nx/docs/shared/recipes/running-tasks/root-level-scripts.md

116 lines
4.0 KiB
Markdown

# Run Root-Level NPM Scripts with Nx
{% youtube
src="https://www.youtube.com/embed/PRURABLaS8s"
title="Run root-level NPM scripts with Nx"
/%}
There are often tasks in a codebase that apply to the whole codebase rather than a single project. Starting with version 15.3.0 of Nx, you can run npm scripts directly from the root `package.json`.
Let's say your root `package.json` looks like this:
```json {% fileName="package.json" %}
{
"name": "myorg",
"scripts": {
"docs": "node ./generateDocsSite.js"
}
}
```
We want to be able to run the `docs` script using Nx to get caching and other benefits.
## Setup
To make Nx aware of the root `package.json` scripts, add an `"nx": {}` property to the root `package.json`
```json {% fileName="package.json" %}
{
"name": "myorg",
"nx": {},
"scripts": {
"docs": "node ./generateDocsSite.js"
}
}
```
## Running a Root-Level Target
Once Nx is aware of your root-level scripts, you can run them the same way you would run any other target. Just use the name of your root `package.json` as the project name, or you can omit the project name and Nx will use the project in the current working directory as the default.
For our example, you would run:
```{% command="nx docs" path="~/myorg" %}
> nx run myorg:docs
yarn run v1.22.19
$ node ./generateDocsSite.js
Documentation site generated in /docs
————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————
> NX Successfully ran target docs for project myorg (5s)
```
## Configuring a Root-Level Target
You can also configure the `inputs` and `outputs` or task pipelines for root-level targets the same way you would for any other target.
Our fully configured example would look like this:
```jsonc {% fileName="package.json" %}
{
"name": "myorg",
"nx": {
// Nx can't infer the project dependency from the docs script,
// so we manually create a dependency on the store app
"implicitDependencies": ["store"],
"targets": {
"docs": {
// generates docs from source code of all dependencies
"inputs": ["^production"],
// the docs site is created under /docs
"outputs": ["{workspaceRoot}/docs"]
}
}
},
"scripts": {
"docs": "node ./generateDocsSite.js"
}
}
```
To cache the `docs` target, you can add `docs` to the `cacheableOperations` in `nx.json` and then your output would look like this:
```{% command="nx docs" path="~/myorg" %}
> nx run myorg:docs [existing outputs match the cache, left as is]
yarn run v1.22.19
$ node ./generateDocsSite.js
Documentation site generated in /docs
————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————
> NX Successfully ran target docs for project myorg (31ms)
Nx read the output from the cache instead of running the command for 1 out of 1 tasks.
```
Read more about [cacheableOperations](/core-features/cache-task-results) and fine-tuning caching with [task inputs](/recipes/running-tasks/customizing-inputs).
## Keep using NPM to run scripts rather than Nx
You can keep using `npm run docs` instead of the new `npx nx docs` version and still leverage the caching. To achieve this you need to wrap your command with `nx exec` s.t. it can be piped through Nx.
```json {% fileName="package.json" %}
{
"name": "myorg",
"nx": {},
"scripts": {
"docs": "nx exec -- node ./generateDocsSite.js"
}
}
```
Read more in the [Nx exec docs](/nx-api/nx/documents/exec).