使用Cobra创建功能强大的命令行CLI

我敢肯定所有人都在使用git，也有很大一部分在使用kuberntes，但是不确定你是否会对每天都在操作使用的这俩命令行CLI感到好奇? 是否想过对它们的实现一探究竟？

毕竟,好奇心是驱使人类进步的一大动力

哈哈，直接进入今天的主题；Cobra，一个可以用来创建强大功能命令行CLI的工具，git / kubectl 都是它的代表作。

就不翻译 文档 了，先说下基本使用方式，然后直接上例子。

Golang 配置和引用 Cobra

go get -u github.com/spf13/cobra/cobra import "github.com/spf13/cobra"

cobra init 创建project

cobra init [app] # this will create your initial application code for you

cobra add 创建cmd

cobra add [command]

基本了解上面这3步就可以上手实验了，下面以一个cmd为例子：

cobra add rexcli : 创建一个叫 rexcli 的 golang project，之后你的命令也就叫 rexcli 了cobra add deploy : 对rexcli 创建一个 sub-command，会直接生成一个 deploy.go 文件，对 deploy.go 添肉 package cmd import ( "fmt" "github.com/spf13/cobra" ) type Deployment struct { modelFramework string modelUrl string apiName string cpu int memory int replicaSet int } var deployment = Deployment{} var deployCmd = &cobra.Command{ Use: "deploy", Short: "Deploy your model in as a model-api (Restful API)", Long: `Deploy your model in as a model-api (Restful API)`, Run: func(cmd *cobra.Command, args []string) { fmt.Println("deploying your model ", deployment.apiName, "in", env, ", from:", deployment.modelUrl, ", model framework: ", deployment.modelFramework, "", ", cpu:", deployment.cpu, ", memory:", deployment.memory, "Gi, how many instances:", deployment.replicaSet, ) fmt.Println("....") }, } func init() { modelCmd.AddCommand(deployCmd) deployCmd.Flags().StringVar(&deployment.modelFramework, "model-framework", "MLflow", "which ML framework you used fro training this model, e.g. MLflow/Tensorflow/XGBoost/SKLearn") deployCmd.Flags().StringVar(&deployment.modelUrl, "model-url", "", "where your model is, a URL") deployCmd.Flags().StringVar(&deployment.apiName, "api-name", "", "your model-api name") deployCmd.Flags().IntVar(&deployment.cpu, "cpu", 1, "how many CPUs") deployCmd.Flags().IntVar(&deployment.memory, "memory", 1, "how many Gi memory for 1 model-api") deployCmd.Flags().IntVar(&deployment.replicaSet, "replica-set", 1, "how many model-api instances for 1 model-api") }

之后， go install build后，就可以直接run了!

怎么样,是不是很神奇?!

更多以及更详细的实例，请移步 我的github，如果能一键三连就更好了 😄