Strip HTML tags in Golang

strip_html_tags.go

import (
	"html"
	"reflect"
	"strings"

	"github.com/microcosm-cc/bluemonday"
)

var p = bluemonday.UGCPolicy()

// StripHTMLTags strips the string field in `input`. The string field can be nested in other
// data types, such as a map or a slice. The method will search through all the fields.
func StripHTMLTags(input interface{}) {
	values := reflect.ValueOf(input)
	stripRecursive(values)
}

func stripRecursive(field reflect.Value) {
	switch field.Kind() {

	// String type, finally at the place we want to be, sanitize
	case reflect.String:
		if field.CanSet() { // Can't set for unexported type, etc.
			sanitized := sanitize(sanitize(field.Interface().(string)))
			field.SetString(sanitized)
		}

	// Pointer type, we check if it's nil. Dereference it and call recursion.
	// Interface type is similar.
	case reflect.Ptr,
		reflect.Interface:
		elem := field.Elem()
		if !elem.IsValid() { // nil pointer stop processing
			return
		}
		stripRecursive(elem)

	// For slice, loop each element and call recursion.
	case reflect.Slice:
		for j := 0; j < field.Len(); j++ {
			sj := field.Index(j)
			stripRecursive(sj)
		}

	// For struct, loop each field and call recursion.
	case reflect.Struct:
		for j := 0; j < field.NumField(); j++ {
			ff := field.Field(j)
			stripRecursive(ff)
		}

	// For map, loop each key and call recursion. Note that we have to construct a new interface
	// for a value in the map, or else the value cannot be able to set.
	case reflect.Map:
		for _, key := range field.MapKeys() {
			val := field.MapIndex(key)
			cloneVal := reflect.New(val.Type()).Elem()
			cloneVal.Set(val)
			stripRecursive(cloneVal)
			field.SetMapIndex(key, cloneVal)
		}
	}
}

func sanitize(input string) string {
	sanitized := p.Sanitize(input)
	sanitized = html.UnescapeString(sanitized)
	sanitized = strings.TrimSpace(sanitized)
	sanitized = strings.ReplaceAll(sanitized, "`", "")
	return sanitized
}