Embrace the Pipe

The unix pipe operator is magical.

As long as you write your application to read input from standard-in and write output to standard-out you can compose your cli tool with thousands of other tools.

Scanners

To read one line at a time I like to use bufio.NewScanner. This uses bufio.ScanLines to split the reader by new lines.

in := bufio.NewScanner(os.Stdin)
// Scan returns false when its out of lines
for in.Scan() {
  // line is a string
  line := in.Text()
  // if you want bytes in.Bytes() exists too
}

Lets build a tool

As an example here is a tiny command line tool I wrote to read urls for images and display them as ascii art images on the terminal.

Full source https://gist.github.com/trashhalo/6154a690bd2612ae33c8252d8051eb9f

func main() {
    in := bufio.NewScanner(os.Stdin)
    for in.Scan() {
        line := in.Text()
        err := convertLineToArt(line)
        if err != nil
    }
}

func convertLineToArt(line string) error {
    fmt.Println(line)
    resp, err := http.Get(line)
    if err != nil {
    defer resp.Body.Close()

    img, _, err := image.Decode(resp.Body)

    convertOptions := convert.DefaultOptions
    convertOptions.FixedWidth = 50
    convertOptions.FixedHeight = 50
    converter := convert.NewImageConverter()
    for _, row := range converter.Image2ASCIIMatrix(img, &convertOptions) {
        fmt.Print(row)
    }
    return nil
}

Use Pipe

Now lets combine together a few tools.

curl -A "test" https://www.reddit.com/r/cats.json | jq -r '.data.children[] | .data.url'|go run .|less

This is saying:

1. curl - download the json for reddit posts on cats. The -A "test" because reddit hates curl user agent so we need to change it.
2. jq -r 'xyz' - extract the urls for the images from the reddit json.
3. go run - run your command!
4. less - Put the output in a little viewer that lets you scroll up and down.