Mercurial > hg > extraction-interface

annotate interface/review_index_xml.php @ 6:63e08b98032f

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
rewrite extraction interface into PHP MVC architecture. (Although js hasn't been rewritten into MVC, it's fitted into the current PHP MVC architecture.) - The root of the new PHP MVC is at 'develop/'. - extraction interface is called "Extractapp" with several action, eg TaggingText, EditWordlist, EditTaglist, ExportTable.
author Zoe Hong <zhong@mpiwg-berlin.mpg.de>
date Thu, 05 Feb 2015 16:07:53 +0100
parents b12c99b7c3f0
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
1 <?php
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
2 header("Content-Type: text/html;charset=utf-8");
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
3
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
4 if (isset($_GET['books_id'])) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
5 $books_id = $_GET['books_id'];
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
6 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
7
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
8 include_once('./Lib_mb_utf8.php');
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
9
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
10 set_time_limit(0);
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
11 ini_set('memory_limit', '-1');
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
12
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
13 $link_mysql = mysql_connect($mysql_server, $mysql_user, $mysql_password);
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
14 mysql_query("SET NAMES utf8");
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
15
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
16 if (!$link_mysql) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
17 die('Could not connect: ' . mysql_error());
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
18 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
19
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
20 $db_selected = mysql_select_db($mysql_database, $link_mysql);
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
21 if (!$db_selected) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
22 die ('Can\'t use foo : ' . mysql_error());
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
23 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
24
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
25 if (!function_exists("GetSQLValueString")) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
26 function GetSQLValueString($theValue, $theType, $theDefinedValue = "", $theNotDefinedValue = "")
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
27 {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
28 $theValue = get_magic_quotes_gpc() ? stripslashes($theValue) : $theValue;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
29
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
30 $theValue = function_exists("mysql_real_escape_string") ? mysql_real_escape_string($theValue) : mysql_escape_string($theValue);
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
31
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
32 switch ($theType) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
33 case "text":
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
34 $theValue = ($theValue != "") ? "'" . $theValue . "'" : "NULL";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
35 break;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
36 case "long":
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
37 case "int":
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
38 $theValue = ($theValue != "") ? intval($theValue) : "NULL";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
39 break;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
40 case "double":
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
41 $theValue = ($theValue != "") ? "'" . doubleval($theValue) . "'" : "NULL";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
42 break;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
43 case "date":
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
44 $theValue = ($theValue != "") ? "'" . $theValue . "'" : "NULL";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
45 break;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
46 case "defined":
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
47 $theValue = ($theValue != "") ? $theDefinedValue : $theNotDefinedValue;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
48 break;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
49 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
50 return $theValue;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
51 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
52 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
53
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
54 $pbCount = 1;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
55 $imgCount = 1;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
56
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
57 function rep_pb_count($matches) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
58 global $pbCount;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
59 return "</page>\n<page n=\"".++$pbCount."\">\n";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
60 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
61 function rep_img_count($matches) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
62 global $imgCount;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
63 return "<img n=\"".$imgCount++."\"/>\n";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
64 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
65
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
66 $counter=0;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
67 $counterError=0;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
68 $counterIndex1=0;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
69 $counterIndex2=0;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
70 $counterMultiIndex=0;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
71 $counterSingleIndex=0;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
72 $counterNullIndex=0;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
73
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
74 $hasWordInIndex=array();
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
75 $hasWordInContent=array();
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
76
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
77 echo "<table width=\"100%\" border=\"1\">";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
78 $query = sprintf("SELECT * FROM books WHERE `id`=%s ORDER BY `id` ASC", $books_id);
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
79 $result = mysql_query($query);
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
80 while ($row = mysql_fetch_assoc($result)) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
81 $startPage = -1;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
82 $endPage = -1;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
83 $fileID = $row['id'];
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
84 $indexFilename = "../bookindexxml/".$fileID.".xml";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
85
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
86 $index = file_get_contents($indexFilename);
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
87 if ( !preg_match("/<Index>(.*?)<\/Index>/sm", $index, $matches) ) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
88 //echo "ERROR!".$fileID."\n";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
89 } else {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
90 $matchString = preg_replace("/\r\n/u", "", $matches[0]);
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
91
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
92 $counter++;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
93 if ( ! ($doc = simplexml_load_string( $matchString )) ) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
94 echo "ERROR!";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
95 exit;
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
96 } else {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
97
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
98 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
99
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
100 foreach ( $doc->children() as $child ) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
101 $tempArray = $child->attributes();
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
102 if ( $tempArray['Type'] == "UNKNOWN" ) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
103 echo "<tr bgcolor=\"#FFE4E1\"><font color=\"#FF0000\">";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
104 } else if ( $tempArray['Type'] == "FROMFULLTEXT" ) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
105 echo "<tr bgcolor=\"#FFFACD\"><font color=\"#000000\">";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
106 } else {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
107 echo "<tr bgcolor=\"#FFFFFF\"><font color=\"#000000\">";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
108 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
109 echo "<td>".$tempArray['Entry']."</td>";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
110 echo "<td>".$tempArray['StartPage']."(".$tempArray['StartRange'].")</td>";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
111 echo "<td>".$tempArray['EndPage']."</td>";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
112 echo "</font></tr>";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
113 foreach ( $child->children() as $child2 ) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
114 $tempArray = $child2->attributes();
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
115 if ( $tempArray['Type'] == "UNKNOWN" ) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
116 echo "<tr bgcolor=\"#FFE4E1\"><font color=\"#FF0000\">";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
117 } else if ( $tempArray['Type'] == "FROMFULLTEXT" ) {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
118 echo "<tr bgcolor=\"#FFFACD\"><font color=\"#000000\">";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
119 } else {
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
120 echo "<tr bgcolor=\"#FFFFFF\"><font color=\"#000000\">";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
121 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
122 echo "<td><span style=\"padding: 0 40px\">&nbsp;</span>".$tempArray['Entry']."</td>";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
123 echo "<td>".$tempArray['StartPage']."(".$tempArray['StartRange'].")</td>";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
124 echo "<td>".$tempArray['EndPage']."</td>";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
125 echo "</font></tr>";
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
126 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
127 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
128 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
129 }
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
130
b12c99b7c3f0 commit for previous development
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
131 ?>