LoadURL loads the HTML document from the specified URL. Default enabling gzip on a HTTP request.
(url string)
| 92 | |
| 93 | // LoadURL loads the HTML document from the specified URL. Default enabling gzip on a HTTP request. |
| 94 | func LoadURL(url string) (*html.Node, error) { |
| 95 | req, err := http.NewRequest("GET", url, nil) |
| 96 | if err != nil { |
| 97 | return nil, err |
| 98 | } |
| 99 | // Enable gzip compression. |
| 100 | req.Header.Add("Accept-Encoding", "gzip") |
| 101 | resp, err := http.DefaultClient.Do(req) |
| 102 | if err != nil { |
| 103 | return nil, err |
| 104 | } |
| 105 | var reader io.ReadCloser |
| 106 | |
| 107 | defer func() { |
| 108 | if reader != nil { |
| 109 | reader.Close() |
| 110 | } |
| 111 | }() |
| 112 | encoding := resp.Header.Get("Content-Encoding") |
| 113 | switch encoding { |
| 114 | case "gzip": |
| 115 | gzipReader, err := gzip.NewReader(resp.Body) |
| 116 | if err != nil { |
| 117 | return nil, err |
| 118 | } |
| 119 | reader = gzipReader |
| 120 | case "deflate": |
| 121 | zlibReader, err := zlib.NewReader(resp.Body) |
| 122 | if err != nil { |
| 123 | return nil, err |
| 124 | } |
| 125 | reader = zlibReader |
| 126 | case "": |
| 127 | reader = resp.Body |
| 128 | default: |
| 129 | return nil, fmt.Errorf("%s compression is not support", encoding) |
| 130 | } |
| 131 | |
| 132 | r, err := charset.NewReader(reader, resp.Header.Get("Content-Type")) |
| 133 | if err != nil { |
| 134 | return nil, err |
| 135 | } |
| 136 | return html.Parse(r) |
| 137 | } |
| 138 | |
| 139 | // LoadDoc loads the HTML document from the specified file path. |
| 140 | func LoadDoc(path string) (*html.Node, error) { |
no outgoing calls
searching dependent graphs…