package spider

import (
	"fmt"
	"io"
	"net/http"
	"strings"

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

// URLs returns all links on a page
func URLs(page io.Reader) ([]string, error) {
	doc, err := html.Parse(page)
	if err != nil {
		return nil, fmt.Errorf("parsing html: %v", err)
	}
	paths := []string{}
	var f func(*html.Node)
	f = func(n *html.Node) {
		if n.Type == html.ElementNode {
			for _, a := range n.Attr {
				switch a.Key {
				case "href", "src":
					paths = append(paths, a.Val)
					break
				}
			}
		}
		for c := n.FirstChild; c != nil; c = c.NextSibling {
			f(c)
		}
	}
	f(doc)
	return paths, nil
}

type Link struct {
	From string
	To   string

	Err error
}

func (l Link) String() string {
	r := fmt.Sprintf("%s > %s", l.From, l.To)
	if l.Err != nil {
		r = fmt.Sprintf("%v (%v)", r, l.Err)
	}
	return r
}

// Pages returns a stream of full urls starting at a given base page.
func Pages(base string) <-chan Link {
	r := make(chan Link)

	base = strings.TrimRight(base, "/")
	visited := map[string]bool{base: true}
	links := []Link{}

	resp, err := http.Get(base)
	if err != nil {
		go func() {
			r <- Link{To: base, From: "start", Err: err}
			close(r)
		}()
		return r
	}
	lks, err := URLs(resp.Body)
	if err != nil {
		go func() {
			r <- Link{To: base, From: "start", Err: err}
			close(r)
		}()
		return r
	}

	for _, l := range lks {
		links = append(links, Link{From: base, To: l})
	}

	go func() {
		for len(links) > 0 {
			l := links[0]
			links = links[1:]
			if strings.HasPrefix(l.To, "/") {
				l.To = base + l.To
			}
			if !strings.HasPrefix(l.To, base) {
				continue
			}

			// drop fragments
			h := strings.Index(l.To, "#")
			if h >= 0 {
				l.To = l.To[:h]
			}

			if _, ok := visited[l.To]; !ok {
				r <- l

				resp, err := http.Get(l.To)
				if err != nil {
					panic(err)
				}
				lks, err := URLs(resp.Body)
				if err != nil {
					panic(err)
				}

				for _, lk := range lks {
					links = append(links, Link{From: l.From, To: lk})
				}
			}
			visited[l.To] = true
		}
		close(r)
	}()
	return r
}