A Crossplane composition function that lets you compose resources using Python.

Provide a Python script that defines a compose function with this signature:

from crossplane.function.proto.v1 import run_function_pb2 as fnv1

def compose(req: fnv1.RunFunctionRequest, rsp: fnv1.RunFunctionResponse)

function-python passes the compose function a request and a response. It pre-populates the response with the results of any previous functions in the pipeline. The compose function should modify the response (rsp), for example to add composed resources.

The RunFunctionRequest and RunFunctionResponse types provided by this SDK are generated from a proto3 protocol buffer schema. Their fields behave similarly to built-in Python types like lists and dictionaries, but there are some differences. Read the generated code documentation to familiarize yourself with the the differences.

Your script has access to function-sdk-python as the crossplane.function module. For example you can import crossplane.function.resource. It also has access to the full Python standard library - use it with care.

apiVersion: apiextensions.crossplane.io/v1
kind: Composition
metadata:
  name: compose-a-resource-with-python
spec:
  compositeTypeRef:
    apiVersion: example.crossplane.io/v1
    kind: XR
  mode: Pipeline
  pipeline:
  - step: compose-a-resource-with-python
    functionRef:
      name: function-python
    input:
      apiVersion: python.fn.crossplane.io/v1beta1
      kind: Script
      script: |
        from crossplane.function.proto.v1 import run_function_pb2 as fnv1

        def compose(req: fnv1.RunFunctionRequest, rsp: fnv1.RunFunctionResponse):
            rsp.desired.resources["bucket"].resource.update({
                "apiVersion": "s3.aws.upbound.io/v1beta2",
                "kind": "Bucket",
                "spec": {
                    "forProvider": {
                        "region": req.observed.composite.resource["spec"]["region"]
                    }
                },
            })
            rsp.desired.resources["bucket"].ready = True

function-python is best for very simple cases. If writing Python inline of YAML becomes unwieldy, consider building a Python function using function-template-python.