xmlstruct.go (for https://www.youtube.com/watch?v=fGwEZmb_HuU)

xmlstruct.go

package main

/////////////////////////////////////////////////////////////////
//Code generated by chidley https://github.com/gnewton/chidley //
/////////////////////////////////////////////////////////////////

import (
	"bufio"
	"compress/bzip2"
	"compress/gzip"
	"encoding/json"
	"encoding/xml"
	"flag"
	"fmt"
	"io"
	"io/ioutil"
	"log"
	"os"
	"runtime"
	"strings"
)

const (
	JsonOut = iota
	XmlOut
	CountAll
)

var toJson bool = false
var toXml bool = false
var oneLevelDown bool = false
var countAll bool = false
var musage bool = false

var uniqueFlags = []*bool{
	&toJson,
	&toXml,
	&countAll}

var filename = ""

func init() {
	flag.BoolVar(&toJson, "j", toJson, "Convert to JSON")
	flag.BoolVar(&toXml, "x", toXml, "Convert to XML")
	flag.BoolVar(&countAll, "c", countAll, "Count each instance of XML tags")
	flag.BoolVar(&oneLevelDown, "s", oneLevelDown, "Stream XML by using XML elements one down from the root tag. Good for huge XML files (see http://blog.davidsingleton.org/parsing-huge-xml-files-with-go/")
	flag.BoolVar(&musage, "h", musage, "Usage")
	flag.StringVar(&filename, "f", filename, "XML file or URL to read in")
}

var out int = -1

var counters map[string]*int

func main() {
	runtime.GOMAXPROCS(runtime.NumCPU())
	flag.Parse()

	if musage {
		flag.Usage()
		return
	}

	numSetBools, outFlag := numberOfBoolsSet(uniqueFlags)
	if numSetBools == 0 {
		flag.Usage()
		return
	}

	if numSetBools != 1 {
		flag.Usage()
		log.Fatal("Only one of ", uniqueFlags, " can be set at once")
	}

	reader, xmlFile, err := genericReader(filename)
	if err != nil {
		log.Fatal(err)
		return
	}

	if true {
		xmlBytes, err := ioutil.ReadAll(reader)

		rss := Rss{}
		err = xml.Unmarshal(xmlBytes, &rss)
		if err != nil {
			fmt.Println("Can't Unmarshal", err)
		}

		//

		for _, i := range rss.Channel.Item {
			fmt.Println(i.Category)
			i.Category.Text = "FunNews"
		}

		//out, err := json.MarshalIndent(rss, "  ", "  ")
		out, err := xml.Marshal(rss)
		fmt.Println(string(out))
		//fmt.Printf("%s", rss.Channel.Item)

		return
	}
	decoder := xml.NewDecoder(reader)

	counters = make(map[string]*int)
	for {
		token, _ := decoder.Token()
		if token == nil {
			break
		}
		switch se := token.(type) {
		case xml.StartElement:
			handleFeed(se, decoder, outFlag)
		}
	}
	if xmlFile != nil {
		defer xmlFile.Close()
	}
	if countAll {
		for k, v := range counters {
			fmt.Println(*v, k)
		}
	}
}

func handleFeed(se xml.StartElement, decoder *xml.Decoder, outFlag *bool) {
	if outFlag == &countAll {
		incrementCounter(se.Name.Space, se.Name.Local)
	} else {
		if !oneLevelDown {
			if se.Name.Local == "rss" && se.Name.Space == "" {
				var item Rss
				decoder.DecodeElement(&item, &se)
				switch outFlag {
				case &toJson:
					writeJson(item)
				case &toXml:
					writeXml(item)
				}
			}
		} else {

			if se.Name.Local == "channel" && se.Name.Space == "" {
				var item Channel
				decoder.DecodeElement(&item, &se)
				switch outFlag {
				case &toJson:
					writeJson(item)
				case &toXml:
					writeXml(item)
				}
			}

		}
	}
}

func makeKey(space string, local string) string {
	if space == "" {
		space = "_"
	}
	return space + ":" + local
}

func incrementCounter(space string, local string) {
	key := makeKey(space, local)

	counter, ok := counters[key]
	if !ok {
		n := 1
		counters[key] = &n
	} else {
		newv := *counter + 1
		counters[key] = &newv
	}
}

func writeJson(item interface{}) {
	b, err := json.MarshalIndent(item, "", " ")
	if err != nil {
		log.Fatal(err)
	}
	fmt.Println(string(b))
}

func writeXml(item interface{}) {
	output, err := xml.MarshalIndent(item, "  ", "    ")
	if err != nil {
		fmt.Printf("error: %v\n", err)
	}
	os.Stdout.Write(output)
}

func genericReader(filename string) (io.Reader, *os.File, error) {
	if filename == "" {
		return bufio.NewReader(os.Stdin), nil, nil
	}
	file, err := os.Open(filename)
	if err != nil {
		return nil, nil, err
	}
	if strings.HasSuffix(filename, "bz2") {
		return bufio.NewReader(bzip2.NewReader(bufio.NewReader(file))), file, err
	}

	if strings.HasSuffix(filename, "gz") {
		reader, err := gzip.NewReader(bufio.NewReader(file))
		if err != nil {
			return nil, nil, err
		}
		return bufio.NewReader(reader), file, err
	}
	return bufio.NewReader(file), file, err
}

func numberOfBoolsSet(a []*bool) (int, *bool) {
	var setBool *bool
	counter := 0
	for i := 0; i < len(a); i++ {
		if *a[i] {
			counter += 1
			setBool = a[i]
		}
	}
	return counter, setBool
}

///////////////////////////
/// structs
///////////////////////////

type Category struct {
	Text string `xml:",chardata" json:",omitempty"`
}

type Channel struct {
	//Atom10_link          []*Atom10_link          `xml:"http://www.w3.org/2005/Atom link,omitempty" json:"link,omitempty"`
	Copyright            *Copyright              `xml:" copyright,omitempty" json:"copyright,omitempty"`
	Description          *Description            `xml:" description,omitempty" json:"description,omitempty"`
	Feedburner_feedFlare []*Feedburner_feedFlare `xml:"http://rssnamespace.org/feedburner/ext/1.0 feedFlare,omitempty" json:"feedFlare,omitempty"`
	Feedburner_info      *Feedburner_info        `xml:"http://rssnamespace.org/feedburner/ext/1.0 info,omitempty" json:"info,omitempty"`
	Image                *Image                  `xml:" image,omitempty" json:"image,omitempty"`
	Item                 []*Item                 `xml:" item,omitempty" json:"item,omitempty"`
	Language             *Language               `xml:" language,omitempty" json:"language,omitempty"`
	LastBuildDate        *LastBuildDate          `xml:" lastBuildDate,omitempty" json:"lastBuildDate,omitempty"`
	Link                 *Link                   `xml:" link,omitempty" json:"link,omitempty"`
	Title                *Title                  `xml:" title,omitempty" json:"title,omitempty"`
}

type Copyright struct {
	Text string `xml:",chardata" json:",omitempty"`
}

type Description struct {
	Text string `xml:",chardata" json:",omitempty"`
}

type Guid struct {
	Attr_isPermaLink string `xml:" isPermaLink,attr"  json:",omitempty"`
	Text             string `xml:",chardata" json:",omitempty"`
}

type Height struct {
	Text string `xml:",chardata" json:",omitempty"`
}

type Image struct {
	Height *Height `xml:" height,omitempty" json:"height,omitempty"`
	Link   *Link   `xml:" link,omitempty" json:"link,omitempty"`
	Title  *Title  `xml:" title,omitempty" json:"title,omitempty"`
	Url    *Url    `xml:" url,omitempty" json:"url,omitempty"`
	Width  *Width  `xml:" width,omitempty" json:"width,omitempty"`
}

type Item struct {
	Category            *Category            `xml:" category,omitempty" json:"category,omitempty"`
	Description         *Description         `xml:" description,omitempty" json:"description,omitempty"`
	Feedburner_origLink *Feedburner_origLink `xml:"http://rssnamespace.org/feedburner/ext/1.0 origLink,omitempty" json:"origLink,omitempty"`
	Guid                *Guid                `xml:" guid,omitempty" json:"guid,omitempty"`
	Link                *Link                `xml:" link,omitempty" json:"link,omitempty"`
	PubDate             *PubDate             `xml:" pubDate,omitempty" json:"pubDate,omitempty"`
	Title               *Title               `xml:" title,omitempty" json:"title,omitempty"`
}

type Language struct {
	Text string `xml:",chardata" json:",omitempty"`
}

type LastBuildDate struct {
	Text string `xml:",chardata" json:",omitempty"`
}

type Link struct {
	Text string `xml:",chardata" json:",omitempty"`
}

type PubDate struct {
	Text string `xml:",chardata" json:",omitempty"`
}

type Root struct {
	Rss *Rss `xml:" rss,omitempty" json:"rss,omitempty"`
}

type Rss struct {
	Attr_feedburner string   `xml:"xmlns feedburner,attr"  json:",omitempty"`
	Attr_version    string   `xml:" version,attr"  json:",omitempty"`
	Channel         *Channel `xml:" channel,omitempty" json:"channel,omitempty"`
}

type Title struct {
	Text string `xml:",chardata" json:",omitempty"`
}

type Url struct {
	Text string `xml:",chardata" json:",omitempty"`
}

type Width struct {
	Text string `xml:",chardata" json:",omitempty"`
}

type Feedburner_feedFlare struct {
	Attr_href string   `xml:" href,attr"  json:",omitempty"`
	Attr_src  string   `xml:" src,attr"  json:",omitempty"`
	Text      string   `xml:",chardata" json:",omitempty"`
	XMLName   xml.Name `xml:"http://rssnamespace.org/feedburner/ext/1.0 feedFlare,omitempty" json:"feedFlare,omitempty"`
}

type Feedburner_info struct {
	Attr_uri string   `xml:" uri,attr"  json:",omitempty"`
	XMLName  xml.Name `xml:"http://rssnamespace.org/feedburner/ext/1.0 info,omitempty" json:"info,omitempty"`
}

type Feedburner_origLink struct {
	Text    string   `xml:",chardata" json:",omitempty"`
	XMLName xml.Name `xml:"http://rssnamespace.org/feedburner/ext/1.0 origLink,omitempty" json:"origLink,omitempty"`
}

/*
type Atom10_link struct {
	Attr_atom10 string   `xml:"xmlns atom10,attr"  json:",omitempty"`
	Attr_href   string   `xml:" href,attr"  json:",omitempty"`
	Attr_rel    string   `xml:" rel,attr"  json:",omitempty"`
	Attr_type   string   `xml:" type,attr"  json:",omitempty"`
	XMLName     xml.Name `xml:"http://www.w3.org/2005/Atom link,omitempty" json:"link,omitempty"`
}
*/
///////////////////////////