In this article, an overview of MRI package is provided with a usage example picked from Changesets source code.
MRI package:
You can use MRI package to quickly scan for flags and arguments. It is an alternative to yargs-parser.
Install
npm install - save mri
Usage
// Example CLI with options
$ demo-cli - foo - bar=baz -mtv - hello world
The following code is picked from MRI npm package documentation.
const mri = require('mri');
const argv = process.argv.slice(2);
mri(argv);
//=> { _: ['hello', 'world'], foo:true, bar:'baz', m:true, t:true, v:true }
mri(argv, { boolean:['bar'] });
//=> { _: ['baz', 'hello', 'world'], foo:true, bar:true, m:true, t:true, v:true }
mri(argv, {
alias: {
b: 'bar',
foo: ['f', 'fuz']
}
});
//=> { _: ['hello', 'world'], foo:true, f:true, fuz:true, b:'baz', bar:'baz', m:true, t:true, v:true }
Read more about the options:
In essence, we get the CLI args converted into an object. Now that we understand the basics of MRI, it is time we look at its usage in Changesets.
MRI usage in Changesets:
MRI is found to be imported in Changesets CLI package
When you add a changeset using the command npx changeset add or npx changeset, these are accessed in the CLI package as shown below.
const args = process.argv.slice(2);
Parsing CLI args in changeset
The below code shows how mri used in Changeset CLI package
const parsed = mri(args, {
boolean: ["sinceMaster", "verbose", "empty", "open", "gitTag", "snapshot"],
string: [
"output",
"otp",
"since",
"ignore",
"tag",
"snapshot",
"snapshotPrereleaseTemplate",
],
alias: {
// Short flags
v: "verbose",
o: "output",
// Support kebab-case flags
"since-master": "sinceMaster",
"git-tag": "gitTag",
"snapshot-prerelease-template": "snapshotPrereleaseTemplate",
// Deprecated flags
"update-changelog": "updateChangelog",
"is-public": "isPublic",
"skip-c-i": "skipCI",
},
default: {
gitTag: true,
},
});
parsed value will look like below, I deduced this based on the documentation:
{
// string value (if you have used 'add' in npx changeset add)
['add'],
// boolean values
"sinceMaster": true,
"verbose": true,
"empty": true,
"open": true,
"gitTag": true,
"snapshot": true
// string values
// Note: if you have passed these options in your CLI, these keys will be parsed as string, no matter the what you pass in
// example: if you pass in - otp=123, 123 here, even though is a number, gets parsed as string since otp is configured to be parsed as
// string in the above code
"output",
"otp",
"since",
"ignore",
"tag",
"snapshot",
"snapshotPrereleaseTemplate",
// The alias option in mri allows you to define alternative names (aliases) for CLI arguments.
// This is particularly useful for supporting:
// Short flags: Such as -v for - verbose.
// Kebab-case flags: Allowing flags like - since-master to map to camelCase variables in JavaScript (e.g., sinceMaster).
// Deprecated flags: If you want to support older names for backward compatibility but still map them to the current property names.
}
parsed variable is used in a function named run imported from /run.ts
// run function call
run(parsed._, parsed, cwd).catch((err)
The first parameter is pared._ because in the documentation, it is provided that parsed commands like ‘add’ looks like { _: [‘add’]}
// run function definition
export async function run(
input: string[],
flags: { [name: string]: any },
cwd: string
) {
parsed contains mri parsed object based on the CLI args and the configuration set for booleans, strings, defaults, aliases.
cwd is a current working directory, you can get this with process.cwd()
About us:
At Thinkthroo, we study large open source projects and provide architectural guides. We have developed reusable Components, built with tailwind, that you can use in your project. We offer Next.js, React and Node development services.
Book a meeting with us to discuss your project.
Top comments (0)