# sparkee	

Sparkee is a publish tool for `monorepo`,just like lerna but support `pnpm`, also support singleRepo.

## About

### What can Sparkee do?
There are six commands in sparkee:

`init` will initialize sparkee workspace.
`init-cliff` will initialize `git-cliff` config file.
`run` will run the script of package.
`info` will show package's info of workspace.
`log` will only generate `changelog.md`.
`publish` will publish any updated packages and generate changelog automatically.

### What diffrences with lerna?
* Sparkee can publish Single package or customize multiple packages.
* Support `workspace` protocol in `package.json`.
* Use `pnpm <--filter> publish` so you can keep the `workspace` keyword of `dependencies`.
* More friendly output of CLI.
* CLI commands are very very simple!

## Getting Started

### Project structure

Sparkee needs a file structure like this(the same as the lerna project):

```
your-repo/
  package.json
  packages/
    package-1/
      src/
      package.json
    package-2/
      src/
      package.json
```

It will track file changes inside the packages `src` folders and `package.json` files.

### Installation

```
# simple install or install it globally with -g
npm install sparkee --save-dev

# if you install globally, use it `sparkee <command>`
sparkee init
sparkee version
sparkee publish

# OR use npx in root of project
npx sparkee init
npx sparkee publish
npx sparkee version
```

### Usage

#### init

```sh
$ sparkee init
```

Initialize `sparkee workspace`. There will create a `spark.json`, you can make sparkee manage all of packages or customized packages.

> Sparkee assumes the repo has already been initialized with `git init`.

Example output on a new git repo:

Example-1. All packages:

```sh
$ sparkee init
? Do you need sparkee to manage all projects of packages folder? yes
✔ Sparkee init successful.
```

Example-2. Custom packages:

```sh
$ sparkee init
? Do you need sparkee to manage all projects of packages folder? No
? What packages do you want to manage? @geocld/pkg1
✔ Sparkee init successful.
``` 

Example-3. single repo:

```sh
$ sparkee init
? Please select repo type: singleRepo
✔ Sparkee init as singleRepo successful.
```
In singleRepo mode, `spark.json` will be:

```json
{
  "singleRepo": true,
  "moduleManager": "npm"
}
```

`singleRepo` : Must be `true` in single repo.
`moduleManager`: Can be `npm` or `yarn` or `pnpm`.

#### run(supported in v1.1.0)

```sh
$ sparkee run
```

Run the `scripts` of `packages.json`:

Example:

```bash
$ sparkee run
? Please select script: (Use arrow keys)
❯ dev -> rimraf dist && tsc -w
  build -> rimraf dist && tsc
```

#### info

```sh
$ sparkee info <--tree>
```

Print local information of packages.

Example:

```bash
$ sparkee info
> Current monorepo packages:
  · @geocld/pkg1: v1.0.0
  · @geocld/pkg2: v1.0.0
```

Or Print detail dependencies tree of packages:

```bash
$ sparkee info --tree

@geocld/pkg1@1.0.0
╰── @geocld/pkg2@1.0.0

----------

@geocld/pkg2@1.0.0
``` 

#### log(supported in v1.3.0)

```sh
$ sparkee log
```
If you only want to generate a changelog, you can execute the above command.

#### publish

```bash
$ sparkee publish
```

`publish` is the core of sparkee. When runing, this command does the following things:

- Find updated packages since the last release(must with `git tag`, otherwise compare with first commit).
- Select what packages that you want to publish.
- Update `version` of `package.json` automatically.
- Generate `CHANGELOG.md` automatically in selected packages.
- Run `git Commit`, `git tag` and `git push` automatically.
- Run `pnpm publish`

> sparkee will never publish packages which do not exits in `spark.json`.

The workflow of `sparkee` is as follows:

![workflow](http://geocld.github.io/img/sparkee/workflow.png)

#### publish --force
Sparkee will not publish unmodified packages, if you want to publish unmodified packages, use `--force`:

```bash
sparkee publish --force
```

#### publish --noPublish
If yout want `sparkee` just genetate changelog and create tag, `publish` command run in other place(such as `pipeline`), use `--noPublish` or `--np`:

```bash
sparkee publish --noPublish

or

sparkee publish --np
```

#### publish --noCommitCheck
If you want to skip commit check before publish or generate changelog, you can use `--noCommitCheck` or `--ncm`:

```bash
sparkee publish --noCommitCheck

or

sparkee publish --ncm
```

#### publish --package <packageName> --ver <packageVersion>
If you dont want to use prompt when publish(such as in CI/CD autodeploy case), you can use `--package` and `--ver` to specify the name and version of package.For example:

```bash
sparkee publish --package @geocld/sparkee --ver 1.2.0
```

#### Custom changelLog

`sparkee` use [git-cliff](https://github.com/orhun/git-cliff) internally to generate `CHANGELOG.md`, If you want to use custom changelog, you can run `sparkee init-cliff` to generate a `cliff` config file, refer to the [configuration](https://git-cliff.org/docs/configuration/) items in detail.


#### JSON $schema

Since version `1.3.0`, a JSON schema file for the `spark.json` is published.

You can specify a relative path to the schema of the Sparkee npm package if Sparkee is installed in the `node_modules` folder:

```json
{
  "$schema": "./node_modules/sparkee/template/schema.json",
  "moduleManager": "pnpm",
  "packages": [
    "@example/docs",
    "@example/cli",
  ]
}
```

## License

MIT © Geocld