment. For example, pipelines in a trusted environment can only use steps of a trusted extension. However, a pipeline can call another pipeline in a different environment.

Extensions and pipelines marked as trusted are carefully reviewed by Shopgate. The review process looks for malicious code.

Naming conventions

Here is the default naming convention for a pipeline:

<organization>.<collection>.<action>.<version>.json

Example: shopgate.catalog.getProducts.v1.json

Pipeline requests

You can run pipeline requests directly to your Sandbox App using the SDK.

When requesting a pipeline, remember there is a 20 second timeout. If the pipeline takes longer than 20 seconds, you will receive an ETIMEOUT error.