Last active
April 9, 2021 04:01
-
-
Save ckhung/7424304627692444febe to your computer and use it in GitHub Desktop.
extract a piece of a web page using css selector
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| #!/usr/bin/php | |
| <?php | |
| require_once 'QueryPath/qp.php'; | |
| // 若是使用舊版 (2.1.2) 的 QueryPath， 應該這樣寫： | |
| // require_once 'QueryPath/QueryPath.php'; | |
| // 詳見 「網頁搜括小工具： 用 extract.php 擷取網頁當中的一小塊」 | |
| // http://newtoypia.blogspot.tw/2016/03/web-scraping.html | |
| // extract.php -s 'div.hentry' < ckhung.html > ckhung2.html | |
| // extract.php -p blogger < ckhung.html > ckhung3.html | |
| // For an incomplete input (no html header) with utf8 contents, | |
| // use -u to wrap the input html so that utf8 characters | |
| // use -t to remove all html tags and retain only text | |
| // can be correctly processed. | |
| // use -c div,table to clothe div (outer) and table (inner) around the extracted piece | |
| // use -w to further wrap html header and body around the wrapped piece | |
| // so that it becomes a complete html. | |
| // convert a predefined name to its search string | |
| $n2ss = array( | |
| 'eff' => '#main', | |
| 'yro' => '.body', | |
| 'blogger' => 'div.hentry', | |
| ); | |
| $options = getopt("p:s:tc:uw"); | |
| # https://github.com/technosophos/querypath/issues/94 | |
| $qp_options = array( | |
| 'convert_from_encoding' => 'UTF-8', | |
| 'convert_to_encoding' => 'UTF-8', | |
| 'strip_low_ascii' => FALSE, | |
| ); | |
| if (array_key_exists("p", $options)) { | |
| if (array_key_exists("s", $options)) | |
| exit("please use either -p <predefined> or -s <selector> (but not both)\n"); | |
| if (! array_key_exists($options["p"], $n2ss)) | |
| exit("'$options[p]' is not a predefined name\n"); | |
| $selector = $n2ss[$options["p"]]; | |
| } else { | |
| if (array_key_exists("s", $options)) | |
| $selector = $options["s"]; | |
| else | |
| $selector = ""; | |
| } | |
| $header = <<<EOT | |
| <html> | |
| <head> | |
| <meta http-equiv='Content-Type' content='text/html; charset=UTF-8' /> | |
| </head> | |
| <body> | |
| EOT; | |
| $footer = <<<EOT | |
| </body> | |
| </html> | |
| EOT; | |
| $in = file_get_contents('php://stdin'); | |
| if (array_key_exists("u", $options)) { | |
| $in = "$header$in$footer"; | |
| } | |
| $qp = htmlqp($in, NULL, $qp_options); | |
| if (strlen($selector) > 0) { | |
| $out = ""; | |
| foreach ($qp->find($selector) as $found) { | |
| $out .= array_key_exists("t", $options) ? | |
| $found->text() : $found->html(); | |
| $out .= "\n"; | |
| } | |
| } else { | |
| $out = array_key_exists("t", $options) ? | |
| $qp->text() : $qp->html(); | |
| } | |
| if (array_key_exists("c", $options)) { | |
| $clothes = array_reverse(explode(",", $options["c"])); | |
| foreach ($clothes as $layer) { | |
| $out = "<$layer>\n$out\n</$layer>"; | |
| } | |
| } | |
| if (array_key_exists("w", $options)) | |
| $out = "$header$out$footer"; | |
| echo $out; | |
| ?> | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment