|
@@ -0,0 +1,61 @@
|
|
1
|
+package main
|
|
2
|
+
|
|
3
|
+import (
|
|
4
|
+ "fmt"
|
|
5
|
+ "log"
|
|
6
|
+ "os"
|
|
7
|
+ "strings"
|
|
8
|
+
|
|
9
|
+ "golang.org/x/net/html"
|
|
10
|
+)
|
|
11
|
+
|
|
12
|
+func main() {
|
|
13
|
+ fi, err := os.Open(os.Args[1])
|
|
14
|
+ if err != nil {
|
|
15
|
+ log.Fatal(err)
|
|
16
|
+ }
|
|
17
|
+
|
|
18
|
+ doc, err := html.Parse(fi)
|
|
19
|
+ if err != nil {
|
|
20
|
+ log.Fatal(err)
|
|
21
|
+ }
|
|
22
|
+
|
|
23
|
+ content := findContent(doc)
|
|
24
|
+
|
|
25
|
+ fmt.Println(extractText(content))
|
|
26
|
+}
|
|
27
|
+
|
|
28
|
+func findContent(n *html.Node) *html.Node {
|
|
29
|
+ for _, attr := range n.Attr {
|
|
30
|
+ if attr.Key == "id" && attr.Val == "content" {
|
|
31
|
+ return n
|
|
32
|
+ }
|
|
33
|
+ }
|
|
34
|
+ for c := n.FirstChild; c != nil; c = c.NextSibling {
|
|
35
|
+ if found := findContent(c); found != nil {
|
|
36
|
+ return found
|
|
37
|
+ }
|
|
38
|
+ }
|
|
39
|
+
|
|
40
|
+ return nil
|
|
41
|
+}
|
|
42
|
+
|
|
43
|
+func extractText(n *html.Node) string {
|
|
44
|
+ text := ""
|
|
45
|
+
|
|
46
|
+ if n.Type == html.TextNode {
|
|
47
|
+ data := strings.TrimSpace(n.Data)
|
|
48
|
+ if data != "" {
|
|
49
|
+ text += data + "\n"
|
|
50
|
+ }
|
|
51
|
+ }
|
|
52
|
+
|
|
53
|
+ for c := n.FirstChild; c != nil; c = c.NextSibling {
|
|
54
|
+ extr := strings.TrimSpace(extractText(c))
|
|
55
|
+ if extr != "" {
|
|
56
|
+ text += extr + "\n"
|
|
57
|
+ }
|
|
58
|
+ }
|
|
59
|
+
|
|
60
|
+ return text
|
|
61
|
+}
|