This change adds support for running substation fmt against jsonnet and libsonnet files, automatically normalizing the format of substation configurations.
Motivation and Context
This change was added to continue making substation more approachable and easy to work with by enhancing the CLI that is used to manage and develop configurations. Right now the CLI supports building and testing. The addition of substation fmt adds a very small piece of that puzzle which is ensuring that all configuration is formatted the same way - this improves quality of life for teams managing many configs that are written by multiple developers.
How Has This Been Tested?
The CLI has been locally built and tested on example jsonnet and libsonnet files to verify functionality.
When running substation fmt, the output styling mimics that of go fmt, for consistency with the stylistic choices in substation test.
Write formatted changes to stdout.
> substation fmt examples/transform/aggregate/sample/config.jsonnet
// This example samples data by aggregating events into an array, then
// selecting the first event in the array as a sample. The sampling rate
// is 1/N, where N is the count of events in the buffer.
local sub = import '../../../../substation.libsonnet';
{
tests: [
{
name: 'sample',
transforms: [
sub.tf.test.message({ value: { a: 'b' } }),
sub.tf.test.message({ value: { c: 'd' } }),
sub.tf.test.message({ value: { e: 'f' } }),
sub.tf.test.message({ value: { g: 'h' } }),
sub.tf.test.message({ value: { i: 'j' } }),
sub.tf.test.message({ value: { k: 'l' } }),
sub.tf.test.message({ value: { m: 'n' } }),
sub.tf.test.message({ value: { o: 'p' } }),
sub.tf.test.message({ value: { q: 'r' } }),
sub.tf.test.message({ value: { s: 't' } }),
sub.tf.test.message({ value: { u: 'v' } }),
sub.tf.test.message({ value: { w: 'x' } }),
sub.tf.test.message({ value: { y: 'z' } }),
sub.tf.test.message({ value: ' ' }),
sub.tf.send.stdout(),
],
// Asserts that the message is '{"c":"d"}'.
condition: sub.cnd.num.len.greater_than({ value: 0 }),
},
],
transforms: [
// Events are aggregated into an array. This example has a sample
// rate of up to 1/5. By default, the sample rate will be lower if
// fewer than 5 events are processed by Substation.
sub.tf.aggregate.to.array({ object: { target_key: 'meta sample' }, batch: { count: 5 } }),
// A strict sample rate can be enforced by dropping any events that
// contain the `sample` key, but do not have a length of 5.
sub.tf.meta.switch(settings={ cases: [
{
condition: sub.cnd.num.len.eq({ object: { source_key: 'meta sample' }, value: 5 }),
transforms: [
sub.tf.object.copy({ object: { source_key: 'meta sample.0' } }),
],
},
{
condition: sub.cnd.num.len.gt({ object: { source_key: 'meta sample' }, value: 0 }),
transforms: [
sub.tf.util.drop(),
],
},
] }),
sub.tf.obj.cp({ object: { source_key: 'meta sample.0' } }),
sub.tf.send.stdout(),
],
}
Description
This change adds support for running
substation fmt
against jsonnet and libsonnet files, automatically normalizing the format of substation configurations.Motivation and Context
This change was added to continue making substation more approachable and easy to work with by enhancing the CLI that is used to manage and develop configurations. Right now the CLI supports building and testing. The addition of
substation fmt
adds a very small piece of that puzzle which is ensuring that all configuration is formatted the same way - this improves quality of life for teams managing many configs that are written by multiple developers.How Has This Been Tested?
The CLI has been locally built and tested on example jsonnet and libsonnet files to verify functionality.
When running
substation fmt
, the output styling mimics that ofgo fmt
, for consistency with the stylistic choices insubstation test
.Write formatted changes to stdout.
Write formatted changes to the file.
Types of changes
Checklist: