@senchou/helm

Generate Helm templates with TypeScript.

Installation

# Using npm.
npm install @senchou/helm

# Using yarn.
yarn add @senchou/helm

Usage

The template function lets you create a template for any API Object or CustomResourceDefinition that you have generated with Senchou. See the following example for some of the different ways you can apply templating to resources.

import { template, Container, ContainerPort } from "@senchou/helm";
import { Pod } from "./senchou/k8s.ts";

const podTemplate = template(Pod, {
	metadata: {
		name: template.string(".Values.customName", {
			default: "default-name",
		}),
		labels: template.object<Record<string, string>>(".Values.customLabels"),
		annotations: template.object<Record<string, string>>(
			".Values.customAnnotations",
			{
				"my-permanent-annotation": "always-included",
			}
		),
	},
	spec: {
		containers: template.array<Container>(".Values.containers", [
			{
				name: "always-included",
				image: template.string(".Values.image", {
					default: "my-default-image:latest",
				}),
				ports: [
					{
						name: "https",
						containerPort: template.number(".Values.httpsPort"),
					},
					template.if<ContainerPort>({
						type: Object,
						condition: ".Values.enableGRPC",
						body: {
							name: "grpc",
							containerPort: 5000,
						},
						else: {
							name: "http",
							containerPort: 80,
						},
					}),
				],
			},
			template.if<Container>({
				type: Object,
				condition: ".Values.versionA",
				body: {
					name: "version-a",
					image: "version-a-image",
				},
				else: template.if<Container>({
					type: Object,
					condition: ".Values.versionB",
					body: {
						name: "version-b",
						image: "version-b-image",
					},
					else: {
						name: "version-c",
						image: "version-c-image",
					},
				}),
			}),
		]),
	},
});

After you've created your template, use the render function to generate a Helm YAML template.

import { render } from "@senchou/helm";

render(podTemplate);

apiVersion: v1
kind: Pod
metadata:
	annotations:
		my-permanent-annotation: always-included
		{{- with .Values.customAnnotations }}
		{{- toYaml . | nindent 4 }}
		{{- end }}
	labels:
		{{- with .Values.customLabels }}
		{{- toYaml . | nindent 4 }}
		{{- end }}
	name: {{ default "default-name" .Values.customName | quote }}
spec:
	containers:
		-
			name: always-included
			image: {{ default "my-default-image:latest" .Values.image | quote }}
			ports:
				-
					name: https
					containerPort: {{ .Values.httpsPort }}
				-
					{{- if .Values.enableGRPC }}
					name: grpc
					containerPort: 5000
					{{- else }}
					name: http
					containerPort: 80
					{{- end }}
		-
			{{- if .Values.versionA }}
			name: version-a
			image: version-a-image
			{{- else if .Values.versionB }}
			name: version-b
			image: version-b-image
			{{- else }}
			name: version-c
			image: version-c-image
			{{- end }}
		{{- toYaml .Values.containers | nindent 4 }}

API

template(factory, templateValue)

The template function takes in a factory function that produces an API object or CustomResourceDefinition. This factory must have been created using Senchou in order to work with the templating system.

Second, the template function takes in a templated value. This can be an object or array that you have used any of the various template.* helpers on.

template(Pod, {
	metadata: {
		name: template.string(".Values.name"),
	},
});

template.string(key, options?)

This helper returns a templated string. The key is the accessor used in the generated YAML. Typically, this will select a value from the user's supplied Helm values.

The options object allows you to customize the behavior of this helper. You can specify a default value with default or disable quoting of the resulting value by setting quote to false.

template(Pod, {
	metadata: {
		name: template.string(".Values.name", {
			default: "my-default-name",
		}),
		labels: {
			"some-label": template.string(".Values.some-label", {
				// This can be useful if you want to inject arbitrary
				// template expressions.
				quote: false,
			}),
		},
	},
});

template.number(key)

This helper returns a templated number. The key is a template expression to get the value of the number.

template(Pod, {
	spec: {
		containers: [
			{
				name: "my-container",
				image: "my-image",
				ports: [
					{
						containerPort: template.number(".Values.port"),
					},
				],
			},
		],
	},
});

template.boolean(key)

This helper returns a templated boolean. The key is a template expression to get the value of the boolean.

template(ConfigMap, {
	immutable: template.boolean(".Values.immutable"),
});

template.object<ObjectType>(key, value?: Partial<ObjectType>)

This helper returns a templated object. The key is a template expression to get the value of the object.

The value here can be highly dynamic, so it is typically required to specify the type of ObjectType as that will be the return type of the helper.

template(Pod, {
	metadata: {
		name: "my-pod",
		labels: template.object<Record<string, string>>(".Values.labels"),
		annotations: template.object<Record<string, string>>(
			".Values.annotations",
			{
				added: "always",
			}
		),
	},
});

template.array<ItemType>(key, items?: Array<ItemType>)

This helper returns a templated array. The key is a template expression to get the value of the array.

The items here can be highly dynamic, so it is typically required to specify the type of ItemType as that will also correct the return type of the helper.

template(Pod, {
	spec: {
		containers: template.array<Container>(".Values.containers", [
			{ name: "always-added-container", image: "some-image" },
		]),
	},
});

template.range<OutputType>(options)

This helper is a bit more unique than the others. Similar to template.if, it can output either an Array or and Object. In order to know which type to generate, you will have to supply it in the options.

template(Pod, {
	metadata: {
		name: "my-pod",
		labels: template.range<Record<string, string>>({
			type: Object,
			expression: "$key, $value := .Values.labels",
			body: {
				"{{ $key }}": template.string("$value"),
			},
		}),
	},
	spec: {
		containers: template.range<Array<Container>>({
			type: Array,
			expression: ".Values.containers",
			body: {
				name: template.string(".name"),
				image: template.string(".image"),
			},
		}),
	},
});

template.if<OutputType>(options)

This helper can let you dynamically output objects or arrays. However, you will need to specify the OutputType due to its dynamic nature.

template(Pod, {
	metadata: {
		name: "my-pod",
		labels: template.if<Record<string, string>>({
			type: Object,
			condition: ".Values.enableLabels",
			body: {
				enabled: "yes",
			},
			else: {
				enabled: "false",
			},
		}),
	},
	spec: {
		containers: [
			{
				name: "my-container",
				image: "my-image",
				ports: template.if<Array<ContainerPort>>({
					type: Array,
					condition: ".Values.enableAllPorts",
					body: [
						{ containerPort: 80 },
						{ containerPort: 443 },
						{ containerPort: 22 },
					],
					else: template.if<Array<Container>>({
						type: Array,
						condition: ".Values.enableHTTP",
						body: [{ containerPort: 80 }],
						else: [],
					}),
				}),
			},
		],
	},
});

template.include<OutputType>(options)

This helper will create an include directive in your template. This makes it easy to drop in existing values or fragments.

template(Pod, {
	spec: {
		affinity: template.include<Affinity>({
			type: Object,
			name: "my.affinity",
			context: ".",
		}),
	},
});

template.define(options)

This helper will create a define directive in your template. This can make it easy to reuse pieces of your template with include.

template(
	Pod,
	template.define({
		type: Object,
		name: "my.pod",
		body: {
			metadata: {
				name: template.string(".Values.name"),
			},
		},
	})
);