/* See LICENSE file for copyright and license details. */

package unxml

import (
	"fmt"
	"io"
	"os"

	"golang.org/x/net/html"
)

type Reader struct {
	reader    io.Reader
	tagmap    map[string]bool
	lastread  []byte
	count     int
	tokenizer *html.Tokenizer
}

func NewReaderKeepTags(r io.Reader, tagstokeep []string) *Reader {
	var tagmap map[string]bool

	if len(tagstokeep) > 0 {
		tagmap = make(map[string]bool, 10)
		for _, tag := range tagstokeep {
			tagmap[tag] = true
		}
	} else {
		tagmap = nil
	}

	return &Reader{reader: r,
		tagmap:    tagmap,
		tokenizer: html.NewTokenizer(r),
	}
}

func (r *Reader) Read(out []byte) (int, error) {
	fmt.Fprintf(os.Stderr, "Read has been called.\n")
	var err error

	r.count = 0
	n := 0
	lenout := len(out)
	if lenout == 0 {
		return r.count, nil
	}

	lenlr := len(r.lastread)
	if lenlr > 0 {
		n = copy(out[0:], r.lastread)
		r.count += n
		r.lastread = make([]byte, len(out))
		lenout -= n
	}

	for {
		tt := r.tokenizer.Next()

		switch tt {
		case html.ErrorToken:
			fmt.Fprintf(os.Stderr, "There was an error when parsing the html: %s, %s\n", tt, r.tokenizer.Err())
			return r.count, io.EOF

		case html.TextToken:
			text := r.tokenizer.Text()
			lentext := len(text)
			if lentext <= lenout {
				n = copy(out[r.count:], text)
				r.count += n
				lenout -= n
				//fmt.Printf("HAD SPACE: %q, count: %d, err: %s\n", text, r.count, err)
			} else {
				n = copy(out[r.count:], text[:lenout-1])
				r.count += n
				r.lastread = text[lenout-1:]
				//fmt.Printf("HAD NO SPACE: count: %d, err: %s\n", r.count, err)
				return r.count, err
			}

		case html.StartTagToken, html.EndTagToken:
			_, _ = r.tokenizer.TagName()
			//fmt.Printf("TagName: %s\n", tn)
		}
	}
}